app-editors/zed
The fast, collaborative code editor
-
zed-1.23.1_pre~amd64+X collab extensions-cli inspector +mimalloc neovim remote-server screen-capture test tracy +wayland +llvm_slot_22 debug
View
Download
Browse License: GPL-3+ BSD Apache-2.0 Apache-2.0-with-LLVM-exceptions BSD-2 BSD Boost-1.0 CC0-1.0 CDLA-Permissive-2.0 ISC LGPL-3 MIT MIT-0 MPL-2.0 UoI-NCSA Unicode-3.0 ZLIB BZIP2 Overlay: bentoo -
zed-1.22.0~amd64+X collab extensions-cli inspector +mimalloc neovim remote-server screen-capture test tracy +wayland +llvm_slot_22 debug
View
Download
Browse License: GPL-3+ BSD Apache-2.0 Apache-2.0-with-LLVM-exceptions BSD-2 BSD Boost-1.0 CC0-1.0 CDLA-Permissive-2.0 ISC LGPL-3 MIT MIT-0 MPL-2.0 UoI-NCSA Unicode-3.0 ZLIB BZIP2 Overlay: bentoo -
zed-1.22.0~amd64 ~arm64+llvm_slot_22 debug
View
Download
Browse License: GPL-3+ BSD Apache-2.0 Apache-2.0-with-LLVM-exceptions BSD-2 BSD Boost-1.0 CC0-1.0 ISC LGPL-3 MIT MIT-0 MPL-2.0 UoI-NCSA Unicode-3.0 ZLIB BZIP2 Overlay: robert7k -
zed-1.15.0~amd64 ~arm64+llvm_slot_22 debug
View
Download
Browse License: GPL-3+ BSD Apache-2.0 Apache-2.0-with-LLVM-exceptions BSD-2 BSD Boost-1.0 CC0-1.0 ISC LGPL-3 MIT MIT-0 MPL-2.0 UoI-NCSA Unicode-3.0 ZLIB BZIP2 Overlay: gentoo
ChangeLog
commit b83e006ffe00e7e34b5fa4a24d2cd90a6059555a
Author: lucascouts <lucascs@protonmail.com>
Date: Thu Oct 1 14:22:26 2026 -0300
app-editors/zeo-bin: name the distfile after $; app-editors/zed: drop a stray probe script
zeo-bin mirrors zeo's full version, and a zeo revbump is a different binary,
so its tarball now carries the revision: SRC_URI and S use $.
zed-patches/scripts/make-bin-release.sh names it the same way and refuses
to reuse a name R2 already serves with other bytes. At -r0 $ is $,
so the published file and the Manifest are unchanged; verified by a full
fetch-verify-unpack-install of zeo-bin-0.1.0_p20261001.
app-editors/zed/.probe.sh was a curl probe left behind on 2026-08-06 for
the long-gone 1.16.0_pre20260806-r1 snapshot; nothing references it.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
commit 001f5e999f6707b9de031d4619482d19d4e1ebd1
Author: lucascouts <lucascs@protonmail.com>
Date: Thu Oct 1 11:07:58 2026 -0300
app-editors/{zed,zeo,zeo-bin}: zed builds releases, the patched snapshot is zeo
app-editors/zed stops carrying the bentoo patch series and becomes the
source twin of zed-bin: upstream's tagged releases, unpatched, as
zed-1.22.0 (stable channel) and zed-1.23.1_pre (preview). The three
claude-agent USE flags leave it.
app-editors/zeo takes over what zed was -- the snapshot at 71456c40 with
the whole series, 0001-0031 -- plus the rebrand (0019-0023, 0025) applied
before every USE group. The rebrand patches had never been refreshed
since 2026-09-12, because they were not in the series refresh.sh reads;
they are now carried onto 71456c40, and 0021 lost one hunk whose function
upstream removed. The distfile is zed-$.tar.gz and ZED_PV is
gone: the autoupdate (base_from = "none") can only move the date and
EGIT_COMMIT, and a stale second version would name new bytes after an
already-manifested distfile.
app-editors/zeo-bin is new: the same build at the same PV, compiled for
x86-64-v3 under zed-patches/release/configroot and hosted at
distfiles.obentoo.org; PROVENANCE.txt inside names the commit, each patch
with its sha256, the USE set and the flags.
Only one of zed, zed-bin, zeo, zeo-bin installs at a time; the blockers
live in zeo and zeo-bin. A news item explains the switch and the
expected downgrade of an installed zed snapshot to a release.
Verified: check-sync (29 patches: series, files/ and source agree);
ebuild prepare for all four ebuilds; cargo check -p zed for zeo and both
releases; a full x86-64-v3 compile+install of zeo (no AVX-512 in the
binary); zeo-bin install from the R2 copy, whose sha256 matches the local
tarball; pkgcheck shows only the pre-existing llvm-r1 deprecation and
zed's stable/preview RedundantVersion; autoupdate --lint passes.
Not merged with emerge: no root here.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
commit 28a06bc71e3862e7ae61a786a271aa67f918e588
Author: lucascouts <lucascs@protonmail.com>
Date: Thu Oct 1 08:59:05 2026 -0300
app-editors/zed: 1.24.0_pre20261001-r1 at 71456c40
Upstream 71456c40 landed the same UTC day as 74c134a3 (06:12 -> 10:15),
so the date component did not change; bumped as -r1. SRC_URI already
renames to $, so the distfile name stays unique.
The commit touches only crates/gpui/src/keymap.rs and
platform/keystroke.rs. pycargoebuild from crates/zed left CRATES (1494),
GIT_CRATES (70) and LICENSE unchanged, and the src_prepare sed turns all
11 [patch.crates-io] entries into path= deps.
The patch series is carried over from 1.23.0_pre20260930, which the
two applier bumps to 1.24.0 skipped: 9 patches moved in offsets only
(+/- lines byte-identical), 14 are identical. Drops 4 md5-cache entries
left behind by earlier bumps.
Verified: ebuild prepare applies 23/23 with no fuzz; cargo check of the
15 patched crates is clean; the 108 tests the patches add pass
(108/108), with RUSTFLAGS=--cap-lints=warn (proptest) and an empty .git
in $ (debug asset lookup). Not built with emerge.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
commit 0bb39ebe93918341f348efc3651f40ada5afde66
Author: lucascouts <lucascs@protonmail.com>
Date: Thu Oct 1 08:05:57 2026 -0300
add(metadata/{md5-cache/app-editors/zed-1.24.0_pre20261001, md5-cache/app-misc/zeo-systray-0.2.1, md5-cache/dev-python/conda-26.9.0, md5-cache/dev-util/ai-memory-2.5.2, md5-cache/dev-util/ai-memory-bin-2.5.2, md5-cache/dev-util/mesa_clc-26.3.0_pre20261001, md5-cache/media-libs/mesa-26.3.0_pre20261001, md5-cache/media-libs/vulkan-layers-1.4.363_p20261001, md5-cache/net-misc/postman-bin-12.30.5, md5-cache/net-p2p/bisq-bin-1.10.9, md5-cache/sci-ml/ik_llama-cpp-0_pre20261001, md5-cache/sci-ml/llama-swap-261, md5-cache/sys-apps/ai-jail-2.5.0, md5-cache/sys-apps/ai-jail-bin-2.5.0}), mod(.autoupdate/packages.toml), up(app-editors/zed-1.24.0_pre20260930 -> 1.24.0_pre20261001, dev-python/conda-26.7.3 -> 26.9.0, dev-util/ai-memory{,-bin}-2.5.0 -> 2.5.2, media-libs/{mesa-26.3.0_pre20260930 -> 26.3.0_pre20261001, vulkan-layers-1.4.363_p20260930 -> 1.4.363_p20261001}, net-misc/postman-bin-12.30.4 -> 12.30.5, net-p2p/bisq-bin-1.10.8 -> 1.10.9, sci-ml/{ik_llama-cpp-0_pre20260930 -> 0_pre20261001, llama-swap-260 -> 261}, sys-apps/ai-jail{,-bin}-2.4.0 -> 2.5.0)
commit c1787b478187140e4b4439f3c363d13e3c35c30f
Author: lucascouts <lucascs@protonmail.com>
Date: Wed Sep 30 23:15:13 2026 -0300
add(metadata/{md5-cache/app-admin/awscli2-2.37.7, md5-cache/app-editors/kiro-1.2.4, md5-cache/app-editors/zed-1.24.0_pre20260930, md5-cache/app-editors/zed-bin-1.22.0, md5-cache/app-editors/zed-bin-1.23.1_pre, md5-cache/dev-util/ai-memory-2.5.0, md5-cache/dev-util/ai-memory-bin-2.5.0, md5-cache/dev-util/antigravity-hub-bin-2.19.1.6046815158665216, md5-cache/dev-util/claude-code-2.1.286, md5-cache/dev-util/cline-bin-3.0.67, md5-cache/dev-util/codex-0.159.3, md5-cache/dev-util/codex-bin-0.159.3, md5-cache/dev-util/codex-desktop-bin-26.928.31416, md5-cache/dev-util/copilot-cli-bin-1.0.90, md5-cache/dev-util/glslang-1.4.363.0_p20260930, md5-cache/dev-util/openchamber-bin-2.1.0, md5-cache/dev-util/opencode-bin-1.18.34, md5-cache/media-libs/harfbuzz-14.5.1, md5-cache/net-analyzer/netdata-2.12.0, md5-cache/sci-ml/lemonade-2026.40.0, md5-cache/sci-ml/lemonade-bin-2026.40.0, md5-cache/sys-apps/ai-jail-2.4.0, md5-cache/sys-apps/ai-jail-bin-2.4.0, md5-cache/www-client/brave-browser-1.96.60}), mod(.autoupdate/packages.toml), up(app-admin/awscli2-2.37.6 -> 2.37.7, app-editors/{kiro-1.1.70 -> 1.2.4, zed-1.23.0_pre20260930 -> 1.24.0_pre20260930, zed-bin-1.21.0 -> 1.22.0, zed-bin-1.22.0_pre -> 1.23.1_pre}, dev-util/{ai-memory{,-bin}-2.4.2 -> 2.5.0, codex{,-bin}-0.159.2 -> 0.159.3}, media-libs/harfbuzz-14.5.0 -> 14.5.1, net-analyzer/netdata-2.11.1 -> 2.12.0, sci-ml/lemonade{,-bin}-2026.39.1 -> 2026.40.0, sys-apps/ai-jail{,-bin}-2.2.1 -> 2.4.0, www-client/brave-browser-1.96.59 -> 1.96.60)
commit 2a69424bdef1b6a0b924305aeed87ef000fa2f64
Author: lucascouts <lucascs@protonmail.com>
Date: Wed Sep 30 16:41:33 2026 -0300
app-editors/zed: carry the patch series onto 1.23.0_pre20260930
The applier bumped to e8c6c67f without touching files/, and two patches
no longer applied, so src_prepare died on eapply:
- 0011: upstream grew the `use acp_thread::` list in thread_view.rs;
AccountUsageStatus is re-inserted into it.
- 0031: upstream appended a field to TurnFields; session_duration now
follows it.
Only their position moved. 14 others changed in offsets and index lines
only; 7 are identical. No revbump: the published ebuild never got past
src_prepare, so nothing broken was installed.
pycargoebuild from crates/zed left CRATES, GIT_CRATES and LICENSE
unchanged, and all seven src_prepare sed commits match Cargo.toml.
Verified: ebuild prepare applies 23/23; cargo check of the 15 patched
crates is clean; the 108 tests the patches add pass (108/108). Tests
ran with RUSTFLAGS=--cap-lints=warn, because proptest denies warnings
and rustc 1.98 now warns in it, and with an empty .git in $, because
debug builds locate their assets through it. Not built with emerge.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
commit 5d924b77b594a2823b9506afb1c25c5c8374ed6c
Author: lucascouts <lucascs@protonmail.com>
Date: Wed Sep 30 14:58:31 2026 -0300
add(metadata/{md5-cache/app-editors/vim-9.2.1152, md5-cache/app-editors/zed-1.23.0_pre20260930, md5-cache/dev-db/sqlitebrowser-3.13.99_p20260930, md5-cache/dev-util/cline-bin-3.0.66, md5-cache/dev-util/codex-desktop-bin-26.928.21956, md5-cache/dev-util/spirv-tools-1.4.363.0_p20260930, md5-cache/dev-util/vulkan-tools-1.4.364_p20260930, md5-cache/media-libs/vulkan-layers-1.4.363_p20260930, md5-cache/media-libs/vulkan-loader-1.4.364_p20260930, md5-cache/media-sound/audacity-bin-4.0.1, md5-cache/net-misc/postman-bin-12.30.4, md5-cache/net-vpn/obscura-1.181_p20260930, md5-cache/net-vpn/obscura-bin-1.181, md5-cache/sci-ml/ik_llama-cpp-0_pre20260930, md5-cache/www-client/librewolf-bin-157.0_p1, md5-cache/www-client/opera-136.0.6008.80}), mod(.autoupdate/packages.toml), up(app-editors/{vim-9.2.1125 -> 9.2.1152, zed-1.23.0_pre20260929-r1 -> 1.23.0_pre20260930}, dev-db/sqlitebrowser-3.13.99_p20260929 -> 3.13.99_p20260930, dev-util/{cline-bin-3.0.65 -> 3.0.66, codex-desktop-bin-26.928.20755 -> 26.928.21956, spirv-tools-1.4.363.0_p20260929 -> 1.4.363.0_p20260930, vulkan-tools-1.4.364_p20260925 -> 1.4.364_p20260930}, media-libs/{vulkan-layers-1.4.363_p20260929 -> 1.4.363_p20260930, vulkan-loader-1.4.364_p20260929 -> 1.4.364_p20260930}, media-sound/audacity-bin-4.0.0 -> 4.0.1, net-misc/postman-bin-12.30.3 -> 12.30.4, net-vpn/{obscura-1.180_p20260929 -> 1.181_p20260930, obscura-bin-1.180 -> 1.181}, sci-ml/ik_llama-cpp-0_pre20260929 -> 0_pre20260930, www-client/{librewolf-bin-156.0.1_p1 -> 157.0_p1, opera-136.0.6008.73 -> 136.0.6008.80})
commit 1b3095c9f1b8f55bfb16772d5803ba11d3c8e628
Author: lucascouts <lucascs@protonmail.com>
Date: Wed Sep 30 00:28:45 2026 -0300
app-editors/zed: bump to decbf641 as 1.23.0_pre20260929-r1
Same-day snapshot, six commits past 1dc88444. pycargoebuild from
crates/zed left CRATES and GIT_CRATES unchanged: 1494/1494 registry
crates and 32/32 git revs match Cargo.lock, and all eleven src_prepare
sed anchors still rewrite to path=.
The patch series is refreshed against decbf641 -- 3 changed (0002,
0008, 0016; offsets only), 20 identical. Also drops the md5-cache entry
of the replaced 0929 ebuild.
Verified: ebuild prepare applies 23/23; cargo metadata --offline exits
0; cargo check of the 15 patched crates is clean; cargo test runs 2865
passed, 3 failed. Two failures assert AcpBetaFeatureFlag is off, which
0001 forces on by design; the third, test_multi_workspace_session_restore,
is flaky upstream (1 in 6 runs without patches). Not built with emerge.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
commit ab641906129c746bb637153ac5f2c744f22de834
Author: lucascouts <lucascs@protonmail.com>
Date: Tue Sep 29 14:27:08 2026 -0300
app-editors/zed: 0024 opens a session in its own project, with its own agent
zed://agent?session=<id> now opens the thread in the window holding its
project and loads it with the agent on its metadata row, instead of the
focused window and the native agent ("no thread found with ID").
0012 and 0027 are regenerated for shifted hunk positions only.
Source: lucascouts/zed-patches db4d4d0.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
commit a43ed83d086645694f313c9eae3ae2eb72352e94
Author: lucascouts <lucascs@protonmail.com>
Date: Tue Sep 29 12:34:36 2026 -0300
app-editors/zed: bump to 1dc88444 as 1.23.0_pre20260929
CRATES regenerated with pycargoebuild from crates/zed: upstream dropped ten
registry crates in this range (portable-pty, nix 0.28, serial2,
x11-clipboard, winreg 0.10, crossbeam, cfg_aliases 0.1, json_dotpath,
shared_library, take-until). GIT_CRATES is unchanged, and all eleven
src_prepare sed anchors still match the [patch.crates-io] revs.
The patch series is refreshed against 1dc88444 -- 11 changed, 12
identical, offsets only plus one context line upstream renamed
(shell -> shell_program). Also drops the stale md5-cache entry of
0927-r1, which had no ebuild behind it.
Verified: CRATES 1494/1494 and GIT_CRATES 70/70 against Cargo.lock;
check-sync clean; ebuild prepare applies 23/23; cargo metadata --offline
exits 0 after prepare. Not compiled.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
commit e9ae0eb296f21bf9e85ccf0334263423b058e1a2
Author: lucascouts <lucascs@protonmail.com>
Date: Mon Sep 28 21:26:28 2026 -0300
app-editors/zed: carry the patch series onto 1.23.0_pre20260928
The autoupdate bumped the ebuild to dd510f99 with the 0927-r1 patches in
files/, and 0008 no longer applied there: upstream began moving the agent UI
onto ACP schema v2 in this range. The series is now refreshed against
dd510f99 -- 18 patches changed, 5 identical -- and 0026 gains a "Scroll to
Next User Message" button beside "previous".
check-sync: series <-> ebuild, patches <-> overlay and patches <-> source
all agree.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
commit e2891460957c2ef4175056790d10509a2611c36d
Author: lucascouts <lucascs@protonmail.com>
Date: Mon Sep 28 18:33:41 2026 -0300
add(metadata/{md5-cache/app-admin/awscli2-2.37.5, md5-cache/app-editors/cursor-3.22.12, md5-cache/app-editors/zed-1.23.0_pre20260928, md5-cache/app-misc/claude-desktop-bin-2.9939.4, md5-cache/dev-python/openai-3.20.0, md5-cache/dev-util/claude-code-2.1.284, md5-cache/dev-util/codex-0.158.0, md5-cache/dev-util/codex-bin-0.158.0, md5-cache/dev-util/codex-desktop-bin-26.924.50649, md5-cache/dev-util/copilot-cli-bin-1.0.89, md5-cache/dev-util/mesa_clc-26.3.0_pre20260928, md5-cache/dev-util/openchamber-bin-2.0.3, md5-cache/dev-util/opencode-bin-1.18.33, md5-cache/dev-util/typescript-language-server-6.0.1-r3, md5-cache/dev-util/vulkan-utility-libraries-1.4.364_p20260927, md5-cache/media-gfx/freecad-1.1.4, md5-cache/media-libs/libopenshot-1.0.1, md5-cache/media-libs/libopenshot-audio-1.0.1, md5-cache/media-libs/mesa-26.3.0_pre20260928, md5-cache/media-libs/vulkan-layers-1.4.363_p20260928, md5-cache/media-video/openshot-4.0.1, md5-cache/media-video/shotcut-26.9.27, md5-cache/net-misc/postman-bin-12.30.0, md5-cache/net-vpn/obscura-1.180_p20260928, md5-cache/sci-ml/ik_llama-cpp-0_pre20260928, md5-cache/sys-apps/flatpak-1.18.4, md5-cache/sys-apps/flatpak-1.19.2}), del(metadata/md5-cache/dev-util/typescript-language-server-6.0.1-r2), mod(.autoupdate/packages.toml), up(app-admin/awscli2-2.37.4 -> 2.37.5, app-editors/{cursor-3.22.7 -> 3.22.12, zed-1.23.0_pre20260927-r1 -> 1.23.0_pre20260928}, app-misc/claude-desktop-bin-2.7032.0 -> 2.9939.4, dev-python/openai-3.19.2 -> 3.20.0, dev-util/codex{,-bin}-0.157.1 -> 0.158.0, media-gfx/freecad-1.1.3-r1 -> 1.1.4, media-libs/{libopenshot-1.0.0 -> 1.0.1, libopenshot-audio-1.0.0 -> 1.0.1, mesa-26.3.0_pre20260927 -> 26.3.0_pre20260928, vulkan-layers-1.4.363_p20260927 -> 1.4.363_p20260928}, media-video/{openshot-4.0.0 -> 4.0.1, shotcut-26.8.1 -> 26.9.27}, net-misc/postman-bin-12.29.5 -> 12.30.0, net-vpn/obscura-1.180_p20260926 -> 1.180_p20260928, sci-ml/ik_llama-cpp-0_pre20260927 -> 0_pre20260928, sys-apps/{flatpak-1.18.3 -> 1.18.4, flatpak-1.19.0 -> 1.19.2})
commit 517e85258d16402e915bf00b3fd886d28ec836d8
Author: lucascouts <lucascs@protonmail.com>
Date: Sun Sep 27 21:07:30 2026 -0300
app-editors/zed: bump to 1a28cff4 as 1.23.0_pre20260927-r1
Same UTC day as the previous snapshot, hence the revision bump.
