Conversation
ee63506 to
3c53233
Compare
Assisted-by: Google Antigravity with Gemini 3 Flash
3c53233 to
9f3aa4b
Compare
|
@chandlerc Moving this way since normally I wouldn't have made this a proposal, but you requested it as part of discussion of #6641, so I'm trying to figure out the expected threshold for toolchain proposals. |
chandlerc
left a comment
There was a problem hiding this comment.
Generally, LG, and nice to have this documented.
Some editorial stuff and a question about where best to put the alternatives discussion.
proposals/p6699.md
Outdated
| ## Alternatives considered | ||
|
|
||
| Alternatives are in the | ||
| [toolchain design](/toolchain/docs/diagnostics.md#diagnostic-sort-alternatives). |
There was a problem hiding this comment.
I had imagined following a similar structure to what we've been doing with the language design, where the design just has a short list of alternatives, and the actual discussion is kept in the proposal doc.
Did I give a different impression in the discussion? If so, I may have just misspoken...
Or is there something that makes you prefer this structure?
There was a problem hiding this comment.
Sorry, this is a detail I heard from you -- I just don't think it came up. I'd assumed we'd continue to document alternatives in the toolchain's design, as has been done in the past. For example:
https://github.com/search?q=repo%3Acarbon-language%2Fcarbon-lang%20path%3Atoolchain%2Fdocs%20%22Alternatives%20considered%22&type=code
Since it sounds like your preference is to use proposals for more, would you suggest moving all the existing alternative documentation out to proposals for consistency? (or is being consistent with the existing documentation good?)
Note too, I think other documentation has included some alternative documents inline.
There was a problem hiding this comment.
Not sure there is a lot of value to retroactively moving stuff around, particularly for the toolchain.
But we should pick the pattern we want going forward and try to be consistent with that maybe? I've been happy with the approach in the language design where the alternatives are just listed in the design but discussed in the proposal in question. But I'm open to other organization if that's better.
WDYT @zygoloid ?
There was a problem hiding this comment.
It's really not much extra work (and I think it'd be useful to move in order to avoid confusion with future contributions, especially since otherwise we'll end up with both styles in the same document) -- #6716
Co-authored-by: Chandler Carruth <[email protected]>
Co-authored-by: Chandler Carruth <[email protected]>
Assisted-by: Google Antigravity with Gemini 3 Flash
chandlerc
left a comment
There was a problem hiding this comment.
Also LGTM and up to you if you'd like to submit or wait for a how we want to organize stuff -- not super worried about a small amount of inconsistency either way.
Assisted-by: Google Antigravity with Gemini 3 Flash
Change
SortingConsumerfrom sorting by last processed token (per-phase) toadditionally allow diagnostics to request sorting by start position (line and
column) when the last processed token is the same.
Assisted-by: Google Antigravity with Gemini 3 Flash