- User Since
- Feb 14 2018, 2:16 AM (228 w, 21 h)
Mon, Jun 27
Fri, Jun 24
hi Serge, sorry for the silence here. we'll be moving this towards a production ready version over the next months (also releasing some comments i've forgotten to send, we've already discussed these offline so no need for action).
we'll probably be leaving some todos/fixmes as we go, if you want to help it'd be great to keep an eye on changes going into the clang-tools-extra/include-cleaner directory (i'll also try to add you as a subscriber to them) and mention things you'd be interested in picking up.
luckily this document links were introduced only recently, hence they didn't make it to clangd-14, but they'll start being around from clangd-15 and such changes will be breaking all the links in existing clangd's.
i wonder if we should actually point these at releases.llvm.org/CLANG_VERSION/... instead of top of the head:
- downside, we don't have a release until it's there, so everything will be broken with a non-released clangd.
i suppose we can make use of different urls based on build configuration, but i couldn't find any existing mechanisms for doing so. hence this needs to be introduced.
another option would be to have a redirection from releases.llvm.org/15.0.0 to clang.llvm.org, until it's released, but the directory structure/subdomain is different so this will also need some adjustments.
Wed, Jun 22
I had another idea about offering the export pragma when in a header file but I don't know if that's going too far
Tue, Jun 21
I also agree with the typo correction verdict. In theory there'll be two cases:
- typo correction helps, in which case it'll be obvious from the warning itself.
- typo correction doesn't help, because the option doesn't exist at all, we'll be just showing a random option. I don't think how this'll be helpful.
as discussed offline I agree that we should have this, as no matter how hard we try there are going to be cases that we can't get right due to ADL/template instantiations or depending on 3rd party code that cannot be edited and also doesn't have relevant pragmas inside.
Get rid of the copy in common case
Mon, Jun 20
Split highlights into multiple tokens rather than trimming
Fri, Jun 17
thanks for the revert, you beat me to it.
Wed, Jun 15
- Leave the Ctx refactoring out of the patch.
thanks! you're right, it was already included in the arcanist metadata, i
was too lazy to check beforehand, sorry.
Just pass optional directly
do you have commit access or should i land this for you? if i should, can you provide an email address for attribution of the commit
Tue, Jun 14
Mon, Jun 13
Also if you put Fixes https://github.com/clangd/clangd/issues/1137 in the commit message, it'll automatically close the issue.
May 19 2022
Can you also update the remote-index protos & marshalling logic in:
May 18 2022
May 16 2022
sorry for the long turn around here, LGTM. let's ship it!
thanks for noticing this, LGTM!
66% win sounds great, it would be nice to have some detailed numbers (but this is clearly a huge win, so no need to reperform the experiments if numbers are gone)
thanks, LGTM. i agree that having this as a config to let users prioritise some sources over others could be useful, but it's too intrusive and unclear how useful that would be in practice (i don't think people would care to write a separate config for those)
May 13 2022
May 11 2022
May 10 2022
- Default initialize paddingLeft/Right
- Don't serialize kind when its null
- Run in vscode, seems to be working.
- Advertise inlayHintProvider
- Get rid of warning and not advertising when client has capability.
- Introduce paddingLeft/paddingRight, get rid of leading/trailing spaces in label, but re-introduce them when serializing for extension.
- Update documentation to reflect LSP docs.
May 9 2022
- Mention range as an extension field.
- Add lit test
Thx, LGTM. I agree that figuring out includes/macros in conditional branches is already best-effort, so I don't think this will result in noticeable regressions.
May 6 2022
May 5 2022
I agree that clangd's indexing should probably be higher priority than whatever FS indexing is going on (because clangd's indexing only start when user actually starts working on a project, hence this indicates some level of "interactiveness"). so I feel like the right default is Utility, rather than Background.
But we had this behavior for a long while and people didn't complain up until M1s. So this is definitely going to result in some behavior change for intel users, even if we agree on this is making the state better for M1s.
Unfortunately I am not in a position to say what does the Utility vs Background imply for intel chips for sure, to tell the behavior change isn't going to be a regression. If this renders next clangd unusable by intel mac folks, it would be really unfortunate.
as discussed offline, let's move forward with what you proposed (i.e. this patch, and then flip the default while making libclang opt-in)
Thanks this LGTM. Just two questions:
- Why not migrate rest of the usages and drop the old function completely in this patch
- What about moving the function from clang/Frontend to Tooling, Support or FrontendTool? this is clearly only inteded to be used by tools and not part of the Frontend. There's actually Tooling/Tooling.h which has a newInvocation that takes in cc1 args, rather than driver args. so these two seem like siblings.
Apr 29 2022
Apr 28 2022
Apr 27 2022
thanks, as discussed offline. this mostly LG. there are concerns about more code re-use, especially around handling pragmas but we should probably address them as we go rather than now.
Apr 26 2022
Apr 25 2022
can you please upload the patch with full context? see https://llvm.org/docs/Phabricator.html#requesting-a-review-via-the-web-interface
Apr 22 2022
Copy-assign to empty containers rather than clear.
Get rid of shrink_to_fit as PostingList representation doesn't really make a copy of the vector, but rather build a new representation.
Apr 21 2022
Apr 20 2022
Landed in 001e88ac83b5c3a4d4f4e61480953ebcabc82b88
Apr 19 2022
Apr 11 2022
Oopsy :/ Thanks for the fix!
- Get rid of leftovers and update comments
- Address comments and more cleanups