- EGIT_COMMIT e683fd7b -> 1a28cff4
- CRATES regenerated with pycargoebuild: add alloca, c-enum,
perf-event{2,-data,-open-sys2}; criterion 0.5 -> 0.8.2,
itertools 0.11 -> 0.13; drop wast, wiggle*, witx
- GIT_CRATES and the src_prepare [patch.crates-io] revs unchanged;
all 11 entries still rewrite to path=
- all 23 patches apply unchanged (zed-patches check-sync in sync)
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
commit 9a4ee5732bbb958e32fda5c3cb62f6852dc3989f
Author: lucascouts <lucascs@protonmail.com>
Date: Sun Sep 27 20:47:55 2026 -0300
mod(app-editors/zed-1.23.0_pre20260927)
commit 499a0becce08ec05a50c29b26dc28f6e0aed5609
Author: lucascouts <lucascs@protonmail.com>
Date: Sun Sep 27 16:48:35 2026 -0300
app-editors/zed, dev-util/ccusage-bin: bump; md5-cache cleanup
- app-editors/zed: 1.23.0_pre20260926-r1 -> 1.23.0_pre20260927
- dev-util/ccusage-bin: 20.0.24 -> 20.0.26
- dev-util/claude-agent-acp-plus: add the missing 0.21.0 md5-cache entry
- drop md5-cache entries left behind by these bumps
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
commit 787ec8988388f7d15f8d96b8a1dcbd05b656c3d5
Author: lucascouts <lucascs@protonmail.com>
Date: Sun Sep 27 14:59:17 2026 -0300
app-editors/zed: 0027 queues several question sets in one card
Several AskUserQuestion calls in one turn now share the pinned card, one
at a time, with "Set N of M" in its header, instead of stacking. Synced
from zed-patches 322f581.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
commit 9bd6b8a70003c06d69ae65f6bf9df6a569bf54eb
Author: lucascouts <lucascs@protonmail.com>
Date: Sun Sep 27 14:14:46 2026 -0300
app-editors/zed: 0030 shows a thread waiting for an answer to a question
Synced from zed-patches bc8ee0e.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
commit a0c3207f3ff2574d4fbd33b0127af8e3ea2d2e12
Author: lucascouts <lucascs@protonmail.com>
Date: Sun Sep 27 07:36:40 2026 -0300
app-editors/zed: 0029 shows a thread whose background work outlives its turn
Synced from zed-patches 2ee40d5. Inert until claude-agent-acp-plus 0.21.0,
which sends the count it reads.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
commit 6ab74ec1f7fe2d720e8d9a38a496c0c7c51f1965
Author: lucascouts <lucascs@protonmail.com>
Date: Sat Sep 26 23:22:57 2026 -0300
app-editors/zed: 0028 opens a folder dropped on the threads sidebar
Synced from zed-patches 2082628.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
commit c68b43c7959c0c12b364417840184da9be8f8ac5
Author: lucascouts <lucascs@protonmail.com>
Date: Sat Sep 26 20:22:39 2026 -0300
app-editors/zed: 0014 test fixture passes clippy
Synced from zed-patches 5fe82c5. Test-only; the built binary is unchanged.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
commit 3048899f8cf0e7186e9f8afac836175363c4b9e7
Author: lucascouts <lucascs@protonmail.com>
Date: Sat Sep 26 19:17:27 2026 -0300
app-editors/zed: bump to 1.23.0_pre20260926-r1 (bda9c0bd)
Same-day snapshot, hence -r1. CRATES and GIT_CRATES regenerated with
pycargoebuild and came out identical; the src_prepare git-dep rewrites
still hit 11/11 entries of [patch.crates-io].
Patch series refreshed onto bda9c0bd: 0016 and 0027 moved by offset
only, the other 17 are byte-identical. cargo check of the 12 crates the
patches touch is clean.
md5-cache: drops the stale _pre20260925-r1 entry left behind with no
ebuild.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
commit 7d567024078e593aebdd12e6283f1c730f3bba23
Author: lucascouts <lucascs@protonmail.com>
Date: Sat Sep 26 18:49:32 2026 -0300
app-editors/zed: 0026 floats copy and scroll-to-bottom too
Synced from zed-patches 4c0e94a.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
commit 0cf8e84ecd8815e5a3025bcbc02ecf2836b213df
Author: lucascouts <lucascs@protonmail.com>
Date: Sat Sep 26 17:12:47 2026 -0300
app-editors/zed: 0002 advertises close_tab
0002 now lists close_tab in tools/list. It was dispatched and never
advertised (parity report item 32): the handler existed, but the CLI only
calls tools the server declares. The descriptor follows the VS Code
extension's registration -- a required `tab_name` string -- and the list
moved into a free tool_descriptors() so a plain test can hold it against
the dispatched names: every_dispatched_tool_is_advertised is red without
the descriptor ("`close_tab` is dispatched but not advertised") and green
with it.
Verified against 70e686c2a377 with the whole series applied, the tree
byte-identical to what the series produces: claude_code_ide 17/17 tests;
cargo check --all-targets -p claude_code_ide -p zed, no errors. Two
clippy lints under -D warnings (result_large_err, manual_inspect) and
rustfmt drift in three of the crate's files predate this change and are
left for a separate commit.
No revbump: the fix rides the next zed version bump rather than forcing
a rebuild of the editor for one tool descriptor.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
commit ba3b5d0571dab0da13f9749be88bef5d373f7a6e
Author: lucascouts <lucascs@protonmail.com>
Date: Sat Sep 26 14:46:25 2026 -0300
app-editors/zed: 0014 tests follow the MessageContent rework
Upstream #64667 (a405fb91d5) made UserMessage.content a MessageContent
that owns the original SDK blocks and dropped the `chunks` field. 0014's
two test fixtures still built the old shape, so `cargo check -p agent_ui
--all-targets` failed with four errors (E0599, E0560) -- in test code
only: verify.sh applies patches, it does not compile them, and the
package build never compiles #[cfg(test)].
The fixtures now use MessageContent::new / MessageContent::default(), as
upstream's own acp_thread test does. No revbump: nothing installed
changes.
Verified against 70e686c2a377 with the whole series applied: cargo check
--all-targets over the 28 crates the series touches, 0 errors, 0 warnings;
0014's four tests pass; the series file reproduces that tree exactly.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
commit 8def768d6922b143d0462fa3a25c8f54fc0bb4a8
Author: lucascouts <lucascs@protonmail.com>
Date: Sat Sep 26 13:10:27 2026 -0300
add(metadata/{md5-cache/app-editors/zed-1.23.0_pre20260926, md5-cache/dev-util/claude-agent-acp-plus-0.20.2, md5-cache/dev-util/claude-agent-acp-tui-0.14.2, md5-cache/dev-util/codex-0.157.1, md5-cache/dev-util/codex-bin-0.157.1, md5-cache/dev-util/codex-desktop-bin-26.924.22138, md5-cache/dev-util/mesa_clc-26.3.0_pre20260926, md5-cache/dev-util/openchamber-bin-2.0.2, md5-cache/dev-util/qwen-code-0.24.6, md5-cache/media-libs/mesa-26.3.0_pre20260926, md5-cache/media-libs/vulkan-layers-1.4.363_p20260926, md5-cache/net-vpn/obscura-1.180_p20260926, md5-cache/sci-ml/ik_llama-cpp-0_pre20260926, md5-cache/sci-ml/koboldcpp-1.122, md5-cache/sci-ml/llama-swap-260}), mod(.autoupdate/packages.toml), up(app-editors/zed-1.23.0_pre20260925-r1 -> 1.23.0_pre20260926, dev-util/codex{,-bin}-0.157.0 -> 0.157.1, media-libs/{mesa-26.3.0_pre20260925 -> 26.3.0_pre20260926, vulkan-layers-1.4.363_p20260925 -> 1.4.363_p20260926}, net-vpn/obscura-1.180_p20260924 -> 1.180_p20260926, sci-ml/{ik_llama-cpp-0_pre20260925 -> 0_pre20260926, koboldcpp-1.121 -> 1.122, llama-swap-258 -> 260})
commit c1194e63c7aaa0871edf1babcab1269753436ae1
Author: lucascouts <lucascs@protonmail.com>
Date: Fri Sep 25 21:45:22 2026 -0300
app-editors/zed: bump to 1.23.0_pre20260925-r1 (933d8d93)
Same-day snapshot, hence -r1. CRATES and GIT_CRATES regenerated with
pycargoebuild and came out identical; the src_prepare git-dep rewrites
still hit 11/11 entries of [patch.crates-io].
Patch series refreshed onto 933d8d93: 0010 needed a manual rebase of one
hunk after upstream turned LanguageModel's provider_id()/id() accessors
into fields (#64794); same shape as before (+516/-157, 52 hunks). The rest
moved by offset only. cargo check of the 12 crates the patches touch is
clean.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
commit ba64ef5c3603ce4bd3de5a7a94d6847b22debf02
Author: lucascouts <lucascs@protonmail.com>
Date: Fri Sep 25 07:27:59 2026 -0300
bump 10 packages from the 2026-09-25 autoupdate sweep
- app-admin/awscli2: 2.37.2 -> 2.37.3
- app-editors/zed: 1.23.0_pre20260924 -> 1.23.0_pre20260925 (a84858ac);
patch series refreshed, yawc moved from GIT_CRATES to CRATES
- app-editors/cursor: 3.21.18 -> 3.22.7
- dev-util/codex{,-bin}: 0.156.1 -> 0.157.0 (crossterm GIT_CRATES rev bumped)
- dev-util/vulkan-utility-libraries: 1.4.363_p20260925 -> 1.4.364_p20260925
(v1.4.364 tags the commit already carried; keeps the _p scheme of its siblings)
- media-libs/vulkan-layers: 1.4.363_p20260924 -> 1.4.363_p20260925 (fd8605f9);
glslang, vulkan-headers and vulkan-utility-libraries floors raised per known_good.json
- net-misc/postman-bin: 12.29.2 -> 12.29.4
- sci-ml/ik_llama-cpp: 0_pre20260924 -> 0_pre20260925 (6475e58b)
- sci-ml/llama-swap: 257 -> 258
media-libs/tlottie is left at 0.1.0_pre20260911: tdesktop 7.2.9 still pins
31f1b542, so the proposed 1.0.6_pre20260917 is main-branch drift the hold blocks.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
commit d1e3a81e58f5014fcb4b1ce52b279dd76a1a9ee7
Author: lucascouts <lucascs@protonmail.com>
Date: Fri Sep 25 07:05:29 2026 -0300
app-editors/zed: refresh the patch series onto 1.23.0_pre20260924
The autoupdate bump to 2c4bc2d7 left two patches conflicting, so the
package died in src_prepare:
- 0009: upstream now assigns session.config_options before applying the
defaults on session load. Build ConfigOptions once right after
config_state() and hand the session a clone; the clone shares the watch
sender through Rc, so default-applied options still wake the UI.
- 0011: upstream dropped PlanEntry from the acp_thread import list in
thread_view.rs; AccountUsageStatus is re-added to the new list.
Both keep their original shape (+59/-26 and +464/-3). The other ten
changed files only moved offsets. Generated from zed-patches
patches/zed-1.23.0_pre20260924.
verify.sh: 19/19 apply cumulatively. check-sync: in sync.
ebuild clean prepare: all 19 applied. cargo check --offline on the 12
crates the patches touch: clean. cargo test -p agent_servers not run:
the git proptest dev-dependency fails to build with rustc 1.98, which
RUST_MIN_VER requires -- unrelated to the patches.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
commit 3d54e2b69dec2a49e62250157d6c88f92d6a2b68
Author: lucascouts <lucascs@protonmail.com>
Date: Thu Sep 24 14:24:16 2026 -0300
bump: doctl, devin-desktop-bin, zed, sentry-native, cline-bin, trufflehog, vulkan-layers, ik_llama-cpp, ollama, ollama-bin, opera; add scripts/go-vendor.sh
dev-vcs/trufflehog 3.97.9 changed go.mod (go-osc52, mimetype), so the
reused 3.97.5 vendor tree no longer matched and -mod=vendor would fail.
Regenerate the vendor tarball for 3.97.9 (uploaded to distfiles.obentoo.org)
and point VENDOR_P at it.
scripts/go-vendor.sh checks whether a Go package's hosted vendor tarball
still matches its go.mod (--check exits 1 when stale) and regenerates it
otherwise. The autoupdate applier cannot see this; run it after bumping any
vendor-tarball Go package. Records and CLAUDE.md point to it; the reviewdog
record no longer claims a vendor tarball it does not use.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
commit 411c9c7a0d28e074dca09d1a95c24ea1d5d0b112
Author: lucascouts <lucascs@protonmail.com>
Date: Wed Sep 23 16:21:47 2026 -0300
add(metadata/{md5-cache/app-editors/zed-1.23.0_pre20260923, md5-cache/app-editors/zed-bin-1.21.0, md5-cache/app-editors/zed-bin-1.22.0_pre, md5-cache/dev-util/claude-agent-acp-plus-0.20.1, md5-cache/dev-util/claude-agent-acp-tui-0.14.0, md5-cache/dev-util/goose-1.52.0, md5-cache/dev-util/goose-bin-1.52.0, md5-cache/dev-util/spirv-headers-1.4.357.0_p20260923, md5-cache/dev-vcs/trufflehog-3.97.8, md5-cache/sci-ml/lemonade-2026.39.1, md5-cache/sci-ml/lemonade-bin-2026.39.1, md5-cache/sys-apps/ai-jail-2.1.0, md5-cache/sys-apps/ai-jail-bin-2.1.0, md5-cache/sys-apps/pnpm-12.6.0}), mod(.autoupdate/packages.toml), up(app-editors/{zed-1.22.0_pre20260923-r1 -> 1.23.0_pre20260923, zed-bin-1.20.2 -> 1.21.0, zed-bin-1.21.0_pre -> 1.22.0_pre}, dev-util/goose{,-bin}-1.51.0 -> 1.52.0, dev-vcs/trufflehog-3.97.6 -> 3.97.8, sci-ml/{lemonade-11.9.0-r1 -> 2026.39.1, lemonade-bin-11.9.0 -> 2026.39.1}, sys-apps/ai-jail{,-bin}-2.0.0 -> 2.1.0)
commit f410b7b63202759b2df79a6ed037d324840a1740
Author: lucascouts <lucascs@protonmail.com>
Date: Wed Sep 23 12:00:36 2026 -0300
app-editors/zed: bump to 1.22.0_pre20260923-r1 (bf89346c)
Same UTC day as 7fecbb2c, so the date component alone does not move and this is
a revision. crates/zed/Cargo.toml still reads 1.22.0, rust-toolchain.toml still
pins 1.98.1 and corgi.toml still pins webrtc 0001d84-4, so PV base, RUST_MIN_VER
and WEBRTC_COMMIT all stand. pycargoebuild re-run from crates/zed reproduced
CRATES and GIT_CRATES byte for byte -- both WARNINGs printed, which is what
tells a real run apart from the silent no-op in the virtual workspace root. All
eleven local *_COMMIT sed anchors still match the new root Cargo.toml exactly
once.
The cost was in the patch series. 0011's agent_diff.rs hunk conflicted because
upstream inserted AcpThreadEvent::NoticesUpdated into the very arm list the hunk
adds AccountUsageUpdated to, so the insertion point lost its trailing context.
The change itself is one line and untouched; only the arm above it moved. Fixed
through a temporary source series rather than by hand-editing @@ counts, so all
nineteen were refreshed from one corrected baseline: 11 changed, 8 identical.
Verified: verify.sh, sync-overlay and check-sync agree on all three relations;
ebuild ... prepare applies all nineteen (>>> Source prepared); the 11 seds leave
zero [patch.crates-io] entries on 'git ='; cargo check --offline -p zed
-p remote_server --features gpui/inspector exits 0 in 2m25 with zero errors;
md5-cache _md5_ matches the ebuild and the two orphan entries were swept;
pkgcheck reports only the pre-existing deprecated-llvm-r1 finding. Not merged
with a build: this host has no sudo.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
commit 687194884611661c9a02a3ad7d0f745a332cb7f4
Author: lucascouts <lucascs@protonmail.com>
Date: Wed Sep 23 05:28:34 2026 -0300
autoupdate: apply the 2026-09-23 batch
app-admin/doctl 1.171.0 -> 1.171.2, app-editors/zed 1.22.0_pre20260922-r1
-> 1.22.0_pre20260923, dev-python/openai 3.18.0 -> 3.19.0,
dev-python/uv-bin 0.12.17 -> 0.12.18, dev-util/codex-bin 0.156.0 ->
0.156.1, dev-util/codex-desktop-bin 26.917.51856 -> 26.917.61114,
dev-util/mesa_clc and media-libs/mesa 26.3.0_pre20260922 ->
26.3.0_pre20260923, net-misc/postman-bin 12.29.1 -> 12.29.2,
sci-ml/ik_llama-cpp 0_pre20260922 -> 0_pre20260923.
