This pull request introduces an additional step to
`SemanticIndex::search_project` that includes the content of buffers
that are modified but haven't been saved yet. In most cases, the buffer
will contain a small portion of changed spans that are potentially not
included in the index. To reuse all the other spans that haven't
changed, we will query the database for embeddings by their digest. This
means we have to index spans by their digest, which means some penalty
when writing, but in our tests this didn't seem to make indexing much
slower.
Release Notes:
- Improved semantic search to include results from modified buffers.
(preview-only)
Fixes
```
thread 'main' panicked at 'byte index 2 is not a char boundary; it is inside '…' (bytes 0..3) of `…)`'
```
panics like
https://zed-industries.slack.com/archives/C04S6T1T7TQ/p1694535396473329
by reworking the inlay hover model:
* avoid storing "hardcoded" coordinates of hovered inlay labels (such as
`InlayOffset`), instead, remember `inlay_id` and apply required
highlights there when hint with the same id is handled
* add randomized tests on inlay highlights
* sped up inlay hint cache lookup by inlay_id
As a downside, background highlights are no long appearing on inlay
hints, but Zed does not receive any tooltips for inlays anyway (r-a does
not send them for some reason, other LSP seem to have no such feature?),
so it does not matter now.
Nontheless, if the logic for displaying hint pop-ups is present and
works for harcoded tooltips in r-a, only background highlight is missing
now.
Release Notes:
- Fixed inlay hint highlights causing panic for certain cases with
"large" characters
This allows users to add a keybind to ToggleReplace from Editor/Pane
contexts.
Release Notes:
- Fixed replace in buffer not reacting to keyboard shortcuts outside of
search bar<preview-only>.
We're using [PostgREST](https://hub.docker.com/r/postgrest/postgrest)
Todo:
* [x] Add instructions for installing postgrest to local development
docs
* [x] Deploy to staging
* [x] Deploy to production
* [x] Add DNS record for `admin-staging.zed.dev` pointing to the staging
db
* [x] Add a DNS record for `admin.zed.dev` pointing to the production db
Small fix to update code for rate limiting status.
Release Notes (Preview only)
- Fixed update to only stop updating status, when the rate limit is
reset to None
Users of keyboard layout with IME complained about the peculiar
behaviour where typing in "sss" and then removing all of it left behind
one 's' and also appended a closing brace. This was not reproducible on
a buffer without language, so I've suspected that brace insertion might
be a problem here. For whatever reason when the user removes the last
character from a run that triggered IME, we receive a notification about
an empty insertion. Sadly, brace completion does not handle an empty
input properly and we erroneously insert a closing brace when deleting
the followup characters. In fact, the brace inserted is always the
closing brace for the first entry in language's config.toml 'brackets'
field (see Scheme vs Markdown). This guard also allows for the proper
removal of the first character.
Closes community tickets zed-industries/community#877zed-industries/community#1329
Z-2869
Release Notes:
- Fixed handling of bracket completion for international keyboard
layouts that use IME. This led to Zed erroneously inserting the `}`
character while removing the first character that triggered IME.
This pull request extracts a separate `Codegen` struct that models the
interaction with OpenAI and takes care of diffing, auto-indenting and
reporting regions to highlight.
As part of this refactoring, we're now relying less on the AI model to
indent code. The new logic lets tree-sitter decide how the first line
should be indented. Then, for every subsequent line reported by ChatGPT,
it calculates an indent delta relative to the first reported line and
applies it to the indent level chosen by tree-sitter.
Release Notes:
- Improved auto-indentation when using the inline assistant.
