We've had several outages with a proximate cause of "vercel is
complicated",
and auto-update is considered a critical feature; so lets not use vercel
for
that.
Release Notes:
- Auto Updates (and remote server binaries) are now downloaded via
https://cloud.zed.dev instead of https://zed.dev. As before, these URLs
redirect to the GitHub release for actual downloads.
Similar to #42246, we'd like to avoid having Vercel on the critical
path.
https://zed.dev/install.sh is served from Cloudflare by intercepting a
route on that page, so this makes the shell-based install flow vercel independent.
Release Notes:
- `./script/install.sh` will now fetch assets via
`https://cloud.zed.dev/`
instead of `https://zed.dev`. As before it will redirect to GitHub
releases
to complete the download.
This hopefully resolves the lingering test failures on linux,
but also adds some logging just in case this isn't the problem...
Release Notes:
- N/A
---------
Co-authored-by: Ben Kunkle <ben@zed.dev>
This is a follow up to #41482. When running `script/bundle-mac`, it will
upload debug symbols to Sentry if you have a `$SENTRY_AUTH_TOKEN` set. I
happen to have one set, so this script was trying to generate and upload
those. Whoops! This change skips the upload entirely if you're running a
local install.
Release Notes:
- N/A
One of the reasons we didn't spot that we were missing the telemetry env
vars for the production builds was that nightly (which was working) had
its own set of build steps. This re-uses those and pushes the env vars
down from the workflow to the job.
It also fixes nightly releases to upload all-in-one go so that all
platforms update in sync.
Closes#41655
Release Notes:
- N/A
I just pulled and ran a local build via `script/bundle-mac -l -i` but
found that the resulting bundle wasn't installed as expected. (me:
"ToggleAllDocks!! Wait! Where is it?!") Looking into, it looks like the
`-l` flag was removed in #41392, leaving the `$local_only` var orphaned,
which then left the `-i/$local_install` flag unreachable. I suspect that
this was unintentional, so this PR re-adds the `-l/$local_only` flag to
`script/bundle-mac`.
I ran the build again and confirmed that local install seemed to work as
expected. (ie "ToggleAllDocks!! 🎉")
While here, I also removed the last reference to `$local_arch`, because
all other references to that were removed in #41392.
/cc @osiewicz
Release Notes:
- N/A
---------
Co-authored-by: Conrad Irwin <conrad.irwin@gmail.com>
Reverts zed-industries/zed#41384
The branch-protection rules work much better when there is a Job that
runs every time and can be depended on to pass, we no longer have this.
Release Notes:
- N/A
Follow up for: #41304
Splits CI tests (cherry-picks and PRs only for now) into separate
workflows using `gh-workflow`. Includes a couple restructures to
- run more things in parallel
- remove our previous shell script based checking to filter tests based
on files changed, instead using the builtin `paths:` workflow filters
Splitting the docs/style/rust tests & checks into separate workflows
means we lose the complete summary showing all the tests in one view,
but it's possible to re-add in the future if we go back to checking what
files changed ourselves or always run everything.
Release Notes:
- N/A *or* Added/Fixed/Improved ...
---------
Co-authored-by: Conrad <conrad@zed.dev>
I noticed we had some typos that were getting through CI, but it looks
like the new version of `typos` catches them. So I updated it and fixed
them.
Release Notes:
- N/A
We've been considering removing workspace-hack for a couple reasons:
- Lukas ran into a situation where its build script seemed to be causing
spurious rebuilds. This seems more likely to be a cargo bug than an
issue with workspace-hack itself (given that it has an empty build
script), but we don't necessarily want to take the time to hunt that
down right now.
- Marshall mentioned hakari interacts poorly with automated crate
updates (in our case provided by rennovate) because you'd need to have
`cargo hakari generate && cargo hakari manage-deps` after their changes
and we prefer to not have actions that make commits.