Each new md5-cache entry was checked against its ebuild: all ten _md5_
lines match md5sum, so no bump carries a stale cache row.
dev-util/codex is not here; its 0.156.1 went in on its own commit because
it needed a source fix to build.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
commit 145bf9d6ca6f70867ceed9f4685d515bb5fb49cf
Author: lucascouts <lucascs@protonmail.com>
Date: Tue Sep 22 21:29:34 2026 -0300
app-editors/zed: the 0017 comment names the crate that is actually linked
The PATCHES comment said the ACP compaction notification was unreachable here
because agent-client-protocol 2.0.0 carries zero occurrences of "compaction".
Measured against the packaged commit 96e95eda: the linked crate is 2.2.0 with
schema 1.9.1, acp_thread.rs matches acp::SessionUpdate::CompactionUpdate, and Zed
constructs acp::CompactionUpdate itself.
The correction is recorded in the comment rather than swapped in silently, so a
reader can tell the old premise was false rather than merely outdated.
files/0017-native-compaction-entry.patch is re-synced from zed-patches in the
same commit: its header carried the same claim, and refresh.sh preserves
everything above the --- on every bump, which is what made that copy permanent.
verify.sh ran before the sync, not after -- 19 of 19 apply onto 96e95eda -- so a
header edit that broke the patch could not be found inside this commit's window.
The patch diff below the --- is byte-identical, 25023 bytes before and after.
Fourth and last copy of the claim. The others: the adapter's
context-compaction-meta.ts, the patch header, and the series comment, which no
plan had budgeted for.
check-sync.sh agrees on all three relations, at 19 patches.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
commit e6a16b72862d879d87301f9ca1b75338cea554e1
Author: lucascouts <lucascs@protonmail.com>
Date: Tue Sep 22 20:46:30 2026 -0300
app-editors/zed: bump to 1.22.0_pre20260922-r1 (96e95eda)
Same UTC day as f25434f3, so the date component alone does not move and this is
a revision. Cargo.lock, the root Cargo.toml, crates/zed/Cargo.toml and
rust-toolchain.toml are all byte-identical between the two commits, and
pycargoebuild re-run from crates/zed reproduced the ebuild byte for byte; all
seven local *_COMMIT still resolve against GIT_CRATES.
The patch series needed a rebase of 0027, whose agent_panel.rs import hunk lost
its context; nine of the nineteen patches changed, the other eight by offset.
Verified: ebuild ... prepare applies all nineteen (>>> Source prepared); 11/11
[patch.crates-io] seds leave zero entries on 'git ='; cargo check --offline
-p zed -p remote_server --features gpui/inspector exits 0; md5-cache matches;
pkgcheck reports only the pre-existing deprecated-llvm-r1 finding. Not merged
with a build: this host has no sudo.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
commit 197a914dd92c344fa0ac77865de5471e59a6eec6
Author: lucascouts <lucascs@protonmail.com>
Date: Mon Sep 21 22:25:52 2026 -0300
mod(.autoupdate/packages.toml), mod(app-editors/zed/files/*), up(app-editors/zed-1.22.0_pre20260921 -> 1.22.0_pre20260922)
bcf6582c -> f25434f3 (2026-09-22T00:11 UTC). CRATES regenerated with
pycargoebuild from crates/zed: three rows moved (agent-client-protocol and
-derive 2.1.0 -> 2.2.0, -schema 1.7.0 -> 1.9.1); GIT_CRATES (71) and the
LICENSE+= block are byte identical, as are rust-toolchain.toml and
crates/zed/Cargo.toml.
Series carried forward through zed-patches: all 19 patches apply, 12 changed by
offset only. ebuild ... prepare applies all nineteen for real; 11/11
src_prepare seds match the new root Cargo.toml exactly once; the four USE-flag
seds yield valid TOML in all sixteen combinations; cargo check --offline
-p zed -p remote_server --features gpui/inspector exits 0 in 1m34; pkgcheck
reports only the pre-existing deprecated-llvm-r1 finding.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
commit e677e5b5843924eee31e8eb332c5c547306b14b8
Author: lucascouts <lucascs@protonmail.com>
Date: Mon Sep 21 20:44:51 2026 -0300
add(metadata/{md5-cache/app-admin/awscli2-2.36.50, md5-cache/app-editors/vim-9.2.1125, md5-cache/app-editors/vim-core-9.2.1125, md5-cache/dev-util/copilot-cli-bin-1.0.87, md5-cache/dev-util/crush-bin-0.96.1, md5-cache/dev-util/kimi-cli-bin-1.51.0, md5-cache/dev-util/opencode-bin-1.18.32, md5-cache/dev-util/qwen-code-0.24.3, md5-cache/dev-util/spirv-tools-1.4.357.0_p20260921, md5-cache/media-libs/vulkan-loader-1.4.363_p20260921, md5-cache/sys-apps/ai-jail-2.0.0, md5-cache/sys-apps/ai-jail-bin-2.0.0}), mod(app-editors/zed-1.22.0_pre20260921), mod(metadata/md5-cache/app-editors/zed-1.22.0_pre20260921, .autoupdate/packages.toml), up(app-admin/awscli2-2.36.49 -> 2.36.50, app-editors/{vim, vim-core}-9.2.1122 -> 9.2.1125, dev-util/{copilot-cli-bin-1.0.86 -> 1.0.87, crush-bin-0.96.0 -> 0.96.1, kimi-cli-bin-1.50.0 -> 1.51.0, opencode-bin-1.18.31 -> 1.18.32, qwen-code-0.24.2 -> 0.24.3, spirv-tools-1.4.357.0_p20260918 -> 1.4.357.0_p20260921}, media-libs/vulkan-loader-1.4.363_p20260918 -> 1.4.363_p20260921, sys-apps/ai-jail{,-bin}-1.22.0 -> 2.0.0)
commit d230559ad96e9a970a221f36d008328540871ddb
Author: lucascouts <lucascs@protonmail.com>
Date: Mon Sep 21 03:55:52 2026 -0300
app-editors/zed: bump to 1.22.0_pre20260921 (bcf6582c)
Cargo.lock, the root Cargo.toml and rust-toolchain.toml are byte-identical
between 916fc2b8 and bcf6582c, and pycargoebuild re-run from crates/zed
reproduced the ebuild byte for byte. All seven src_prepare git-dep revisions
still resolve against GIT_CRATES, and EGIT_COMMIT occurs exactly once: this was
a manual bump, so the applier's 40-hex rewrite never ran.
Fix a pre-existing defect the bump uncovered. USE=inspector passed the cargo
feature zed/inspector, which no longer exists: upstream reverted "Implement
`inspector` flag" (#62920) on 2026-09-15, so crates/zed/Cargo.toml has carried
no such key since _pre20260916 and cargo answers "none of the selected packages
contains this feature". The flag is default-off, which is why four bumps went by
without it surfacing. Point it at gpui/inspector, where the switch lived before
that PR and still does; gpui is a direct dependency of the zed crate, so the
dep/feature form resolves. zed/tracy and zed/mimalloc were re-checked and are
both still valid.
The patch series was three versions behind (files/ was generated against
_pre20260918-r1 by the applier) and was refreshed through zed-patches; all 18
apply and only 0018 changed, by offset.
Verified: `ebuild ... prepare` applies all eighteen for real; 11/11 sed patterns
match the new Cargo.toml exactly once, leaving zero [patch.crates-io] entries on
`git =` and all ten WORKDIR targets present; the four USE-flag seds produce
valid TOML in all sixteen wayland/X/screen-capture/neovim combinations;
`cargo check --offline -p zed -p remote_server --features gpui/inspector` exits
0 in 1m49 with zero errors; md5-cache _md5_ matches the ebuild and egencache
also swept the orphan _pre20260918-r1 entry; pkgcheck reports only the
pre-existing deprecated-llvm-r1 finding. Not merged with a build: this host has
no sudo, so no emerge ran.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
commit 273eb22a2316f7196deac76a37cd1fab9cece7a0
Author: lucascouts <lucascs@protonmail.com>
Date: Sat Sep 19 14:38:37 2026 -0300
add(metadata/{md5-cache/app-editors/vim-9.2.1122, md5-cache/app-editors/vim-core-9.2.1122, md5-cache/app-editors/zed-1.22.0_pre20260919, md5-cache/dev-libs/sentry-native-0.16.8, md5-cache/dev-util/claude-code-2.1.278, md5-cache/dev-util/qwen-code-0.24.1, md5-cache/media-libs/vulkan-layers-1.4.363_p20260919, md5-cache/sci-ml/ik_llama-cpp-0_pre20260919, md5-cache/sys-apps/ai-jail-1.22.0, md5-cache/sys-apps/ai-jail-bin-1.22.0}), mod(.autoupdate/packages.toml), up(app-editors/{vim-9.2.1119 -> 9.2.1122, vim-core-9.2.1119 -> 9.2.1122, zed-1.22.0_pre20260918-r1 -> 1.22.0_pre20260919}, dev-libs/sentry-native-0.16.6 -> 0.16.8, dev-util/{claude-code-2.1.277 -> 2.1.278, qwen-code-0.24.0 -> 0.24.1}, media-libs/vulkan-layers-1.4.363_p20260918 -> 1.4.363_p20260919, sci-ml/ik_llama-cpp-0_pre20260918 -> 0_pre20260919, sys-apps/ai-jail{,-bin}-1.21.0 -> 1.22.0)
commit 2306b89eab662662167959d7197955ec7b0ab831
Author: lucascouts <lucascs@protonmail.com>
Date: Fri Sep 18 19:52:20 2026 -0300
app-editors/zed: bump to 1.22.0_pre20260918-r1 (72b060af)
Same-day revision: upstream moved from 650a8d1b to 72b060af on 2026-09-18,
so the snapshot date is unchanged and the bump is a -r1.
CRATES regenerated with pycargoebuild from crates/zed (the tree root is a
virtual workspace and would be a silent no-op). Five crates dropped, none
added: hyper-rustls@0.24.2, rustls@0.21.12, rustls-webpki@0.101.7,
sct@0.7.1, tokio-rustls@0.24.1. GIT_CRATES is unchanged, and so are the
seven local *_COMMIT values in src_prepare -- the sha appears exactly once
in the ebuild.
Patch series refreshed onto the new source (all 18). Upstream added
Focusable to the gpui import block in conversation_view/elicitation.rs,
which conflicted with hunk 1 of 0006-elicitation-option-previews; the whole
series was re-refreshed from a corrected temporary source rather than
rebasing that one patch in files/, so the four patches that had only drifted
in offsets are carried forward too.
Verified without a full build:
- CRATES and GIT_CRATES diffed against the new Cargo.lock: no additions,
no removals, no stale revs.
- The src_prepare sed block run against the new Cargo.toml rewrites 11 of
11 [patch.crates-io] deps to path=, leaving no git= behind.
- The wayland/x11/screen-capture seds produce valid TOML in all four
combinations.
- ebuild ... clean prepare applies every patch, and cargo metadata
--offline then resolves the whole graph (exit 0).
Not covered: src_compile was not run.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
commit 14d7823bd90004bfa2ffdc73920a7574443badad
Author: lucascouts <lucascs@protonmail.com>
Date: Fri Sep 18 16:40:22 2026 -0300
add(metadata/{md5-cache/app-editors/cursor-3.21.13, md5-cache/app-editors/vim-9.2.1119, md5-cache/app-editors/vim-core-9.2.1119, md5-cache/app-editors/zed-1.22.0_pre20260918, md5-cache/app-misc/claude-desktop-bin-2.2553.1, md5-cache/app-text/calibre-9.15.0, md5-cache/dev-python/openai-3.16.0, md5-cache/dev-python/tavily-python-0.8.4, md5-cache/dev-python/uv-bin-0.12.16, md5-cache/dev-util/ccusage-bin-20.0.23, md5-cache/dev-util/claude-code-2.1.277, md5-cache/dev-util/codex-0.155.0, md5-cache/dev-util/codex-bin-0.155.0, md5-cache/dev-util/codex-desktop-bin-26.915.31945, md5-cache/dev-util/copilot-cli-bin-1.0.86, md5-cache/dev-util/glslang-1.4.357.0_p20260918, md5-cache/dev-util/kilocode-bin-7.7.5, md5-cache/dev-util/mesa_clc-26.3.0_pre20260918, md5-cache/dev-util/openchamber-bin-1.24.1, md5-cache/dev-util/reviewdog-0.21.2, md5-cache/dev-util/spirv-headers-1.4.357.0_p20260917, md5-cache/dev-util/spirv-tools-1.4.357.0_p20260918, md5-cache/dev-util/vulkan-headers-1.4.363_p20260918, md5-cache/dev-util/vulkan-tools-1.4.363_p20260918, md5-cache/dev-util/vulkan-utility-libraries-1.4.363, md5-cache/media-libs/mesa-26.3.0_pre20260918, md5-cache/media-libs/vulkan-layers-1.4.363_p20260918, md5-cache/media-libs/vulkan-loader-1.4.363_p20260918, md5-cache/net-misc/postman-bin-12.28.6, md5-cache/sci-ml/ik_llama-cpp-0_pre20260918, md5-cache/sci-ml/localai-4.10.0, md5-cache/sci-ml/ollama-0.34.2, md5-cache/sci-ml/ollama-bin-0.34.2, md5-cache/www-client/brave-browser-1.95.104, md5-cache/www-client/chromium-153.0.8010.52}), mod(.autoupdate/packages.toml), up(app-editors/{cursor-3.20.21 -> 3.21.13, vim-9.2.1117 -> 9.2.1119, vim-core-9.2.1117 -> 9.2.1119, zed-1.22.0_pre20260917-r1 -> 1.22.0_pre20260918}, app-misc/claude-desktop-bin-2.110.1 -> 2.2553.1, app-text/calibre-9.14.0 -> 9.15.0, dev-python/{openai-3.14.1 -> 3.16.0, tavily-python-0.8.3 -> 0.8.4, uv-bin-0.12.15 -> 0.12.16}, dev-util/codex{,-bin}-0.154.0 -> 0.155.0, media-libs/{mesa-26.3.0_pre20260917 -> 26.3.0_pre20260918, vulkan-layers-1.4.362_p20260917 -> 1.4.363_p20260918, vulkan-loader-1.4.362_p20260915 -> 1.4.363_p20260918}, net-misc/postman-bin-12.28.3 -> 12.28.6, sci-ml/ollama{,-bin}-0.34.1 -> 0.34.2, www-client/{brave-browser-1.95.102 -> 1.95.104, chromium-153.0.8010.47 -> 153.0.8010.52})
commit c6f4d45316ba48130f83303a2262f7c3238ab149
Author: lucascouts <lucascs@protonmail.com>
Date: Thu Sep 17 07:27:50 2026 -0300
app-editors/zed: bump to 1.22.0_pre20260917-r1 (b6171bcc)
Same UTC day as the packaged _pre20260917 (4b47ceb9), so this is a same-day
revision rather than a new date. SRC_URI already renames to $, so the two
distfiles do not collide.
CRATES and GIT_CRATES regenerated with pycargoebuild from crates/zed (the repo
root is a virtual workspace), EGIT_COMMIT edited afterwards because the
in-place mode rewrites everything else from the input. Against the new
Cargo.lock both blocks match exactly: 0 registry crates missing or extra, 71
git crates with identical revs. CRATES did move - the include-flate/libflate/
zstd group came in, the symphonia-* set went out.
The seven src_prepare *_COMMIT literals needed no change (grep -c of
EGIT_COMMIT is 1, and every rev the sed anchors on is still in the source
Cargo.toml): running the sed block over the new Cargo.toml turns 11 of 11
[patch.crates-io] entries into path=, leaving no git= behind. The USE-flag seds
still yield valid TOML in all eight wayland/X/screen-capture combinations.
Patch series refreshed from zed-patches onto b6171bcc: 18 of 18 apply, four
changed on offsets only, no orphans.
Verified: ebuild clean prepare applies all 18 patches, and cargo metadata
--offline in $ exits 0, so the whole graph resolves without network.
src_compile was not run.
