fix: windows can now paste non-ascii multiline text #8774
Conversation
dylan-hurd-oai
changed the title
dylan-hurd-oai
force-pushed
the
fix-windows-paste
branch
from
9803d17 to
2ce8cc7
Compare
![chatgpt-codex-connector[bot]](https://avatars.githubusercontent.com/in/1144995?s=60&v=4){kind=small}
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 2617290692
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
dylan-hurd-oai
force-pushed
the
fix-windows-paste
branch
from
c2c624f to
1d604f5
Compare
| } | ||
|
|
||
| #[test] | ||
| fn question_mark_does_not_toggle_during_paste_burst() { |
There was a problem hiding this comment.
exists in tui/chat_composer.rs, copying over
| const PASTE_BURST_MIN_CHARS: u16 = 3; | ||
| const PASTE_BURST_CHAR_INTERVAL: Duration = Duration::from_millis(8); | ||
| const PASTE_ENTER_SUPPRESS_WINDOW: Duration = Duration::from_millis(120); | ||
| const PASTE_BURST_ACTIVE_IDLE_TIMEOUT: Duration = Duration::from_millis(60); |
There was a problem hiding this comment.
Would be good to capture the reason for the differences in these intervals - are they based on testing / docs / ? Do we have more context on why we might increase or decrease? Are they likely to cause differences in operating systems / remote systems / multiplexers?
Maybe just something here that captures why we chose is enough and we can come back to the deeper questions.
There was a problem hiding this comment.
In Windows, if we use an 8ms interval, copying long non-ASCII text might cause the Pasted Content Block to be split into multiple chunks, indicating that the paste burst was temporarily exited midway through the paste process. Sometimes, it might even cause line breaks to be recognized as commits. I'm not sure if there's a better way.
The following images show the results caused by two different delays.
8ms:
60ms:
There was a problem hiding this comment.
Thanks - mostly I want a short summary in the code, so we know why / how to tweak it if needed.
There was a problem hiding this comment.
Thanks @occurrent! I added a comment here and made this platform-specific
|
@occurrent thanks again for your contribution here - we really appreciate the time you took to draft this change. I'll make sure you're credited in the next release notes for this PR |
Summary
This PR builds heavily on the work from @occurrent in #8021 - I've only added a small fix, added additional tests, and propagated the changes to tui2.
From the original PR:
Testing