Currently removing workspace-hack causes our workspace to grow from
~1700 to ~2000 crates being built (depending on platform), which is
mainly a problem when you're building the whole workspace or running
tests across the the normal and remote binaries (which is where
feature-unification nets us the most sharing). It doesn't impact
incremental times noticeably when you're just iterating on `-p zed`, and
we'll hopefully get these savings back in the future when
rust-lang/cargo#14774 (which re-implements the functionality of hakari)
is finished.
Release Notes:
- N/A
Retrieval stats will now use polars to build a big data frame for
references with the cartesian product of LSP declarations and retrieved
declaration candidates (with all their score components) and rebuilds
the stats summary on top of it.
This data frame is written to a `.parquet` file, which we can load into
advanced analytics tools (such as Metabase), so we can explore our
scoring distributions and find ways to improve retrieval, and then train
the decision tree.
Release Notes:
- N/A
This moves some of the changes made in
https://github.com/zed-industries/zed/pull/39543 to the `publish_gpui`
script.
This PR also updates that script to use `gpui_` instead of `zed-` (where
possible)
Release Notes:
- N/A
This PR is a follow-up to #39090 and addresses two issues:
* Moves `conpty.dll` and `OpenConsole.exe` out of the `bin` folder to
prevent other programs from using them.
* Updates these files only after Zed exits, avoiding update failures due
to file locks.
Release Notes:
- N/A
---------
Co-authored-by: Jakub Konka <kubkon@jakubkonka.com>
Closes#22657Closes#37863
# Background
Several users have noted that the terminal shipped with Zed on Windows
is either misbehaving or missing several features including lack of
consistent clearing behaviour. After some investigation which included
digging into the Microsoft Terminal project and VSCode editor, it turns
out that the pseudoconsole provided by Windows OS is severely outdated
which manifests itself in problems such as lack of clearing behaviour,
etc. Interestingly however, neither MS Terminal nor VSCode exhibit this
limitation so the question was why. Enter custom `conpty.dll` and
`OpenConsole.exe` runtime. These are updated, developed in MS Terminal
tree subprojects that aim to replace native Windows API as well as
augment the `conhost.exe` process that runs by default in Windows. They
also fix all the woes we had with the terminal on Windows (there is a
chance that ctrl-c behaviour is also fixed with these, but still need to
double check that this is indeed the case). This PR ensures that Zed
also benefits from the update pseudoconsole API.
# Proposed approach
It is possible to fork MS Terminal and instrument the necessary
subprojects for Rust-awareness (using `cc-rs` or otherwise to compile
the C++ code and then embed it in Rust-produced binaries for easier
inclusion in projects) but it comes at a cost of added complexity,
maintenance burden, etc. An alternative approach was proposed by
@reflectronic to download the binary from the official Nuget repo and
bundle it for release/local use. This PR aims to do just that.
There are two bits to this PR:
1. ~~when building Zed locally, and more specifically, when the `zed`
crate is being built, we will strive to download and unpack the binaries
into `OUT_DIR` provided by `cargo`. We will then set
`ZED_CONPTY_INSTALL_PATH=${OUT_DIR}/conpty` and use it at runtime in Zed
binary to tweak the loader's search path with that additional path. This
effectively ensures that Zed built from source on Windows has full
terminal support.~~ EDIT: after several discussions offline, we've
decided that keeping it minimal will serve us best, meaning: when
developing locally it is up to the developer of Zed to install
`conpty.dll` and put it in the loader's search path.
2. when bundling Windows release, we will download and unpack the nuget
package into Zed's bundle which will ensure it is installed in the same
directory as Zed by the installer.
**Note** I realise that 1. may actually not be needed - instead we could
leave that bit for the user who wants to run Zed from source to ensure
that they have `conpty.dll` in the loader's search path. I'd love to
hear opinions on this!