Also drops the stale md5-cache entry for _pre20260916, which the previous bump
left behind.
commit acf840a35a3b098459c5c0be21582bb5a8da7420
Author: lucascouts <lucascs@protonmail.com>
Date: Thu Sep 17 00:16:42 2026 -0300
add(metadata/{md5-cache/app-editors/zed-1.22.0_pre20260917, md5-cache/dev-util/claude-code-2.1.274, md5-cache/dev-util/mesa_clc-26.3.0_pre20260917, md5-cache/dev-util/openchamber-bin-1.24.0, md5-cache/media-libs/mesa-26.3.0_pre20260917}), mod(.autoupdate/packages.toml), up(app-editors/zed-1.22.0_pre20260916 -> 1.22.0_pre20260917, dev-util/{claude-code-2.1.273 -> 2.1.274, mesa_clc-26.3.0_pre20260916 -> 26.3.0_pre20260917, openchamber-bin-1.23.2 -> 1.24.0}, media-libs/mesa-26.3.0_pre20260916 -> 26.3.0_pre20260917)
commit cd631aeb82cf3004b6097b922f3342a3f5d4b7bf
Author: lucascouts <lucascs@protonmail.com>
Date: Wed Sep 16 18:34:40 2026 -0300
add(metadata/{md5-cache/app-admin/awscli2-2.36.47, md5-cache/app-editors/devin-desktop-bin-3.10.31, md5-cache/app-editors/vim-9.2.1116, md5-cache/app-editors/vim-core-9.2.1116, md5-cache/app-editors/zed-1.22.0_pre20260916, md5-cache/app-editors/zed-bin-1.20.1, md5-cache/app-editors/zed-bin-1.21.0_pre, md5-cache/dev-util/ai-memory-2.3.0, md5-cache/dev-util/ai-memory-bin-2.3.0, md5-cache/dev-util/glslang-1.4.357.0_p20260916, md5-cache/dev-vcs/trufflehog-3.97.5, md5-cache/net-analyzer/netdata-2.11.1, md5-cache/net-libs/nodejs-26.9.0, md5-cache/net-p2p/bisq-bin-1.10.8, md5-cache/www-client/brave-browser-1.95.102}), mod(.autoupdate/packages.toml), up(app-admin/awscli2-2.36.46 -> 2.36.47, app-editors/{devin-desktop-bin-3.10.27 -> 3.10.31, vim-9.2.1112 -> 9.2.1116, vim-core-9.2.1112 -> 9.2.1116, zed-1.21.0_pre20260916 -> 1.22.0_pre20260916, zed-bin-1.19.2 -> 1.20.1, zed-bin-1.20.1_pre -> 1.21.0_pre}, dev-util/ai-memory{,-bin}-2.2.2 -> 2.3.0, dev-vcs/trufflehog-3.97.4 -> 3.97.5, net-analyzer/netdata-2.11.0 -> 2.11.1, net-libs/nodejs-26.8.2 -> 26.9.0, net-p2p/bisq-bin-1.10.7 -> 1.10.8, www-client/brave-browser-1.95.101 -> 1.95.102)
commit dabdc8c16e28ffdcdf15f2e68100f15fadfd55da
Author: lucascouts <lucascs@protonmail.com>
Date: Wed Sep 16 11:54:03 2026 -0300
add(mail-client/{thunderbird, thunderbird-l10n}-156.0), add(metadata/{md5-cache/app-editors/zed-1.21.0_pre20260916, md5-cache/dev-db/sqlitebrowser-3.13.99_p20260916, md5-cache/dev-util/copilot-cli-bin-1.0.85, md5-cache/dev-util/crush-bin-0.95.0, md5-cache/dev-util/mesa_clc-26.3.0_pre20260916, md5-cache/dev-util/qwen-code-0.24.0, md5-cache/mail-client/thunderbird-156.0, md5-cache/mail-client/thunderbird-l10n-156.0, md5-cache/media-gfx/openvdb-13.1.0, md5-cache/media-libs/mesa-26.3.0_pre20260916, md5-cache/net-dns/bind-tools-9.20.29, md5-cache/net-misc/postman-bin-12.28.2, md5-cache/sci-ml/amd-gaia-0.24.1, md5-cache/sci-ml/ik_llama-cpp-0_pre20260916, md5-cache/sys-kernel/linux-firmware-20260916, md5-cache/www-client/chromium-153.0.8010.47}), mod(.autoupdate/packages.toml), up(app-editors/zed-1.21.0_pre20260915-r1 -> 1.21.0_pre20260916, dev-db/sqlitebrowser-3.13.99_p20260913 -> 3.13.99_p20260916, dev-util/{copilot-cli-bin-1.0.83 -> 1.0.85, crush-bin-0.94.2 -> 0.95.0, mesa_clc-26.3.0_pre20260915 -> 26.3.0_pre20260916, qwen-code-0.23.4 -> 0.24.0}, media-gfx/openvdb-13.0.0 -> 13.1.0, media-libs/mesa-26.3.0_pre20260915 -> 26.3.0_pre20260916, net-dns/bind-tools-9.20.27 -> 9.20.29, net-misc/postman-bin-12.28.1 -> 12.28.2, sci-ml/{amd-gaia-0.23.0 -> 0.24.1, ik_llama-cpp-0_pre20260915 -> 0_pre20260916}, sys-kernel/linux-firmware-20260915 -> 20260916, www-client/chromium-153.0.8010.36 -> 153.0.8010.47)
commit 375bea2c592c51537dcfbfe86b737a93a475ec36
Author: lucascouts <lucascs@protonmail.com>
Date: Wed Sep 16 00:31:16 2026 -0300
app-editors/zed: revbump 1.21.0_pre20260915-r1, refresh patch series
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
commit 1728ef48f7752667ff89a23fcba82aaab5b2c46a
Author: lucascouts <lucascs@protonmail.com>
Date: Tue Sep 15 01:07:53 2026 -0300
app-editors/zed: bump to 1.21.0_pre20260915, and fix the inherited FTBFS
7f00507e0156 landed 2026-09-15 03:06 UTC; d62802d456ac was 2026-09-14
16:38 UTC, so this is a new date and not a revision. Upstream's own
version stays 1.21.0 and rust-toolchain.toml stays 1.98.1, which is
already the RUST_MIN_VER here.
pycargoebuild run from crates/zed left CRATES (1494 entries) and
GIT_CRATES (71) byte-identical against the ebuild -- the Cargo.lock is
byte-identical between the two commits, and running it anyway is what
proves the ebuild matches that lock rather than merely that the lock did
not move. The seven revisions hardcoded in the src_prepare sed still name
what [patch.crates-io] pins upstream, and clean prepare rewrites 11 of 11
lines to path=, leaving no git= behind.
The 1.21.0_pre20260914-r1 commit message recorded a known FTBFS and said
the fix would land through zed-patches. It has: 0017 now writes
`error: None` + `summary: Vec::new()` for the ContextCompaction shape
upstream moved to in 25b5569d, and clones the status, which stopped being
Copy when ContextCompactionStatus gained Other(Arc<str>). With that,
cargo check --offline passes on acp_thread, agent, agent_ui,
agent_servers and claude_code_ide -- the first green on this series since
_pre20260912-r1.
0026 joins PATCHES under claude-agent-acp-plus: the two end-of-turn
navigation controls float over the conversation instead of scrolling away
with the turn that owned them. It was authored after the last bump and
had never been wired in.
files/ is resynced from zed-patches, all three relations in sync. That
also drops 0024 from 13069 lines to 308: the published copy had swept
acp_thread.rs.orig and .rej into itself from a conflicted run, so it was
creating two junk files in $ on every merge.
pkgcheck reports only the pre-existing DeprecatedEclass notice for
llvm-r1. gentoo-parity exits 0 with all 9 rows JUSTIFIED.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
commit 45dc22b63133d26abf506f398c93437a6521b1d9
Author: lucascouts <lucascs@protonmail.com>
Date: Mon Sep 14 15:56:09 2026 -0300
add(metadata/md5-cache/app-misc/zeo-systray-0.1.2), del(metadata/md5-cache/app-misc/zeo-systray-0.1.1), up(app-misc/zeo-systray-0.1.1 -> 0.1.2)
commit 59d99dd229680f98ea2c9f6f2578a96a7f4cba9b
Author: lucascouts <lucascs@protonmail.com>
Date: Mon Sep 14 15:54:37 2026 -0300
app-editors/zed: bump to 1.21.0_pre20260914-r1
Same-day snapshot, so this is a revision and not a new date: d62802d456ac
landed 2026-09-14 16:38 UTC, after the 25b5569dd231 the applier packaged
that morning.
pycargoebuild run from crates/zed left CRATES (1494), GIT_CRATES and the
crate LICENSE+= block byte-identical -- the Cargo.lock did not move at all
since _pre20260912-r1. The seven revisions hardcoded in the src_prepare sed
still match [patch.crates-io] upstream, and ebuild clean prepare rewrites
11 of 11 lines to path=, leaving no git= behind.
files/ is resynced from zed-patches: 17 of 17 apply to d62802d456ac with
zero fuzz, nine drifted in offsets only. The two stale md5-cache entries
(_pre20260912-r1 and _pre20260914) went with the egencache pass.
KNOWN FTBFS, inherited, not introduced: cargo check --offline -p acp_thread
fails on files/0017 -- upstream changed ContextCompaction between 7960b2a7
and 25b5569d (summary: Option<Entity<Markdown>> became error: Option<...>
plus summary: Vec<ContentBlock>), and the patch still builds `summary:
None` in two places. The published _pre20260914 breaks identically; the
fix lands through zed-patches, not here.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
commit 47bd6a6307c631219999d2276cc6af970d0d714c
Author: lucascouts <lucascs@protonmail.com>
Date: Mon Sep 14 11:29:36 2026 -0300
add(metadata/{md5-cache/app-editors/zed-1.21.0_pre20260914, md5-cache/dev-util/crush-bin-0.94.2, md5-cache/media-libs/vulkan-layers-1.4.362_p20260914, md5-cache/net-misc/postman-bin-12.28.0, md5-cache/sci-ml/ik_llama-cpp-0_pre20260914}), mod(.autoupdate/packages.toml), up(app-editors/zed-1.21.0_pre20260912-r1 -> 1.21.0_pre20260914, dev-util/crush-bin-0.94.1 -> 0.94.2, media-libs/vulkan-layers-1.4.362_p20260913 -> 1.4.362_p20260914, net-misc/postman-bin-12.27.7 -> 12.28.0, sci-ml/ik_llama-cpp-0_pre20260910 -> 0_pre20260914)
commit fa55f9c1deb15341729f1bfec59738b5ea87f351
Author: lucascouts <lucascs@protonmail.com>
Date: Sun Sep 13 23:02:49 2026 -0300
app-editors/zed: bump to 1.21.0_pre20260912-r1
Same-day snapshot, so this is a revision and not a new date: 7960b2a7c956
landed 2026-09-12 20:38 UTC, hours after the 9d272b0363 this replaces.
pycargoebuild run from crates/zed left CRATES and GIT_CRATES byte-identical.
The Cargo.lock did move, but only in internal edges -- the zed crate gained
buffer_diff and git and dropped itertools -- so no registry or git package
entered or left. The eight revisions hardcoded in the src_prepare sed still
match [patch.crates-io] upstream, and replaying that block over the new
Cargo.toml rewrites 11 of 11 lines to path=, leaving no git= behind.
Evidence, not inference: ebuild clean prepare applies 17 of 17 patches, and
cargo metadata --offline inside $ exits 0, so the whole graph resolves
without the network. It was not compiled -- an FTBFS in code would still be
invisible here.
files/0002 and files/0024 are resynced from zed-patches, which is the source
of truth for both; each drifted in offsets only. The stale md5-cache entry
for 1.21.0_pre20260911-r2 went with the egencache pass.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
commit 4faa2152b55e6e237aed0016b5c5a7a55114b312
Author: lucascouts <lucascs@protonmail.com>
Date: Sun Sep 13 21:24:10 2026 -0300
add(metadata/md5-cache/app-misc/zeo-systray-0.1.1), del(metadata/md5-cache/app-misc/zeo-systray-0.1.0), mod(app-editors/zed-1.21.0_pre20260912), mod(metadata/{md5-cache/app-editors/zed-1.21.0_pre20260912, md5-cache/app-editors/zeo-0.1.0_p20260912}), up(app-misc/zeo-systray-0.1.0 -> 0.1.1)
commit 83eab793b8ba89f6d655b200a5e9ecc896b245cd
Author: lucascouts <lucascs@protonmail.com>
Date: Sat Sep 12 13:36:49 2026 -0300
add(metadata/{md5-cache/app-editors/cursor-3.20.17, md5-cache/app-editors/zed-1.21.0_pre20260912, md5-cache/dev-util/ai-memory-2.2.0, md5-cache/dev-util/ai-memory-bin-2.2.0, md5-cache/dev-util/crush-bin-0.94.1, md5-cache/media-libs/vulkan-layers-1.4.362_p20260912, md5-cache/sys-kernel/linux-firmware-20260912}), mod(.autoupdate/packages.toml), up(app-editors/{cursor-3.20.14 -> 3.20.17, zed-1.21.0_pre20260911-r2 -> 1.21.0_pre20260912}, dev-util/ai-memory{,-bin}-2.1.2 -> 2.2.0, media-libs/vulkan-layers-1.4.362_p20260911 -> 1.4.362_p20260912, sys-kernel/linux-firmware-20260911 -> 20260912)
commit d9889079af47a8959fcb075c66ab5396a81f0d33
Author: lucascouts <lucascs@protonmail.com>
Date: Fri Sep 11 11:02:54 2026 -0300
up(app-editors/zed-1.21.0_pre20260911-r1 -> 1.21.0_pre20260911-r2), mod(.autoupdate/packages.toml)
Same-day snapshot, so the bump is a revision: a9cdfc9936c4 lands on 2026-09-11
like d9e1c024f393 did.
This -r2 was first written against 3cef31688ae2 (11:10 UTC) and never published;
the packaged commit is now a9cdfc9936c4 (21:33 UTC), nine commits later.
Retargeted in place rather than published and followed by an -r3: a revision for
an intermediate commit nobody would install buys nothing, and SRC_URI names the
tarball after $, so reusing -r2 reuses the DIST slot too. That reuse is also
the Manifest trap -- the stale DIST line sat under exactly the right filename, as
did the stale distfile in DISTDIR, and both had to be replaced by hand before
`pkgdev manifest`, which otherwise has no reason to look at either.
Three lines of the ebuild move. CRATES gains ashpd 0.13.5 for 0.13.2 -- the whole
of what pycargoebuild, re-run from crates/zed, says changed: GIT_CRATES (71) and
the LICENSE+= block come out byte-identical, and unindent stays in CRATES even
though it left crates/migrator's dev-dependencies, other consumers holding it.
All seven `local *_COMMIT` were re-checked against the regenerated GIT_CRATES:
all seven resolve, none moved.
RUST_MIN_VER 1.97.1 -> 1.98.1 is not required by the new commit; it was already
wrong. rust-toolchain.toml had moved to 1.98.1 at 3cef3168, the very commit the
unpublished -r2 packaged. That axis is invisible to every local check -- the
applier never reads the file, `ebuild ... prepare` never consults the floor, and
cargo check passes precisely because this host's active toolchain is already what
the ebuild forgot to ask for. Only someone installing with an older rust finds it.
Six patches changed, offsets only (conversation_view.rs +7, crates/zed/src/main.rs
+3), refreshed through lucascouts/zed-patches, where the series is the source of
truth; files/ here is the generated copy and check-sync reports all three
relations in sync.
Verified: `ebuild ... prepare` applies all sixteen for real (>>> Source prepared);
11/11 sed patterns match the new Cargo.toml exactly once, leaving 0
[patch.crates-io] entries on `git =` and all ten WORKDIR targets present;
`cargo check --locked --offline -p zed -p remote_server --features zed/inspector`
exits 0 with zero errors, --locked on purpose because 0002 adds a Cargo.lock entry
whose pin is silent without it; md5-cache _md5_ matches the ebuild and egencache
touched only that one entry; pkgcheck reports only the pre-existing
deprecated-llvm-r1 finding. Not merged with a build: this host has no sudo, so no
emerge ran.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
commit 9d95d6c73efb6c55c9a27f2bc53ae20458224cf1
Author: lucascouts <lucascs@protonmail.com>
Date: Fri Sep 11 04:43:57 2026 -0300
up(app-editors/zed-1.21.0_pre20260911 -> 1.21.0_pre20260911-r1), mod(.autoupdate/packages.toml)
2518540 (01:07 UTC) -> d9e1c024 (05:25 UTC), the same UTC day, so the date
component alone does not move: -r1. An in-place swap was not available either,
because _pre20260911 was already pushed.
Cargo.lock and the root Cargo.toml are byte-identical between the two commits,
so the eleven src_prepare seds and the seven `local *_COMMIT` needed nothing.
All eighteen were re-checked rather than assumed.
What re-running pycargoebuild uncovered was in the two ebuilds ALREADY
PUBLISHED, neither of which would have built offline:
- _pre20260910 (86b2cf96) added objc2-core-graphics, objc2-io-surface and
objc2-screen-capture-kit to Cargo.lock; all three were missing from CRATES,
so they never reached SRC_URI.
- _pre20260911 (2518540) moved pet from microsoft/python-environment-tools
@bb8e0460 to zed-industries/python-environment-tools @5aad7164, and all 27
pet-* entries in GIT_CRATES kept naming the old repo and the dead rev.
Both are the applier defect recorded on 2026-08-25, recurring. An identical
Cargo.lock proves the SOURCE did not move between two commits; it never proves
the EBUILD ever matched it.
The patch series was 17 commits behind, not 2 -- files/ was generated against
52b2927a and the two applier bumps since never touched it. Refreshed through
zed-patches: all 16 apply, and only 0002 and 0012 changed, both purely
re-anchored (blob index hashes, Cargo.lock @@ -22769 -> -22807 from the objc2
crates, main.rs @@ -696 -> -698 from the two lines d9e1c024 inserted above the
hunk). No semantic rebase.
Verified: check-sync agrees on all three relations; `ebuild ... prepare` applies
all sixteen for real; 11/11 seds match exactly once, leaving 0 [patch.crates-io]
entries on `git =` with all ten WORKDIR targets present; `cargo check --offline
-p zed -p remote_server --features zed/inspector` exits 0 in 1m33 with zero
errors; md5-cache _md5_ matches; pkgcheck reports only the pre-existing
deprecated-llvm-r1 finding. Three orphan md5-cache entries swept.
Not merged with a build: this host has no sudo, so no emerge ran.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
commit 53b15b2081566a545a5066fe2801dbfa87355882
Author: lucascouts <lucascs@protonmail.com>
Date: Thu Sep 10 23:44:40 2026 -0300
add(metadata/{md5-cache/app-admin/awscli2-2.36.43, md5-cache/app-editors/cursor-3.20.10, md5-cache/app-editors/vim-9.2.1067, md5-cache/app-editors/vim-core-9.2.1067, md5-cache/app-editors/zed-1.21.0_pre20260911, md5-cache/app-misc/claude-desktop-bin-1.52386.0, md5-cache/dev-games/godot-4.8_alpha5, md5-cache/dev-python/openai-3.13.0, md5-cache/dev-python/uv-bin-0.12.13, md5-cache/dev-util/claude-code-2.1.268, md5-cache/dev-util/glslang-1.4.357.0_p20260910, md5-cache/kde-frameworks/kwindowsystem-6.30.0, md5-cache/net-im/telegram-desktop-7.2.8}), mod(.autoupdate/packages.toml), up(app-admin/awscli2-2.36.42 -> 2.36.43, app-editors/{cursor-3.19.19 -> 3.20.10, vim-9.2.1064 -> 9.2.1067, vim-core-9.2.1064 -> 9.2.1067, zed-1.21.0_pre20260910 -> 1.21.0_pre20260911}, app-misc/claude-desktop-bin-1.49585.0 -> 1.52386.0, dev-games/godot-4.8_alpha4-r1 -> 4.8_alpha5, dev-python/{openai-3.12.0 -> 3.13.0, uv-bin-0.12.12 -> 0.12.13}, dev-util/{claude-code-2.1.267 -> 2.1.268, glslang-1.4.357.0_p20260909 -> 1.4.357.0_p20260910}, net-im/telegram-desktop-7.2.7 -> 7.2.8)
commit e59e7c11f946617395df9aa0f66485aa460bd9ce
Author: lucascouts <lucascs@protonmail.com>
Date: Thu Sep 10 09:59:55 2026 -0300
up(app-editors/zed-1.21.0_pre20260909 -> 1.21.0_pre20260910, app-emulation/bochs-3.0-r1 -> 3.1, dev-util/{codex-desktop-bin-26.903.61454 -> 26.903.71938, mesa_clc-26.3.0_pre20260909 -> 26.3.0_pre20260910}, media-libs/{mesa-26.3.0_pre20260909 -> 26.3.0_pre20260910, vulkan-layers-1.4.362_p20260909 -> 1.4.362_p20260910}, net-misc/postman-bin-12.27.2 -> 12.27.5, sci-libs/onnxruntime{,-bin}-1.29.0 -> 1.29.1, sci-ml/ollama-bin-0.33.3 -> 0.34.0, sys-kernel/linux-firmware-20260831 -> 20260910), mod(.autoupdate/packages.toml)
The applier re-enabled three records that were pinned OFF and resurrected
the packages behind them. All three live in ::gentoo, which is why they
were dropped here in the first place, and two came back unfetchable:
sci-ml/ollama-0.34.0 Manifest carried ollama-0.33.3-deps.tar.xz's
size and hashes under the 0.34.0 name, and
LLAMA_CPP_tag was still 0.33.3's b10760
app-admin/calamares-3.4.3 same rename-without-recompute, this time on
calamares-3.4.2b-gentoo-patches.patch
sys-power/upower-1.91.4 fetchable, but byte-identical to ::gentoo's
1.91.3 ebuild: a duplicate, not a divergence
Only the ollama record tripped check-autoupdate-damage.sh, because it is
the one carrying a bare `enabled = false`. The other two also wore
`disabled_by = "auto"`, which the check reads as "the package came back
on purpose" -- true when a human re-adds it, false here. Restored all
three records verbatim from HEAD and dropped the three package
directories again.
Also
Author: lucascouts <lucascs@protonmail.com>
Date: Thu Oct 1 14:22:26 2026 -0300
app-editors/zeo-bin: name the distfile after $; app-editors/zed: drop a stray probe script
zeo-bin mirrors zeo's full version, and a zeo revbump is a different binary,
so its tarball now carries the revision: SRC_URI and S use $.
zed-patches/scripts/make-bin-release.sh names it the same way and refuses
to reuse a name R2 already serves with other bytes. At -r0 $ is $,
so the published file and the Manifest are unchanged; verified by a full
fetch-verify-unpack-install of zeo-bin-0.1.0_p20261001.
app-editors/zed/.probe.sh was a curl probe left behind on 2026-08-06 for
the long-gone 1.16.0_pre20260806-r1 snapshot; nothing references it.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
commit 001f5e999f6707b9de031d4619482d19d4e1ebd1
Author: lucascouts <lucascs@protonmail.com>
Date: Thu Oct 1 11:07:58 2026 -0300
app-editors/{zed,zeo,zeo-bin}: zed builds releases, the patched snapshot is zeo
app-editors/zed stops carrying the bentoo patch series and becomes the
source twin of zed-bin: upstream's tagged releases, unpatched, as
zed-1.22.0 (stable channel) and zed-1.23.1_pre (preview). The three
claude-agent USE flags leave it.
app-editors/zeo takes over what zed was -- the snapshot at 71456c40 with
the whole series, 0001-0031 -- plus the rebrand (0019-0023, 0025) applied
before every USE group. The rebrand patches had never been refreshed
since 2026-09-12, because they were not in the series refresh.sh reads;
they are now carried onto 71456c40, and 0021 lost one hunk whose function
upstream removed. The distfile is zed-$.tar.gz and ZED_PV is
gone: the autoupdate (base_from = "none") can only move the date and
EGIT_COMMIT, and a stale second version would name new bytes after an
already-manifested distfile.
app-editors/zeo-bin is new: the same build at the same PV, compiled for
x86-64-v3 under zed-patches/release/configroot and hosted at
distfiles.obentoo.org; PROVENANCE.txt inside names the commit, each patch
with its sha256, the USE set and the flags.
Only one of zed, zed-bin, zeo, zeo-bin installs at a time; the blockers
live in zeo and zeo-bin. A news item explains the switch and the
expected downgrade of an installed zed snapshot to a release.
Verified: check-sync (29 patches: series, files/ and source agree);
ebuild prepare for all four ebuilds; cargo check -p zed for zeo and both
releases; a full x86-64-v3 compile+install of zeo (no AVX-512 in the
binary); zeo-bin install from the R2 copy, whose sha256 matches the local
tarball; pkgcheck shows only the pre-existing llvm-r1 deprecation and
zed's stable/preview RedundantVersion; autoupdate --lint passes.
Not merged with emerge: no root here.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
commit 28a06bc71e3862e7ae61a786a271aa67f918e588
Author: lucascouts <lucascs@protonmail.com>
Date: Thu Oct 1 08:59:05 2026 -0300
app-editors/zed: 1.24.0_pre20261001-r1 at 71456c40
Upstream 71456c40 landed the same UTC day as 74c134a3 (06:12 -> 10:15),
so the date component did not change; bumped as -r1. SRC_URI already
renames to $, so the distfile name stays unique.
The commit touches only crates/gpui/src/keymap.rs and
platform/keystroke.rs. pycargoebuild from crates/zed left CRATES (1494),
GIT_CRATES (70) and LICENSE unchanged, and the src_prepare sed turns all
11 [patch.crates-io] entries into path= deps.
The patch series is carried over from 1.23.0_pre20260930, which the
two applier bumps to 1.24.0 skipped: 9 patches moved in offsets only
(+/- lines byte-identical), 14 are identical. Drops 4 md5-cache entries
left behind by earlier bumps.
Verified: ebuild prepare applies 23/23 with no fuzz; cargo check of the
15 patched crates is clean; the 108 tests the patches add pass
(108/108), with RUSTFLAGS=--cap-lints=warn (proptest) and an empty .git
in $ (debug asset lookup). Not built with emerge.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
commit 0bb39ebe93918341f348efc3651f40ada5afde66
Author: lucascouts <lucascs@protonmail.com>
Date: Thu Oct 1 08:05:57 2026 -0300
add(metadata/{md5-cache/app-editors/zed-1.24.0_pre20261001, md5-cache/app-misc/zeo-systray-0.2.1, md5-cache/dev-python/conda-26.9.0, md5-cache/dev-util/ai-memory-2.5.2, md5-cache/dev-util/ai-memory-bin-2.5.2, md5-cache/dev-util/mesa_clc-26.3.0_pre20261001, md5-cache/media-libs/mesa-26.3.0_pre20261001, md5-cache/media-libs/vulkan-layers-1.4.363_p20261001, md5-cache/net-misc/postman-bin-12.30.5, md5-cache/net-p2p/bisq-bin-1.10.9, md5-cache/sci-ml/ik_llama-cpp-0_pre20261001, md5-cache/sci-ml/llama-swap-261, md5-cache/sys-apps/ai-jail-2.5.0, md5-cache/sys-apps/ai-jail-bin-2.5.0}), mod(.autoupdate/packages.toml), up(app-editors/zed-1.24.0_pre20260930 -> 1.24.0_pre20261001, dev-python/conda-26.7.3 -> 26.9.0, dev-util/ai-memory{,-bin}-2.5.0 -> 2.5.2, media-libs/{mesa-26.3.0_pre20260930 -> 26.3.0_pre20261001, vulkan-layers-1.4.363_p20260930 -> 1.4.363_p20261001}, net-misc/postman-bin-12.30.4 -> 12.30.5, net-p2p/bisq-bin-1.10.8 -> 1.10.9, sci-ml/{ik_llama-cpp-0_pre20260930 -> 0_pre20261001, llama-swap-260 -> 261}, sys-apps/ai-jail{,-bin}-2.4.0 -> 2.5.0)
commit c1787b478187140e4b4439f3c363d13e3c35c30f
Author: lucascouts <lucascs@protonmail.com>
Date: Wed Sep 30 23:15:13 2026 -0300
add(metadata/{md5-cache/app-admin/awscli2-2.37.7, md5-cache/app-editors/kiro-1.2.4, md5-cache/app-editors/zed-1.24.0_pre20260930, md5-cache/app-editors/zed-bin-1.22.0, md5-cache/app-editors/zed-bin-1.23.1_pre, md5-cache/dev-util/ai-memory-2.5.0, md5-cache/dev-util/ai-memory-bin-2.5.0, md5-cache/dev-util/antigravity-hub-bin-2.19.1.6046815158665216, md5-cache/dev-util/claude-code-2.1.286, md5-cache/dev-util/cline-bin-3.0.67, md5-cache/dev-util/codex-0.159.3, md5-cache/dev-util/codex-bin-0.159.3, md5-cache/dev-util/codex-desktop-bin-26.928.31416, md5-cache/dev-util/copilot-cli-bin-1.0.90, md5-cache/dev-util/glslang-1.4.363.0_p20260930, md5-cache/dev-util/openchamber-bin-2.1.0, md5-cache/dev-util/opencode-bin-1.18.34, md5-cache/media-libs/harfbuzz-14.5.1, md5-cache/net-analyzer/netdata-2.12.0, md5-cache/sci-ml/lemonade-2026.40.0, md5-cache/sci-ml/lemonade-bin-2026.40.0, md5-cache/sys-apps/ai-jail-2.4.0, md5-cache/sys-apps/ai-jail-bin-2.4.0, md5-cache/www-client/brave-browser-1.96.60}), mod(.autoupdate/packages.toml), up(app-admin/awscli2-2.37.6 -> 2.37.7, app-editors/{kiro-1.1.70 -> 1.2.4, zed-1.23.0_pre20260930 -> 1.24.0_pre20260930, zed-bin-1.21.0 -> 1.22.0, zed-bin-1.22.0_pre -> 1.23.1_pre}, dev-util/{ai-memory{,-bin}-2.4.2 -> 2.5.0, codex{,-bin}-0.159.2 -> 0.159.3}, media-libs/harfbuzz-14.5.0 -> 14.5.1, net-analyzer/netdata-2.11.1 -> 2.12.0, sci-ml/lemonade{,-bin}-2026.39.1 -> 2026.40.0, sys-apps/ai-jail{,-bin}-2.2.1 -> 2.4.0, www-client/brave-browser-1.96.59 -> 1.96.60)
commit 2a69424bdef1b6a0b924305aeed87ef000fa2f64
Author: lucascouts <lucascs@protonmail.com>
Date: Wed Sep 30 16:41:33 2026 -0300
app-editors/zed: carry the patch series onto 1.23.0_pre20260930
The applier bumped to e8c6c67f without touching files/, and two patches
no longer applied, so src_prepare died on eapply:
- 0011: upstream grew the `use acp_thread::` list in thread_view.rs;
AccountUsageStatus is re-inserted into it.
- 0031: upstream appended a field to TurnFields; session_duration now
follows it.
Only their position moved. 14 others changed in offsets and index lines
only; 7 are identical. No revbump: the published ebuild never got past
src_prepare, so nothing broken was installed.
pycargoebuild from crates/zed left CRATES, GIT_CRATES and LICENSE
unchanged, and all seven src_prepare sed commits match Cargo.toml.
Verified: ebuild prepare applies 23/23; cargo check of the 15 patched
crates is clean; the 108 tests the patches add pass (108/108). Tests
ran with RUSTFLAGS=--cap-lints=warn, because proptest denies warnings
and rustc 1.98 now warns in it, and with an empty .git in $, because
debug builds locate their assets through it. Not built with emerge.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
commit 5d924b77b594a2823b9506afb1c25c5c8374ed6c
Author: lucascouts <lucascs@protonmail.com>
Date: Wed Sep 30 14:58:31 2026 -0300
add(metadata/{md5-cache/app-editors/vim-9.2.1152, md5-cache/app-editors/zed-1.23.0_pre20260930, md5-cache/dev-db/sqlitebrowser-3.13.99_p20260930, md5-cache/dev-util/cline-bin-3.0.66, md5-cache/dev-util/codex-desktop-bin-26.928.21956, md5-cache/dev-util/spirv-tools-1.4.363.0_p20260930, md5-cache/dev-util/vulkan-tools-1.4.364_p20260930, md5-cache/media-libs/vulkan-layers-1.4.363_p20260930, md5-cache/media-libs/vulkan-loader-1.4.364_p20260930, md5-cache/media-sound/audacity-bin-4.0.1, md5-cache/net-misc/postman-bin-12.30.4, md5-cache/net-vpn/obscura-1.181_p20260930, md5-cache/net-vpn/obscura-bin-1.181, md5-cache/sci-ml/ik_llama-cpp-0_pre20260930, md5-cache/www-client/librewolf-bin-157.0_p1, md5-cache/www-client/opera-136.0.6008.80}), mod(.autoupdate/packages.toml), up(app-editors/{vim-9.2.1125 -> 9.2.1152, zed-1.23.0_pre20260929-r1 -> 1.23.0_pre20260930}, dev-db/sqlitebrowser-3.13.99_p20260929 -> 3.13.99_p20260930, dev-util/{cline-bin-3.0.65 -> 3.0.66, codex-desktop-bin-26.928.20755 -> 26.928.21956, spirv-tools-1.4.363.0_p20260929 -> 1.4.363.0_p20260930, vulkan-tools-1.4.364_p20260925 -> 1.4.364_p20260930}, media-libs/{vulkan-layers-1.4.363_p20260929 -> 1.4.363_p20260930, vulkan-loader-1.4.364_p20260929 -> 1.4.364_p20260930}, media-sound/audacity-bin-4.0.0 -> 4.0.1, net-misc/postman-bin-12.30.3 -> 12.30.4, net-vpn/{obscura-1.180_p20260929 -> 1.181_p20260930, obscura-bin-1.180 -> 1.181}, sci-ml/ik_llama-cpp-0_pre20260929 -> 0_pre20260930, www-client/{librewolf-bin-156.0.1_p1 -> 157.0_p1, opera-136.0.6008.73 -> 136.0.6008.80})
commit 1b3095c9f1b8f55bfb16772d5803ba11d3c8e628
Author: lucascouts <lucascs@protonmail.com>
Date: Wed Sep 30 00:28:45 2026 -0300
app-editors/zed: bump to decbf641 as 1.23.0_pre20260929-r1
Same-day snapshot, six commits past 1dc88444. pycargoebuild from
crates/zed left CRATES and GIT_CRATES unchanged: 1494/1494 registry
crates and 32/32 git revs match Cargo.lock, and all eleven src_prepare
sed anchors still rewrite to path=.
The patch series is refreshed against decbf641 -- 3 changed (0002,
0008, 0016; offsets only), 20 identical. Also drops the md5-cache entry
of the replaced 0929 ebuild.
Verified: ebuild prepare applies 23/23; cargo metadata --offline exits
0; cargo check of the 15 patched crates is clean; cargo test runs 2865
passed, 3 failed. Two failures assert AcpBetaFeatureFlag is off, which
0001 forces on by design; the third, test_multi_workspace_session_restore,
is flaky upstream (1 in 6 runs without patches). Not built with emerge.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
commit ab641906129c746bb637153ac5f2c744f22de834
Author: lucascouts <lucascs@protonmail.com>
Date: Tue Sep 29 14:27:08 2026 -0300
app-editors/zed: 0024 opens a session in its own project, with its own agent
zed://agent?session=<id> now opens the thread in the window holding its
project and loads it with the agent on its metadata row, instead of the
focused window and the native agent ("no thread found with ID").
0012 and 0027 are regenerated for shifted hunk positions only.
Source: lucascouts/zed-patches db4d4d0.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
commit a43ed83d086645694f313c9eae3ae2eb72352e94
Author: lucascouts <lucascs@protonmail.com>
Date: Tue Sep 29 12:34:36 2026 -0300
app-editors/zed: bump to 1dc88444 as 1.23.0_pre20260929
CRATES regenerated with pycargoebuild from crates/zed: upstream dropped ten
registry crates in this range (portable-pty, nix 0.28, serial2,
x11-clipboard, winreg 0.10, crossbeam, cfg_aliases 0.1, json_dotpath,
shared_library, take-until). GIT_CRATES is unchanged, and all eleven
src_prepare sed anchors still match the [patch.crates-io] revs.
The patch series is refreshed against 1dc88444 -- 11 changed, 12
identical, offsets only plus one context line upstream renamed
(shell -> shell_program). Also drops the stale md5-cache entry of
0927-r1, which had no ebuild behind it.
Verified: CRATES 1494/1494 and GIT_CRATES 70/70 against Cargo.lock;
check-sync clean; ebuild prepare applies 23/23; cargo metadata --offline
exits 0 after prepare. Not compiled.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
commit e9ae0eb296f21bf9e85ccf0334263423b058e1a2
Author: lucascouts <lucascs@protonmail.com>
Date: Mon Sep 28 21:26:28 2026 -0300
app-editors/zed: carry the patch series onto 1.23.0_pre20260928
The autoupdate bumped the ebuild to dd510f99 with the 0927-r1 patches in
files/, and 0008 no longer applied there: upstream began moving the agent UI
onto ACP schema v2 in this range. The series is now refreshed against
dd510f99 -- 18 patches changed, 5 identical -- and 0026 gains a "Scroll to
Next User Message" button beside "previous".
check-sync: series <-> ebuild, patches <-> overlay and patches <-> source
all agree.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
commit e2891460957c2ef4175056790d10509a2611c36d
Author: lucascouts <lucascs@protonmail.com>
Date: Mon Sep 28 18:33:41 2026 -0300
add(metadata/{md5-cache/app-admin/awscli2-2.37.5, md5-cache/app-editors/cursor-3.22.12, md5-cache/app-editors/zed-1.23.0_pre20260928, md5-cache/app-misc/claude-desktop-bin-2.9939.4, md5-cache/dev-python/openai-3.20.0, md5-cache/dev-util/claude-code-2.1.284, md5-cache/dev-util/codex-0.158.0, md5-cache/dev-util/codex-bin-0.158.0, md5-cache/dev-util/codex-desktop-bin-26.924.50649, md5-cache/dev-util/copilot-cli-bin-1.0.89, md5-cache/dev-util/mesa_clc-26.3.0_pre20260928, md5-cache/dev-util/openchamber-bin-2.0.3, md5-cache/dev-util/opencode-bin-1.18.33, md5-cache/dev-util/typescript-language-server-6.0.1-r3, md5-cache/dev-util/vulkan-utility-libraries-1.4.364_p20260927, md5-cache/media-gfx/freecad-1.1.4, md5-cache/media-libs/libopenshot-1.0.1, md5-cache/media-libs/libopenshot-audio-1.0.1, md5-cache/media-libs/mesa-26.3.0_pre20260928, md5-cache/media-libs/vulkan-layers-1.4.363_p20260928, md5-cache/media-video/openshot-4.0.1, md5-cache/media-video/shotcut-26.9.27, md5-cache/net-misc/postman-bin-12.30.0, md5-cache/net-vpn/obscura-1.180_p20260928, md5-cache/sci-ml/ik_llama-cpp-0_pre20260928, md5-cache/sys-apps/flatpak-1.18.4, md5-cache/sys-apps/flatpak-1.19.2}), del(metadata/md5-cache/dev-util/typescript-language-server-6.0.1-r2), mod(.autoupdate/packages.toml), up(app-admin/awscli2-2.37.4 -> 2.37.5, app-editors/{cursor-3.22.7 -> 3.22.12, zed-1.23.0_pre20260927-r1 -> 1.23.0_pre20260928}, app-misc/claude-desktop-bin-2.7032.0 -> 2.9939.4, dev-python/openai-3.19.2 -> 3.20.0, dev-util/codex{,-bin}-0.157.1 -> 0.158.0, media-gfx/freecad-1.1.3-r1 -> 1.1.4, media-libs/{libopenshot-1.0.0 -> 1.0.1, libopenshot-audio-1.0.0 -> 1.0.1, mesa-26.3.0_pre20260927 -> 26.3.0_pre20260928, vulkan-layers-1.4.363_p20260927 -> 1.4.363_p20260928}, media-video/{openshot-4.0.0 -> 4.0.1, shotcut-26.8.1 -> 26.9.27}, net-misc/postman-bin-12.29.5 -> 12.30.0, net-vpn/obscura-1.180_p20260926 -> 1.180_p20260928, sci-ml/ik_llama-cpp-0_pre20260927 -> 0_pre20260928, sys-apps/{flatpak-1.18.3 -> 1.18.4, flatpak-1.19.0 -> 1.19.2})
commit 517e85258d16402e915bf00b3fd886d28ec836d8
Author: lucascouts <lucascs@protonmail.com>
Date: Sun Sep 27 21:07:30 2026 -0300
app-editors/zed: bump to 1a28cff4 as 1.23.0_pre20260927-r1
Same UTC day as the previous snapshot, hence the revision bump.