Release Notes:
- N/A
---------
Co-authored-by: Cole Miller <cole@zed.dev>
Follows on from
https://github.com/zed-industries/zed/pull/37717#discussion_r2376739687
@dvdsk suggested this but I didn't get to it in the previous PR.
# Tested
```
; sudo rm /usr/local/bin/wild
; ./script/install-wild
Downloading from https://github.com/davidlattimore/wild/releases/download/0.6.0/wild-linker-0.6.0-x86_64-unknown-linux-gnu.tar.gz
Wild is installed to /usr/local/bin/wild
To make it your default, add or merge these lines into your ~/.cargo/config.toml:
[target.x86_64-unknown-linux-gnu]
linker = "clang"
rustflags = ["-C", "link-arg=--ld-path=wild"]
[target.aarch64-unknown-linux-gnu]
linker = "clang"
rustflags = ["-C", "link-arg=--ld-path=wild"]
```
```
; sudo rm /usr/local/bin/mold
; ./script/install-mold 2.34.0
Downloading from https://github.com/rui314/mold/releases/download/v2.34.0/mold-2.34.0-x86_64-linux.tar.gz
Mold is installed to /usr/local/bin/mold
To make it your default, add or merge these lines into your ~/.cargo/config.toml:
[target.'cfg(target_os = "linux")']
linker = "clang"
rustflags = ["-C", "link-arg=-fuse-ld=mold"]
```
Release Notes:
- N/A
Wild changed in 0.6.0 to using gzip rather than xz, and changed the
format of the package name.
Follows on from and fixes
https://github.com/zed-industries/zed/pull/37717
cc @dvdsk @mati865
Release Notes:
- N/A
There's a mismatch between the URL used here and the one that's referred
to in `build_zed_cloud_url`, which prevents using the script on Windows.
A previous PR changed the script to use `127.0.0.1` instead of
`localhost` because of supposed URL parsing issues, but we were unable
to reproduce those.
Release Notes:
- N/A
# Summary
Today, Zed uses Mold on Linux, but Wild can be significantly faster.
On my machine, Wild is 14% faster at a whole-tree clean build, 20%
faster on an incremental build with a minimal change, and makes no
measurable effect on runtime performance of tests.
However, Wild's page says it's not yet ready for production, so it seems
to early to switch for production and CI builds.
This PR keeps using Mold in CI and lets developers choose in their own
config what linker to use. (The downside of this is that after landing
this change, developers will have to do some local config or it will
fall back to the default linker which may be slower.)
[Wild 0.6 is out, and their announcement has some
benchmarks](https://davidlattimore.github.io/posts/2025/09/23/wild-update-0.6.0.html).
cc @davidlattimore from Wild, just fyi
# Tasks
- [x] Measure Wild build, incremental build, and runtime performance in
different scenarios
- [x] Remove the Linux linker config from `.cargo/config.toml` in the
tree
- [x] Test rope benchmarks etc
- [x] Set the linker to Mold in CI
- [x] Add instructions to use Wild or Mold into `linux.md`
- [x] Add a script to download Wild
- [x] Measure binary size
- [x] Recommend Wild from `scripts/linux`
# Benchmarks
| | wild 0.6 (rust 1.89) | mold 2.37.1 (1.89) | lld (rust 1.90) | wild
advantage |
| -- | -- | -- | -- | -- |
| clean workspace build | 176s | 184s | 182s | 5% faster than mold |
| nextest run workspace after build | 137s | 142s | 137s | in the noise?
|
| incremental rebuild | 3.9s | 5.0s | 6.6s | 22% faster than mold |
I didn't observe any apparent significant change in runtime performance
or binary size, or in the in-tree microbenchmarks.
Release Notes:
- N/A
---------
Co-authored-by: Mateusz Mikuła <oss@mateuszmikula.dev>
Also removed the symbolicate script, which we could replace with a
`minidump-stackwalk` wrapper that downloaded sources/unstripped binaries
from github releases if that's helpful for folks.
Release Notes:
- N/A