- EGIT_COMMIT e683fd7b -> 1a28cff4
- CRATES regenerated with pycargoebuild: add alloca, c-enum,
perf-event{2,-data,-open-sys2}; criterion 0.5 -> 0.8.2,
itertools 0.11 -> 0.13; drop wast, wiggle*, witx
- GIT_CRATES and the src_prepare [patch.crates-io] revs unchanged;
all 11 entries still rewrite to path=
- all 23 patches apply unchanged (zed-patches check-sync in sync)
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
commit 9a4ee5732bbb958e32fda5c3cb62f6852dc3989f
Author: lucascouts <lucascs@protonmail.com>
Date: Sun Sep 27 20:47:55 2026 -0300
mod(app-editors/zed-1.23.0_pre20260927)
commit 499a0becce08ec05a50c29b26dc28f6e0aed5609
Author: lucascouts <lucascs@protonmail.com>
Date: Sun Sep 27 16:48:35 2026 -0300
app-editors/zed, dev-util/ccusage-bin: bump; md5-cache cleanup
- app-editors/zed: 1.23.0_pre20260926-r1 -> 1.23.0_pre20260927
- dev-util/ccusage-bin: 20.0.24 -> 20.0.26
- dev-util/claude-agent-acp-plus: add the missing 0.21.0 md5-cache entry
- drop md5-cache entries left behind by these bumps
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
commit 787ec8988388f7d15f8d96b8a1dcbd05b656c3d5
Author: lucascouts <lucascs@protonmail.com>
Date: Sun Sep 27 14:59:17 2026 -0300
app-editors/zed: 0027 queues several question sets in one card
Several AskUserQuestion calls in one turn now share the pinned card, one
at a time, with "Set N of M" in its header, instead of stacking. Synced
from zed-patches 322f581.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
commit 9bd6b8a70003c06d69ae65f6bf9df6a569bf54eb
Author: lucascouts <lucascs@protonmail.com>
Date: Sun Sep 27 14:14:46 2026 -0300
app-editors/zed: 0030 shows a thread waiting for an answer to a question
Synced from zed-patches bc8ee0e.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
commit a0c3207f3ff2574d4fbd33b0127af8e3ea2d2e12
Author: lucascouts <lucascs@protonmail.com>
Date: Sun Sep 27 07:36:40 2026 -0300
app-editors/zed: 0029 shows a thread whose background work outlives its turn
Synced from zed-patches 2ee40d5. Inert until claude-agent-acp-plus 0.21.0,
which sends the count it reads.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
commit 6ab74ec1f7fe2d720e8d9a38a496c0c7c51f1965
Author: lucascouts <lucascs@protonmail.com>
Date: Sat Sep 26 23:22:57 2026 -0300
app-editors/zed: 0028 opens a folder dropped on the threads sidebar
Synced from zed-patches 2082628.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
commit c68b43c7959c0c12b364417840184da9be8f8ac5
Author: lucascouts <lucascs@protonmail.com>
Date: Sat Sep 26 20:22:39 2026 -0300
app-editors/zed: 0014 test fixture passes clippy
Synced from zed-patches 5fe82c5. Test-only; the built binary is unchanged.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
commit 3048899f8cf0e7186e9f8afac836175363c4b9e7
Author: lucascouts <lucascs@protonmail.com>
Date: Sat Sep 26 19:17:27 2026 -0300
app-editors/zed: bump to 1.23.0_pre20260926-r1 (bda9c0bd)
Same-day snapshot, hence -r1. CRATES and GIT_CRATES regenerated with
pycargoebuild and came out identical; the src_prepare git-dep rewrites
still hit 11/11 entries of [patch.crates-io].
Patch series refreshed onto bda9c0bd: 0016 and 0027 moved by offset
only, the other 17 are byte-identical. cargo check of the 12 crates the
patches touch is clean.
md5-cache: drops the stale _pre20260925-r1 entry left behind with no
ebuild.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
commit 7d567024078e593aebdd12e6283f1c730f3bba23
Author: lucascouts <lucascs@protonmail.com>
Date: Sat Sep 26 18:49:32 2026 -0300
app-editors/zed: 0026 floats copy and scroll-to-bottom too
Synced from zed-patches 4c0e94a.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
commit 0cf8e84ecd8815e5a3025bcbc02ecf2836b213df
Author: lucascouts <lucascs@protonmail.com>
Date: Sat Sep 26 17:12:47 2026 -0300
app-editors/zed: 0002 advertises close_tab
0002 now lists close_tab in tools/list. It was dispatched and never
advertised (parity report item 32): the handler existed, but the CLI only
calls tools the server declares. The descriptor follows the VS Code
extension's registration -- a required `tab_name` string -- and the list
moved into a free tool_descriptors() so a plain test can hold it against
the dispatched names: every_dispatched_tool_is_advertised is red without
the descriptor ("`close_tab` is dispatched but not advertised") and green
with it.
Verified against 70e686c2a377 with the whole series applied, the tree
byte-identical to what the series produces: claude_code_ide 17/17 tests;
cargo check --all-targets -p claude_code_ide -p zed, no errors. Two
clippy lints under -D warnings (result_large_err, manual_inspect) and
rustfmt drift in three of the crate's files predate this change and are
left for a separate commit.
No revbump: the fix rides the next zed version bump rather than forcing
a rebuild of the editor for one tool descriptor.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
commit ba3b5d0571dab0da13f9749be88bef5d373f7a6e
Author: lucascouts <lucascs@protonmail.com>
Date: Sat Sep 26 14:46:25 2026 -0300
app-editors/zed: 0014 tests follow the MessageContent rework
Upstream #64667 (a405fb91d5) made UserMessage.content a MessageContent
that owns the original SDK blocks and dropped the `chunks` field. 0014's
two test fixtures still built the old shape, so `cargo check -p agent_ui
--all-targets` failed with four errors (E0599, E0560) -- in test code
only: verify.sh applies patches, it does not compile them, and the
package build never compiles #[cfg(test)].
The fixtures now use MessageContent::new / MessageContent::default(), as
upstream's own acp_thread test does. No revbump: nothing installed
changes.
Verified against 70e686c2a377 with the whole series applied: cargo check
--all-targets over the 28 crates the series touches, 0 errors, 0 warnings;
0014's four tests pass; the series file reproduces that tree exactly.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
commit 8def768d6922b143d0462fa3a25c8f54fc0bb4a8
Author: lucascouts <lucascs@protonmail.com>
Date: Sat Sep 26 13:10:27 2026 -0300
add(metadata/{md5-cache/app-editors/zed-1.23.0_pre20260926, md5-cache/dev-util/claude-agent-acp-plus-0.20.2, md5-cache/dev-util/claude-agent-acp-tui-0.14.2, md5-cache/dev-util/codex-0.157.1, md5-cache/dev-util/codex-bin-0.157.1, md5-cache/dev-util/codex-desktop-bin-26.924.22138, md5-cache/dev-util/mesa_clc-26.3.0_pre20260926, md5-cache/dev-util/openchamber-bin-2.0.2, md5-cache/dev-util/qwen-code-0.24.6, md5-cache/media-libs/mesa-26.3.0_pre20260926, md5-cache/media-libs/vulkan-layers-1.4.363_p20260926, md5-cache/net-vpn/obscura-1.180_p20260926, md5-cache/sci-ml/ik_llama-cpp-0_pre20260926, md5-cache/sci-ml/koboldcpp-1.122, md5-cache/sci-ml/llama-swap-260}), mod(.autoupdate/packages.toml), up(app-editors/zed-1.23.0_pre20260925-r1 -> 1.23.0_pre20260926, dev-util/codex{,-bin}-0.157.0 -> 0.157.1, media-libs/{mesa-26.3.0_pre20260925 -> 26.3.0_pre20260926, vulkan-layers-1.4.363_p20260925 -> 1.4.363_p20260926}, net-vpn/obscura-1.180_p20260924 -> 1.180_p20260926, sci-ml/{ik_llama-cpp-0_pre20260925 -> 0_pre20260926, koboldcpp-1.121 -> 1.122, llama-swap-258 -> 260})
commit c1194e63c7aaa0871edf1babcab1269753436ae1
Author: lucascouts <lucascs@protonmail.com>
Date: Fri Sep 25 21:45:22 2026 -0300
app-editors/zed: bump to 1.23.0_pre20260925-r1 (933d8d93)
Same-day snapshot, hence -r1. CRATES and GIT_CRATES regenerated with
pycargoebuild and came out identical; the src_prepare git-dep rewrites
still hit 11/11 entries of [patch.crates-io].
Patch series refreshed onto 933d8d93: 0010 needed a manual rebase of one
hunk after upstream turned LanguageModel's provider_id()/id() accessors
into fields (#64794); same shape as before (+516/-157, 52 hunks). The rest
moved by offset only. cargo check of the 12 crates the patches touch is
clean.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
commit ba64ef5c3603ce4bd3de5a7a94d6847b22debf02
Author: lucascouts <lucascs@protonmail.com>
Date: Fri Sep 25 07:27:59 2026 -0300
bump 10 packages from the 2026-09-25 autoupdate sweep
- app-admin/awscli2: 2.37.2 -> 2.37.3
- app-editors/zed: 1.23.0_pre20260924 -> 1.23.0_pre20260925 (a84858ac);
patch series refreshed, yawc moved from GIT_CRATES to CRATES
- app-editors/cursor: 3.21.18 -> 3.22.7
- dev-util/codex{,-bin}: 0.156.1 -> 0.157.0 (crossterm GIT_CRATES rev bumped)
- dev-util/vulkan-utility-libraries: 1.4.363_p20260925 -> 1.4.364_p20260925
(v1.4.364 tags the commit already carried; keeps the _p scheme of its siblings)
- media-libs/vulkan-layers: 1.4.363_p20260924 -> 1.4.363_p20260925 (fd8605f9);
glslang, vulkan-headers and vulkan-utility-libraries floors raised per known_good.json
- net-misc/postman-bin: 12.29.2 -> 12.29.4
- sci-ml/ik_llama-cpp: 0_pre20260924 -> 0_pre20260925 (6475e58b)
- sci-ml/llama-swap: 257 -> 258
media-libs/tlottie is left at 0.1.0_pre20260911: tdesktop 7.2.9 still pins
31f1b542, so the proposed 1.0.6_pre20260917 is main-branch drift the hold blocks.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
commit d1e3a81e58f5014fcb4b1ce52b279dd76a1a9ee7
Author: lucascouts <lucascs@protonmail.com>
Date: Fri Sep 25 07:05:29 2026 -0300
app-editors/zed: refresh the patch series onto 1.23.0_pre20260924
The autoupdate bump to 2c4bc2d7 left two patches conflicting, so the
package died in src_prepare:
- 0009: upstream now assigns session.config_options before applying the
defaults on session load. Build ConfigOptions once right after
config_state() and hand the session a clone; the clone shares the watch
sender through Rc, so default-applied options still wake the UI.
- 0011: upstream dropped PlanEntry from the acp_thread import list in
thread_view.rs; AccountUsageStatus is re-added to the new list.
Both keep their original shape (+59/-26 and +464/-3). The other ten
changed files only moved offsets. Generated from zed-patches
patches/zed-1.23.0_pre20260924.
verify.sh: 19/19 apply cumulatively. check-sync: in sync.
ebuild clean prepare: all 19 applied. cargo check --offline on the 12
crates the patches touch: clean. cargo test -p agent_servers not run:
the git proptest dev-dependency fails to build with rustc 1.98, which
RUST_MIN_VER requires -- unrelated to the patches.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
commit 3d54e2b69dec2a49e62250157d6c88f92d6a2b68
Author: lucascouts <lucascs@protonmail.com>
Date: Thu Sep 24 14:24:16 2026 -0300
bump: doctl, devin-desktop-bin, zed, sentry-native, cline-bin, trufflehog, vulkan-layers, ik_llama-cpp, ollama, ollama-bin, opera; add scripts/go-vendor.sh
dev-vcs/trufflehog 3.97.9 changed go.mod (go-osc52, mimetype), so the
reused 3.97.5 vendor tree no longer matched and -mod=vendor would fail.
Regenerate the vendor tarball for 3.97.9 (uploaded to distfiles.obentoo.org)
and point VENDOR_P at it.
scripts/go-vendor.sh checks whether a Go package's hosted vendor tarball
still matches its go.mod (--check exits 1 when stale) and regenerates it
otherwise. The autoupdate applier cannot see this; run it after bumping any
vendor-tarball Go package. Records and CLAUDE.md point to it; the reviewdog
record no longer claims a vendor tarball it does not use.
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
commit 411c9c7a0d28e074dca09d1a95c24ea1d5d0b112
Author: lucascouts <lucascs@protonmail.com>
Date: Wed Sep 23 16:21:47 2026 -0300
add(metadata/{md5-cache/app-editors/zed-1.23.0_pre20260923, md5-cache/app-editors/zed-bin-1.21.0, md5-cache/app-editors/zed-bin-1.22.0_pre, md5-cache/dev-util/claude-agent-acp-plus-0.20.1, md5-cache/dev-util/claude-agent-acp-tui-0.14.0, md5-cache/dev-util/goose-1.52.0, md5-cache/dev-util/goose-bin-1.52.0, md5-cache/dev-util/spirv-headers-1.4.357.0_p20260923, md5-cache/dev-vcs/trufflehog-3.97.8, md5-cache/sci-ml/lemonade-2026.39.1, md5-cache/sci-ml/lemonade-bin-2026.39.1, md5-cache/sys-apps/ai-jail-2.1.0, md5-cache/sys-apps/ai-jail-bin-2.1.0, md5-cache/sys-apps/pnpm-12.6.0}), mod(.autoupdate/packages.toml), up(app-editors/{zed-1.22.0_pre20260923-r1 -> 1.23.0_pre20260923, zed-bin-1.20.2 -> 1.21.0, zed-bin-1.21.0_pre -> 1.22.0_pre}, dev-util/goose{,-bin}-1.51.0 -> 1.52.0, dev-vcs/trufflehog-3.97.6 -> 3.97.8, sci-ml/{lemonade-11.9.0-r1 -> 2026.39.1, lemonade-bin-11.9.0 -> 2026.39.1}, sys-apps/ai-jail{,-bin}-2.0.0 -> 2.1.0)
commit f410b7b63202759b2df79a6ed037d324840a1740
Author: lucascouts <lucascs@protonmail.com>
Date: Wed Sep 23 12:00:36 2026 -0300
app-editors/zed: bump to 1.22.0_pre20260923-r1 (bf89346c)
Same UTC day as 7fecbb2c, so the date component alone does not move and this is
a revision. crates/zed/Cargo.toml still reads 1.22.0, rust-toolchain.toml still
pins 1.98.1 and corgi.toml still pins webrtc 0001d84-4, so PV base, RUST_MIN_VER
and WEBRTC_COMMIT all stand. pycargoebuild re-run from crates/zed reproduced
CRATES and GIT_CRATES byte for byte -- both WARNINGs printed, which is what
tells a real run apart from the silent no-op in the virtual workspace root. All
eleven local *_COMMIT sed anchors still match the new root Cargo.toml exactly
once.
The cost was in the patch series. 0011's agent_diff.rs hunk conflicted because
upstream inserted AcpThreadEvent::NoticesUpdated into the very arm list the hunk
adds AccountUsageUpdated to, so the insertion point lost its trailing context.
The change itself is one line and untouched; only the arm above it moved. Fixed
through a temporary source series rather than by hand-editing @@ counts, so all
nineteen were refreshed from one corrected baseline: 11 changed, 8 identical.
Verified: verify.sh, sync-overlay and check-sync agree on all three relations;
ebuild ... prepare applies all nineteen (>>> Source prepared); the 11 seds leave
zero [patch.crates-io] entries on 'git ='; cargo check --offline -p zed
-p remote_server --features gpui/inspector exits 0 in 2m25 with zero errors;
md5-cache _md5_ matches the ebuild and the two orphan entries were swept;
pkgcheck reports only the pre-existing deprecated-llvm-r1 finding. Not merged
with a build: this host has no sudo.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
commit 687194884611661c9a02a3ad7d0f745a332cb7f4
Author: lucascouts <lucascs@protonmail.com>
Date: Wed Sep 23 05:28:34 2026 -0300
autoupdate: apply the 2026-09-23 batch
app-admin/doctl 1.171.0 -> 1.171.2, app-editors/zed 1.22.0_pre20260922-r1
-> 1.22.0_pre20260923, dev-python/openai 3.18.0 -> 3.19.0,
dev-python/uv-bin 0.12.17 -> 0.12.18, dev-util/codex-bin 0.156.0 ->
0.156.1, dev-util/codex-desktop-bin 26.917.51856 -> 26.917.61114,
dev-util/mesa_clc and media-libs/mesa 26.3.0_pre20260922 ->
26.3.0_pre20260923, net-misc/postman-bin 12.29.1 -> 12.29.2,
sci-ml/ik_llama-cpp 0_pre20260922 -> 0_pre20260923.
Each new md5-cache entry was checked against its ebuild: all ten _md5_
lines match md5sum, so no bump carries a stale cache row.
dev-util/codex is not here; its 0.156.1 went in on its own commit because
it needed a source fix to build.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
commit 145bf9d6ca6f70867ceed9f4685d515bb5fb49cf
Author: lucascouts <lucascs@protonmail.com>
Date: Tue Sep 22 21:29:34 2026 -0300
app-editors/zed: the 0017 comment names the crate that is actually linked
The PATCHES comment said the ACP compaction notification was unreachable here
because agent-client-protocol 2.0.0 carries zero occurrences of "compaction".
Measured against the packaged commit 96e95eda: the linked crate is 2.2.0 with
schema 1.9.1, acp_thread.rs matches acp::SessionUpdate::CompactionUpdate, and Zed
constructs acp::CompactionUpdate itself.
The correction is recorded in the comment rather than swapped in silently, so a
reader can tell the old premise was false rather than merely outdated.
files/0017-native-compaction-entry.patch is re-synced from zed-patches in the
same commit: its header carried the same claim, and refresh.sh preserves
everything above the --- on every bump, which is what made that copy permanent.
verify.sh ran before the sync, not after -- 19 of 19 apply onto 96e95eda -- so a
header edit that broke the patch could not be found inside this commit's window.
The patch diff below the --- is byte-identical, 25023 bytes before and after.
Fourth and last copy of the claim. The others: the adapter's
context-compaction-meta.ts, the patch header, and the series comment, which no
plan had budgeted for.
check-sync.sh agrees on all three relations, at 19 patches.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
commit e6a16b72862d879d87301f9ca1b75338cea554e1
Author: lucascouts <lucascs@protonmail.com>
Date: Tue Sep 22 20:46:30 2026 -0300
app-editors/zed: bump to 1.22.0_pre20260922-r1 (96e95eda)
Same UTC day as f25434f3, so the date component alone does not move and this is
a revision. Cargo.lock, the root Cargo.toml, crates/zed/Cargo.toml and
rust-toolchain.toml are all byte-identical between the two commits, and
pycargoebuild re-run from crates/zed reproduced the ebuild byte for byte; all
seven local *_COMMIT still resolve against GIT_CRATES.
The patch series needed a rebase of 0027, whose agent_panel.rs import hunk lost
its context; nine of the nineteen patches changed, the other eight by offset.
Verified: ebuild ... prepare applies all nineteen (>>> Source prepared); 11/11
[patch.crates-io] seds leave zero entries on 'git ='; cargo check --offline
-p zed -p remote_server --features gpui/inspector exits 0; md5-cache matches;
pkgcheck reports only the pre-existing deprecated-llvm-r1 finding. Not merged
with a build: this host has no sudo.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
commit 197a914dd92c344fa0ac77865de5471e59a6eec6
Author: lucascouts <lucascs@protonmail.com>
Date: Mon Sep 21 22:25:52 2026 -0300
mod(.autoupdate/packages.toml), mod(app-editors/zed/files/*), up(app-editors/zed-1.22.0_pre20260921 -> 1.22.0_pre20260922)
bcf6582c -> f25434f3 (2026-09-22T00:11 UTC). CRATES regenerated with
pycargoebuild from crates/zed: three rows moved (agent-client-protocol and
-derive 2.1.0 -> 2.2.0, -schema 1.7.0 -> 1.9.1); GIT_CRATES (71) and the
LICENSE+= block are byte identical, as are rust-toolchain.toml and
crates/zed/Cargo.toml.
Series carried forward through zed-patches: all 19 patches apply, 12 changed by
offset only. ebuild ... prepare applies all nineteen for real; 11/11
src_prepare seds match the new root Cargo.toml exactly once; the four USE-flag
seds yield valid TOML in all sixteen combinations; cargo check --offline
-p zed -p remote_server --features gpui/inspector exits 0 in 1m34; pkgcheck
reports only the pre-existing deprecated-llvm-r1 finding.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
commit e677e5b5843924eee31e8eb332c5c547306b14b8
Author: lucascouts <lucascs@protonmail.com>
Date: Mon Sep 21 20:44:51 2026 -0300
add(metadata/{md5-cache/app-admin/awscli2-2.36.50, md5-cache/app-editors/vim-9.2.1125, md5-cache/app-editors/vim-core-9.2.1125, md5-cache/dev-util/copilot-cli-bin-1.0.87, md5-cache/dev-util/crush-bin-0.96.1, md5-cache/dev-util/kimi-cli-bin-1.51.0, md5-cache/dev-util/opencode-bin-1.18.32, md5-cache/dev-util/qwen-code-0.24.3, md5-cache/dev-util/spirv-tools-1.4.357.0_p20260921, md5-cache/media-libs/vulkan-loader-1.4.363_p20260921, md5-cache/sys-apps/ai-jail-2.0.0, md5-cache/sys-apps/ai-jail-bin-2.0.0}), mod(app-editors/zed-1.22.0_pre20260921), mod(metadata/md5-cache/app-editors/zed-1.22.0_pre20260921, .autoupdate/packages.toml), up(app-admin/awscli2-2.36.49 -> 2.36.50, app-editors/{vim, vim-core}-9.2.1122 -> 9.2.1125, dev-util/{copilot-cli-bin-1.0.86 -> 1.0.87, crush-bin-0.96.0 -> 0.96.1, kimi-cli-bin-1.50.0 -> 1.51.0, opencode-bin-1.18.31 -> 1.18.32, qwen-code-0.24.2 -> 0.24.3, spirv-tools-1.4.357.0_p20260918 -> 1.4.357.0_p20260921}, media-libs/vulkan-loader-1.4.363_p20260918 -> 1.4.363_p20260921, sys-apps/ai-jail{,-bin}-1.22.0 -> 2.0.0)
commit d230559ad96e9a970a221f36d008328540871ddb
Author: lucascouts <lucascs@protonmail.com>
Date: Mon Sep 21 03:55:52 2026 -0300
app-editors/zed: bump to 1.22.0_pre20260921 (bcf6582c)
Cargo.lock, the root Cargo.toml and rust-toolchain.toml are byte-identical
between 916fc2b8 and bcf6582c, and pycargoebuild re-run from crates/zed
reproduced the ebuild byte for byte. All seven src_prepare git-dep revisions
still resolve against GIT_CRATES, and EGIT_COMMIT occurs exactly once: this was
a manual bump, so the applier's 40-hex rewrite never ran.
Fix a pre-existing defect the bump uncovered. USE=inspector passed the cargo
feature zed/inspector, which no longer exists: upstream reverted "Implement
`inspector` flag" (#62920) on 2026-09-15, so crates/zed/Cargo.toml has carried
no such key since _pre20260916 and cargo answers "none of the selected packages
contains this feature". The flag is default-off, which is why four bumps went by
without it surfacing. Point it at gpui/inspector, where the switch lived before
that PR and still does; gpui is a direct dependency of the zed crate, so the
dep/feature form resolves. zed/tracy and zed/mimalloc were re-checked and are
both still valid.
The patch series was three versions behind (files/ was generated against
_pre20260918-r1 by the applier) and was refreshed through zed-patches; all 18
apply and only 0018 changed, by offset.
Verified: `ebuild ... prepare` applies all eighteen for real; 11/11 sed patterns
match the new Cargo.toml exactly once, leaving zero [patch.crates-io] entries on
`git =` and all ten WORKDIR targets present; the four USE-flag seds produce
valid TOML in all sixteen wayland/X/screen-capture/neovim combinations;
`cargo check --offline -p zed -p remote_server --features gpui/inspector` exits
0 in 1m49 with zero errors; md5-cache _md5_ matches the ebuild and egencache
also swept the orphan _pre20260918-r1 entry; pkgcheck reports only the
pre-existing deprecated-llvm-r1 finding. Not merged with a build: this host has
no sudo, so no emerge ran.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
commit 273eb22a2316f7196deac76a37cd1fab9cece7a0
Author: lucascouts <lucascs@protonmail.com>
Date: Sat Sep 19 14:38:37 2026 -0300
add(metadata/{md5-cache/app-editors/vim-9.2.1122, md5-cache/app-editors/vim-core-9.2.1122, md5-cache/app-editors/zed-1.22.0_pre20260919, md5-cache/dev-libs/sentry-native-0.16.8, md5-cache/dev-util/claude-code-2.1.278, md5-cache/dev-util/qwen-code-0.24.1, md5-cache/media-libs/vulkan-layers-1.4.363_p20260919, md5-cache/sci-ml/ik_llama-cpp-0_pre20260919, md5-cache/sys-apps/ai-jail-1.22.0, md5-cache/sys-apps/ai-jail-bin-1.22.0}), mod(.autoupdate/packages.toml), up(app-editors/{vim-9.2.1119 -> 9.2.1122, vim-core-9.2.1119 -> 9.2.1122, zed-1.22.0_pre20260918-r1 -> 1.22.0_pre20260919}, dev-libs/sentry-native-0.16.6 -> 0.16.8, dev-util/{claude-code-2.1.277 -> 2.1.278, qwen-code-0.24.0 -> 0.24.1}, media-libs/vulkan-layers-1.4.363_p20260918 -> 1.4.363_p20260919, sci-ml/ik_llama-cpp-0_pre20260918 -> 0_pre20260919, sys-apps/ai-jail{,-bin}-1.21.0 -> 1.22.0)
commit 2306b89eab662662167959d7197955ec7b0ab831
Author: lucascouts <lucascs@protonmail.com>
Date: Fri Sep 18 19:52:20 2026 -0300
app-editors/zed: bump to 1.22.0_pre20260918-r1 (72b060af)
Same-day revision: upstream moved from 650a8d1b to 72b060af on 2026-09-18,
so the snapshot date is unchanged and the bump is a -r1.
CRATES regenerated with pycargoebuild from crates/zed (the tree root is a
virtual workspace and would be a silent no-op). Five crates dropped, none
added: hyper-rustls@0.24.2, rustls@0.21.12, rustls-webpki@0.101.7,
sct@0.7.1, tokio-rustls@0.24.1. GIT_CRATES is unchanged, and so are the
seven local *_COMMIT values in src_prepare -- the sha appears exactly once
in the ebuild.
Patch series refreshed onto the new source (all 18). Upstream added
Focusable to the gpui import block in conversation_view/elicitation.rs,
which conflicted with hunk 1 of 0006-elicitation-option-previews; the whole
series was re-refreshed from a corrected temporary source rather than
rebasing that one patch in files/, so the four patches that had only drifted
in offsets are carried forward too.
Verified without a full build:
- CRATES and GIT_CRATES diffed against the new Cargo.lock: no additions,
no removals, no stale revs.
- The src_prepare sed block run against the new Cargo.toml rewrites 11 of
11 [patch.crates-io] deps to path=, leaving no git= behind.
- The wayland/x11/screen-capture seds produce valid TOML in all four
combinations.
- ebuild ... clean prepare applies every patch, and cargo metadata
--offline then resolves the whole graph (exit 0).
Not covered: src_compile was not run.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
commit 14d7823bd90004bfa2ffdc73920a7574443badad
Author: lucascouts <lucascs@protonmail.com>
Date: Fri Sep 18 16:40:22 2026 -0300
add(metadata/{md5-cache/app-editors/cursor-3.21.13, md5-cache/app-editors/vim-9.2.1119, md5-cache/app-editors/vim-core-9.2.1119, md5-cache/app-editors/zed-1.22.0_pre20260918, md5-cache/app-misc/claude-desktop-bin-2.2553.1, md5-cache/app-text/calibre-9.15.0, md5-cache/dev-python/openai-3.16.0, md5-cache/dev-python/tavily-python-0.8.4, md5-cache/dev-python/uv-bin-0.12.16, md5-cache/dev-util/ccusage-bin-20.0.23, md5-cache/dev-util/claude-code-2.1.277, md5-cache/dev-util/codex-0.155.0, md5-cache/dev-util/codex-bin-0.155.0, md5-cache/dev-util/codex-desktop-bin-26.915.31945, md5-cache/dev-util/copilot-cli-bin-1.0.86, md5-cache/dev-util/glslang-1.4.357.0_p20260918, md5-cache/dev-util/kilocode-bin-7.7.5, md5-cache/dev-util/mesa_clc-26.3.0_pre20260918, md5-cache/dev-util/openchamber-bin-1.24.1, md5-cache/dev-util/reviewdog-0.21.2, md5-cache/dev-util/spirv-headers-1.4.357.0_p20260917, md5-cache/dev-util/spirv-tools-1.4.357.0_p20260918, md5-cache/dev-util/vulkan-headers-1.4.363_p20260918, md5-cache/dev-util/vulkan-tools-1.4.363_p20260918, md5-cache/dev-util/vulkan-utility-libraries-1.4.363, md5-cache/media-libs/mesa-26.3.0_pre20260918, md5-cache/media-libs/vulkan-layers-1.4.363_p20260918, md5-cache/media-libs/vulkan-loader-1.4.363_p20260918, md5-cache/net-misc/postman-bin-12.28.6, md5-cache/sci-ml/ik_llama-cpp-0_pre20260918, md5-cache/sci-ml/localai-4.10.0, md5-cache/sci-ml/ollama-0.34.2, md5-cache/sci-ml/ollama-bin-0.34.2, md5-cache/www-client/brave-browser-1.95.104, md5-cache/www-client/chromium-153.0.8010.52}), mod(.autoupdate/packages.toml), up(app-editors/{cursor-3.20.21 -> 3.21.13, vim-9.2.1117 -> 9.2.1119, vim-core-9.2.1117 -> 9.2.1119, zed-1.22.0_pre20260917-r1 -> 1.22.0_pre20260918}, app-misc/claude-desktop-bin-2.110.1 -> 2.2553.1, app-text/calibre-9.14.0 -> 9.15.0, dev-python/{openai-3.14.1 -> 3.16.0, tavily-python-0.8.3 -> 0.8.4, uv-bin-0.12.15 -> 0.12.16}, dev-util/codex{,-bin}-0.154.0 -> 0.155.0, media-libs/{mesa-26.3.0_pre20260917 -> 26.3.0_pre20260918, vulkan-layers-1.4.362_p20260917 -> 1.4.363_p20260918, vulkan-loader-1.4.362_p20260915 -> 1.4.363_p20260918}, net-misc/postman-bin-12.28.3 -> 12.28.6, sci-ml/ollama{,-bin}-0.34.1 -> 0.34.2, www-client/{brave-browser-1.95.102 -> 1.95.104, chromium-153.0.8010.47 -> 153.0.8010.52})
commit c6f4d45316ba48130f83303a2262f7c3238ab149
Author: lucascouts <lucascs@protonmail.com>
Date: Thu Sep 17 07:27:50 2026 -0300
app-editors/zed: bump to 1.22.0_pre20260917-r1 (b6171bcc)
Same UTC day as the packaged _pre20260917 (4b47ceb9), so this is a same-day
revision rather than a new date. SRC_URI already renames to $, so the two
distfiles do not collide.
CRATES and GIT_CRATES regenerated with pycargoebuild from crates/zed (the repo
root is a virtual workspace), EGIT_COMMIT edited afterwards because the
in-place mode rewrites everything else from the input. Against the new
Cargo.lock both blocks match exactly: 0 registry crates missing or extra, 71
git crates with identical revs. CRATES did move - the include-flate/libflate/
zstd group came in, the symphonia-* set went out.
The seven src_prepare *_COMMIT literals needed no change (grep -c of
EGIT_COMMIT is 1, and every rev the sed anchors on is still in the source
Cargo.toml): running the sed block over the new Cargo.toml turns 11 of 11
[patch.crates-io] entries into path=, leaving no git= behind. The USE-flag seds
still yield valid TOML in all eight wayland/X/screen-capture combinations.
Patch series refreshed from zed-patches onto b6171bcc: 18 of 18 apply, four
changed on offsets only, no orphans.
Verified: ebuild clean prepare applies all 18 patches, and cargo metadata
--offline in $ exits 0, so the whole graph resolves without network.
src_compile was not run.
Also drops the stale md5-cache entry for _pre20260916, which the previous bump
left behind.
commit acf840a35a3b098459c5c0be21582bb5a8da7420
Author: lucascouts <lucascs@protonmail.com>
Date: Thu Sep 17 00:16:42 2026 -0300
add(metadata/{md5-cache/app-editors/zed-1.22.0_pre20260917, md5-cache/dev-util/claude-code-2.1.274, md5-cache/dev-util/mesa_clc-26.3.0_pre20260917, md5-cache/dev-util/openchamber-bin-1.24.0, md5-cache/media-libs/mesa-26.3.0_pre20260917}), mod(.autoupdate/packages.toml), up(app-editors/zed-1.22.0_pre20260916 -> 1.22.0_pre20260917, dev-util/{claude-code-2.1.273 -> 2.1.274, mesa_clc-26.3.0_pre20260916 -> 26.3.0_pre20260917, openchamber-bin-1.23.2 -> 1.24.0}, media-libs/mesa-26.3.0_pre20260916 -> 26.3.0_pre20260917)
commit cd631aeb82cf3004b6097b922f3342a3f5d4b7bf
Author: lucascouts <lucascs@protonmail.com>
Date: Wed Sep 16 18:34:40 2026 -0300
add(metadata/{md5-cache/app-admin/awscli2-2.36.47, md5-cache/app-editors/devin-desktop-bin-3.10.31, md5-cache/app-editors/vim-9.2.1116, md5-cache/app-editors/vim-core-9.2.1116, md5-cache/app-editors/zed-1.22.0_pre20260916, md5-cache/app-editors/zed-bin-1.20.1, md5-cache/app-editors/zed-bin-1.21.0_pre, md5-cache/dev-util/ai-memory-2.3.0, md5-cache/dev-util/ai-memory-bin-2.3.0, md5-cache/dev-util/glslang-1.4.357.0_p20260916, md5-cache/dev-vcs/trufflehog-3.97.5, md5-cache/net-analyzer/netdata-2.11.1, md5-cache/net-libs/nodejs-26.9.0, md5-cache/net-p2p/bisq-bin-1.10.8, md5-cache/www-client/brave-browser-1.95.102}), mod(.autoupdate/packages.toml), up(app-admin/awscli2-2.36.46 -> 2.36.47, app-editors/{devin-desktop-bin-3.10.27 -> 3.10.31, vim-9.2.1112 -> 9.2.1116, vim-core-9.2.1112 -> 9.2.1116, zed-1.21.0_pre20260916 -> 1.22.0_pre20260916, zed-bin-1.19.2 -> 1.20.1, zed-bin-1.20.1_pre -> 1.21.0_pre}, dev-util/ai-memory{,-bin}-2.2.2 -> 2.3.0, dev-vcs/trufflehog-3.97.4 -> 3.97.5, net-analyzer/netdata-2.11.0 -> 2.11.1, net-libs/nodejs-26.8.2 -> 26.9.0, net-p2p/bisq-bin-1.10.7 -> 1.10.8, www-client/brave-browser-1.95.101 -> 1.95.102)
commit dabdc8c16e28ffdcdf15f2e68100f15fadfd55da
Author: lucascouts <lucascs@protonmail.com>
Date: Wed Sep 16 11:54:03 2026 -0300
add(mail-client/{thunderbird, thunderbird-l10n}-156.0), add(metadata/{md5-cache/app-editors/zed-1.21.0_pre20260916, md5-cache/dev-db/sqlitebrowser-3.13.99_p20260916, md5-cache/dev-util/copilot-cli-bin-1.0.85, md5-cache/dev-util/crush-bin-0.95.0, md5-cache/dev-util/mesa_clc-26.3.0_pre20260916, md5-cache/dev-util/qwen-code-0.24.0, md5-cache/mail-client/thunderbird-156.0, md5-cache/mail-client/thunderbird-l10n-156.0, md5-cache/media-gfx/openvdb-13.1.0, md5-cache/media-libs/mesa-26.3.0_pre20260916, md5-cache/net-dns/bind-tools-9.20.29, md5-cache/net-misc/postman-bin-12.28.2, md5-cache/sci-ml/amd-gaia-0.24.1, md5-cache/sci-ml/ik_llama-cpp-0_pre20260916, md5-cache/sys-kernel/linux-firmware-20260916, md5-cache/www-client/chromium-153.0.8010.47}), mod(.autoupdate/packages.toml), up(app-editors/zed-1.21.0_pre20260915-r1 -> 1.21.0_pre20260916, dev-db/sqlitebrowser-3.13.99_p20260913 -> 3.13.99_p20260916, dev-util/{copilot-cli-bin-1.0.83 -> 1.0.85, crush-bin-0.94.2 -> 0.95.0, mesa_clc-26.3.0_pre20260915 -> 26.3.0_pre20260916, qwen-code-0.23.4 -> 0.24.0}, media-gfx/openvdb-13.0.0 -> 13.1.0, media-libs/mesa-26.3.0_pre20260915 -> 26.3.0_pre20260916, net-dns/bind-tools-9.20.27 -> 9.20.29, net-misc/postman-bin-12.28.1 -> 12.28.2, sci-ml/{amd-gaia-0.23.0 -> 0.24.1, ik_llama-cpp-0_pre20260915 -> 0_pre20260916}, sys-kernel/linux-firmware-20260915 -> 20260916, www-client/chromium-153.0.8010.36 -> 153.0.8010.47)
commit 375bea2c592c51537dcfbfe86b737a93a475ec36
Author: lucascouts <lucascs@protonmail.com>
Date: Wed Sep 16 00:31:16 2026 -0300
app-editors/zed: revbump 1.21.0_pre20260915-r1, refresh patch series
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
commit 1728ef48f7752667ff89a23fcba82aaab5b2c46a
Author: lucascouts <lucascs@protonmail.com>
Date: Tue Sep 15 01:07:53 2026 -0300
app-editors/zed: bump to 1.21.0_pre20260915, and fix the inherited FTBFS
7f00507e0156 landed 2026-09-15 03:06 UTC; d62802d456ac was 2026-09-14
16:38 UTC, so this is a new date and not a revision. Upstream's own
version stays 1.21.0 and rust-toolchain.toml stays 1.98.1, which is
already the RUST_MIN_VER here.
pycargoebuild run from crates/zed left CRATES (1494 entries) and
GIT_CRATES (71) byte-identical against the ebuild -- the Cargo.lock is
byte-identical between the two commits, and running it anyway is what
proves the ebuild matches that lock rather than merely that the lock did
not move. The seven revisions hardcoded in the src_prepare sed still name
what [patch.crates-io] pins upstream, and clean prepare rewrites 11 of 11
lines to path=, leaving no git= behind.
The 1.21.0_pre20260914-r1 commit message recorded a known FTBFS and said
the fix would land through zed-patches. It has: 0017 now writes
`error: None` + `summary: Vec::new()` for the ContextCompaction shape
upstream moved to in 25b5569d, and clones the status, which stopped being
Copy when ContextCompactionStatus gained Other(Arc<str>). With that,
cargo check --offline passes on acp_thread, agent, agent_ui,
agent_servers and claude_code_ide -- the first green on this series since
_pre20260912-r1.
0026 joins PATCHES under claude-agent-acp-plus: the two end-of-turn
navigation controls float over the conversation instead of scrolling away
with the turn that owned them. It was authored after the last bump and
had never been wired in.
files/ is resynced from zed-patches, all three relations in sync. That
also drops 0024 from 13069 lines to 308: the published copy had swept
acp_thread.rs.orig and .rej into itself from a conflicted run, so it was
creating two junk files in $ on every merge.
pkgcheck reports only the pre-existing DeprecatedEclass notice for
llvm-r1. gentoo-parity exits 0 with all 9 rows JUSTIFIED.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
commit 45dc22b63133d26abf506f398c93437a6521b1d9
Author: lucascouts <lucascs@protonmail.com>
Date: Mon Sep 14 15:56:09 2026 -0300
add(metadata/md5-cache/app-misc/zeo-systray-0.1.2), del(metadata/md5-cache/app-misc/zeo-systray-0.1.1), up(app-misc/zeo-systray-0.1.1 -> 0.1.2)
commit 59d99dd229680f98ea2c9f6f2578a96a7f4cba9b
Author: lucascouts <lucascs@protonmail.com>
Date: Mon Sep 14 15:54:37 2026 -0300
app-editors/zed: bump to 1.21.0_pre20260914-r1
Same-day snapshot, so this is a revision and not a new date: d62802d456ac
landed 2026-09-14 16:38 UTC, after the 25b5569dd231 the applier packaged
that morning.
pycargoebuild run from crates/zed left CRATES (1494), GIT_CRATES and the
crate LICENSE+= block byte-identical -- the Cargo.lock did not move at all
since _pre20260912-r1. The seven revisions hardcoded in the src_prepare sed
still match [patch.crates-io] upstream, and ebuild clean prepare rewrites
11 of 11 lines to path=, leaving no git= behind.
files/ is resynced from zed-patches: 17 of 17 apply to d62802d456ac with
zero fuzz, nine drifted in offsets only. The two stale md5-cache entries
(_pre20260912-r1 and _pre20260914) went with the egencache pass.
KNOWN FTBFS, inherited, not introduced: cargo check --offline -p acp_thread
fails on files/0017 -- upstream changed ContextCompaction between 7960b2a7
and 25b5569d (summary: Option<Entity<Markdown>> became error: Option<...>
plus summary: Vec<ContentBlock>), and the patch still builds `summary:
None` in two places. The published _pre20260914 breaks identically; the
fix lands through zed-patches, not here.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
commit 47bd6a6307c631219999d2276cc6af970d0d714c
Author: lucascouts <lucascs@protonmail.com>
Date: Mon Sep 14 11:29:36 2026 -0300
add(metadata/{md5-cache/app-editors/zed-1.21.0_pre20260914, md5-cache/dev-util/crush-bin-0.94.2, md5-cache/media-libs/vulkan-layers-1.4.362_p20260914, md5-cache/net-misc/postman-bin-12.28.0, md5-cache/sci-ml/ik_llama-cpp-0_pre20260914}), mod(.autoupdate/packages.toml), up(app-editors/zed-1.21.0_pre20260912-r1 -> 1.21.0_pre20260914, dev-util/crush-bin-0.94.1 -> 0.94.2, media-libs/vulkan-layers-1.4.362_p20260913 -> 1.4.362_p20260914, net-misc/postman-bin-12.27.7 -> 12.28.0, sci-ml/ik_llama-cpp-0_pre20260910 -> 0_pre20260914)
commit fa55f9c1deb15341729f1bfec59738b5ea87f351
Author: lucascouts <lucascs@protonmail.com>
Date: Sun Sep 13 23:02:49 2026 -0300
app-editors/zed: bump to 1.21.0_pre20260912-r1
Same-day snapshot, so this is a revision and not a new date: 7960b2a7c956
landed 2026-09-12 20:38 UTC, hours after the 9d272b0363 this replaces.
pycargoebuild run from crates/zed left CRATES and GIT_CRATES byte-identical.
The Cargo.lock did move, but only in internal edges -- the zed crate gained
buffer_diff and git and dropped itertools -- so no registry or git package
entered or left. The eight revisions hardcoded in the src_prepare sed still
match [patch.crates-io] upstream, and replaying that block over the new
Cargo.toml rewrites 11 of 11 lines to path=, leaving no git= behind.
Evidence, not inference: ebuild clean prepare applies 17 of 17 patches, and
cargo metadata --offline inside $ exits 0, so the whole graph resolves
without the network. It was not compiled -- an FTBFS in code would still be
invisible here.
files/0002 and files/0024 are resynced from zed-patches, which is the source
of truth for both; each drifted in offsets only. The stale md5-cache entry
for 1.21.0_pre20260911-r2 went with the egencache pass.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
commit 4faa2152b55e6e237aed0016b5c5a7a55114b312
Author: lucascouts <lucascs@protonmail.com>
Date: Sun Sep 13 21:24:10 2026 -0300
add(metadata/md5-cache/app-misc/zeo-systray-0.1.1), del(metadata/md5-cache/app-misc/zeo-systray-0.1.0), mod(app-editors/zed-1.21.0_pre20260912), mod(metadata/{md5-cache/app-editors/zed-1.21.0_pre20260912, md5-cache/app-editors/zeo-0.1.0_p20260912}), up(app-misc/zeo-systray-0.1.0 -> 0.1.1)
commit 83eab793b8ba89f6d655b200a5e9ecc896b245cd
Author: lucascouts <lucascs@protonmail.com>
Date: Sat Sep 12 13:36:49 2026 -0300
add(metadata/{md5-cache/app-editors/cursor-3.20.17, md5-cache/app-editors/zed-1.21.0_pre20260912, md5-cache/dev-util/ai-memory-2.2.0, md5-cache/dev-util/ai-memory-bin-2.2.0, md5-cache/dev-util/crush-bin-0.94.1, md5-cache/media-libs/vulkan-layers-1.4.362_p20260912, md5-cache/sys-kernel/linux-firmware-20260912}), mod(.autoupdate/packages.toml), up(app-editors/{cursor-3.20.14 -> 3.20.17, zed-1.21.0_pre20260911-r2 -> 1.21.0_pre20260912}, dev-util/ai-memory{,-bin}-2.1.2 -> 2.2.0, media-libs/vulkan-layers-1.4.362_p20260911 -> 1.4.362_p20260912, sys-kernel/linux-firmware-20260911 -> 20260912)
commit d9889079af47a8959fcb075c66ab5396a81f0d33
Author: lucascouts <lucascs@protonmail.com>
Date: Fri Sep 11 11:02:54 2026 -0300
up(app-editors/zed-1.21.0_pre20260911-r1 -> 1.21.0_pre20260911-r2), mod(.autoupdate/packages.toml)
Same-day snapshot, so the bump is a revision: a9cdfc9936c4 lands on 2026-09-11
like d9e1c024f393 did.
This -r2 was first written against 3cef31688ae2 (11:10 UTC) and never published;
the packaged commit is now a9cdfc9936c4 (21:33 UTC), nine commits later.
Retargeted in place rather than published and followed by an -r3: a revision for
an intermediate commit nobody would install buys nothing, and SRC_URI names the
tarball after $, so reusing -r2 reuses the DIST slot too. That reuse is also
the Manifest trap -- the stale DIST line sat under exactly the right filename, as
did the stale distfile in DISTDIR, and both had to be replaced by hand before
`pkgdev manifest`, which otherwise has no reason to look at either.
Three lines of the ebuild move. CRATES gains ashpd 0.13.5 for 0.13.2 -- the whole
of what pycargoebuild, re-run from crates/zed, says changed: GIT_CRATES (71) and
the LICENSE+= block come out byte-identical, and unindent stays in CRATES even
though it left crates/migrator's dev-dependencies, other consumers holding it.
All seven `local *_COMMIT` were re-checked against the regenerated GIT_CRATES:
all seven resolve, none moved.
RUST_MIN_VER 1.97.1 -> 1.98.1 is not required by the new commit; it was already
wrong. rust-toolchain.toml had moved to 1.98.1 at 3cef3168, the very commit the
unpublished -r2 packaged. That axis is invisible to every local check -- the
applier never reads the file, `ebuild ... prepare` never consults the floor, and
cargo check passes precisely because this host's active toolchain is already what
the ebuild forgot to ask for. Only someone installing with an older rust finds it.
Six patches changed, offsets only (conversation_view.rs +7, crates/zed/src/main.rs
+3), refreshed through lucascouts/zed-patches, where the series is the source of
truth; files/ here is the generated copy and check-sync reports all three
relations in sync.
Verified: `ebuild ... prepare` applies all sixteen for real (>>> Source prepared);
11/11 sed patterns match the new Cargo.toml exactly once, leaving 0
[patch.crates-io] entries on `git =` and all ten WORKDIR targets present;
`cargo check --locked --offline -p zed -p remote_server --features zed/inspector`
exits 0 with zero errors, --locked on purpose because 0002 adds a Cargo.lock entry
whose pin is silent without it; md5-cache _md5_ matches the ebuild and egencache
touched only that one entry; pkgcheck reports only the pre-existing
deprecated-llvm-r1 finding. Not merged with a build: this host has no sudo, so no
emerge ran.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
commit 9d95d6c73efb6c55c9a27f2bc53ae20458224cf1
Author: lucascouts <lucascs@protonmail.com>
Date: Fri Sep 11 04:43:57 2026 -0300
up(app-editors/zed-1.21.0_pre20260911 -> 1.21.0_pre20260911-r1), mod(.autoupdate/packages.toml)
2518540 (01:07 UTC) -> d9e1c024 (05:25 UTC), the same UTC day, so the date
component alone does not move: -r1. An in-place swap was not available either,
because _pre20260911 was already pushed.
Cargo.lock and the root Cargo.toml are byte-identical between the two commits,
so the eleven src_prepare seds and the seven `local *_COMMIT` needed nothing.
All eighteen were re-checked rather than assumed.
What re-running pycargoebuild uncovered was in the two ebuilds ALREADY
PUBLISHED, neither of which would have built offline:
- _pre20260910 (86b2cf96) added objc2-core-graphics, objc2-io-surface and
objc2-screen-capture-kit to Cargo.lock; all three were missing from CRATES,
so they never reached SRC_URI.
- _pre20260911 (2518540) moved pet from microsoft/python-environment-tools
@bb8e0460 to zed-industries/python-environment-tools @5aad7164, and all 27
pet-* entries in GIT_CRATES kept naming the old repo and the dead rev.
Both are the applier defect recorded on 2026-08-25, recurring. An identical
Cargo.lock proves the SOURCE did not move between two commits; it never proves
the EBUILD ever matched it.
The patch series was 17 commits behind, not 2 -- files/ was generated against
52b2927a and the two applier bumps since never touched it. Refreshed through
zed-patches: all 16 apply, and only 0002 and 0012 changed, both purely
re-anchored (blob index hashes, Cargo.lock @@ -22769 -> -22807 from the objc2
crates, main.rs @@ -696 -> -698 from the two lines d9e1c024 inserted above the
hunk). No semantic rebase.
Verified: check-sync agrees on all three relations; `ebuild ... prepare` applies
all sixteen for real; 11/11 seds match exactly once, leaving 0 [patch.crates-io]
entries on `git =` with all ten WORKDIR targets present; `cargo check --offline
-p zed -p remote_server --features zed/inspector` exits 0 in 1m33 with zero
errors; md5-cache _md5_ matches; pkgcheck reports only the pre-existing
deprecated-llvm-r1 finding. Three orphan md5-cache entries swept.
Not merged with a build: this host has no sudo, so no emerge ran.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
commit 53b15b2081566a545a5066fe2801dbfa87355882
Author: lucascouts <lucascs@protonmail.com>
Date: Thu Sep 10 23:44:40 2026 -0300
add(metadata/{md5-cache/app-admin/awscli2-2.36.43, md5-cache/app-editors/cursor-3.20.10, md5-cache/app-editors/vim-9.2.1067, md5-cache/app-editors/vim-core-9.2.1067, md5-cache/app-editors/zed-1.21.0_pre20260911, md5-cache/app-misc/claude-desktop-bin-1.52386.0, md5-cache/dev-games/godot-4.8_alpha5, md5-cache/dev-python/openai-3.13.0, md5-cache/dev-python/uv-bin-0.12.13, md5-cache/dev-util/claude-code-2.1.268, md5-cache/dev-util/glslang-1.4.357.0_p20260910, md5-cache/kde-frameworks/kwindowsystem-6.30.0, md5-cache/net-im/telegram-desktop-7.2.8}), mod(.autoupdate/packages.toml), up(app-admin/awscli2-2.36.42 -> 2.36.43, app-editors/{cursor-3.19.19 -> 3.20.10, vim-9.2.1064 -> 9.2.1067, vim-core-9.2.1064 -> 9.2.1067, zed-1.21.0_pre20260910 -> 1.21.0_pre20260911}, app-misc/claude-desktop-bin-1.49585.0 -> 1.52386.0, dev-games/godot-4.8_alpha4-r1 -> 4.8_alpha5, dev-python/{openai-3.12.0 -> 3.13.0, uv-bin-0.12.12 -> 0.12.13}, dev-util/{claude-code-2.1.267 -> 2.1.268, glslang-1.4.357.0_p20260909 -> 1.4.357.0_p20260910}, net-im/telegram-desktop-7.2.7 -> 7.2.8)
commit e59e7c11f946617395df9aa0f66485aa460bd9ce
Author: lucascouts <lucascs@protonmail.com>
Date: Thu Sep 10 09:59:55 2026 -0300
up(app-editors/zed-1.21.0_pre20260909 -> 1.21.0_pre20260910, app-emulation/bochs-3.0-r1 -> 3.1, dev-util/{codex-desktop-bin-26.903.61454 -> 26.903.71938, mesa_clc-26.3.0_pre20260909 -> 26.3.0_pre20260910}, media-libs/{mesa-26.3.0_pre20260909 -> 26.3.0_pre20260910, vulkan-layers-1.4.362_p20260909 -> 1.4.362_p20260910}, net-misc/postman-bin-12.27.2 -> 12.27.5, sci-libs/onnxruntime{,-bin}-1.29.0 -> 1.29.1, sci-ml/ollama-bin-0.33.3 -> 0.34.0, sys-kernel/linux-firmware-20260831 -> 20260910), mod(.autoupdate/packages.toml)
The applier re-enabled three records that were pinned OFF and resurrected
the packages behind them. All three live in ::gentoo, which is why they
were dropped here in the first place, and two came back unfetchable:
sci-ml/ollama-0.34.0 Manifest carried ollama-0.33.3-deps.tar.xz's
size and hashes under the 0.34.0 name, and
LLAMA_CPP_tag was still 0.33.3's b10760
app-admin/calamares-3.4.3 same rename-without-recompute, this time on
calamares-3.4.2b-gentoo-patches.patch
sys-power/upower-1.91.4 fetchable, but byte-identical to ::gentoo's
1.91.3 ebuild: a duplicate, not a divergence
Only the ollama record tripped check-autoupdate-damage.sh, because it is
the one carrying a bare `enabled = false`. The other two also wore
`disabled_by = "auto"`, which the check reads as "the package came back
on purpose" -- true when a human re-adds it, false here. Restored all
three records verbatim from HEAD and dropped the three package
directories again.
Also

