- User Since
- May 24 2017, 5:35 AM (73 w, 6 d)
Fri, Oct 19
Do you know the better way to accomplish my aim than adding an option to libFormat? For example making a dependent library which serves a little different purpose than libFormat itself or something simpler?
Wed, Oct 10
Sep 21 2018
Provide CorrectedBase = Base for C code
You're right actually. The fix is much simpler than I expected :)
Sep 19 2018
I tried that first but did not I find a way just to copy an expression (we basically need the same expr for such case). Do you know how to properly generate a copy of expression or some other way to get the same expression?
Test is added
Sep 7 2018
Sep 5 2018
Aug 28 2018
You are completely right! Thanks!
I did not think that -test-print-type also checks for the pointee.
Aug 26 2018
Aug 23 2018
Let's proceed with this one. I really see that it's going to be useful.
Jul 31 2018
File locking is the first one. Another one comes with "plugin mode" of tidy.
Anyways, if c++ part does not seem relevant I'm fine if we only have libclang part from D48116.
And we already saw, that -isystem is not the best choice for that. And by the way - clang-tidy has this filtering in consumer which does not exist in it's plugin-mode so it spits all system diagnostics all the time...
I already mentioned in the review inherited by this one that this way covers also loaded plugins with different consumers. So let's assume that driver loads clang-tidy and some other plugins and runs over each file in the project.
It makes sense not to include clang-tidy diagnostics coming from headers, at least for third-party headers (you can still run clang over specific headers themselves to get diagnostics for them).
Restore missing tests
Jul 25 2018
I'm mostly fine with your set of patches. Let me double check and we can get it in.
Jul 24 2018
@ilya-biryukov I would name this revision "for test purpose". It works at some extent with the current clangd version and can be used by someone else. But since we had a discussion about YAML being a temporary solution I never planned to commit this revision.
So yes, I totally agree that using a binary format would be ideal for the "index while build' use case.
Jul 13 2018
Jul 11 2018
Jul 6 2018
Jul 4 2018
@ilya-biryukov Sorry. I didn't have time to post comments here.
The usecase that we have is a supportive translation unit for code completion. Probably you use something similar in clangd not to wait for the TU to be reparsed after a change?
The gain from this change is both performance and memory but I don't have measurements under my hand right now. And of course they are only valid for this usecase.
Jul 3 2018
So I have that working on windows with clang-cl (probably broke something else during the porting, need to recheck.
One more commit I have in my working tree is YAML export to be able to use it for clangd. The quality of code is currently not so good so I probably need some extra time to make it ok-ish. But i hope to have it posted to reviews this week (unless I have something urgent)
Jul 1 2018
@nathawes I've already started changing so I can hopefully share my code with you when I fix all undefined symbols :)
Jun 29 2018
Will it work if I replace every block-related code here with llvm::function_ref<>?
I can't really test this one because INDEXSTORE_HAS_BLOCKS is apple-specific and it's required to rewrite all block-related stuff to use that code
Jun 27 2018
But this one misses a way to set this flag for everything except libclang. We can probably change that (Nikolai is in vacation till the end of summer so it's probably my part now) and add some tests outside of Index for it (probably Frontend)
The idea was to ignore everything including notes and remarks so that only errors from system headers are still collected.
Jun 21 2018
Thanks! I probably forgot to check line endings for this file before committing which I do for source files and headers...
@ilya-biryukov what do you think?
Ok, I reconsidered, let's remove IsTemplate :)
tests are back :)
Jun 20 2018
Can you probably update part 2 and 3 so they can be applied on top of the first patch?
Jun 19 2018
Thanks for the feedback :) I will return to this one and address comments a little bit later.
Jun 13 2018
Diff with unix line endings
Address comments about Index.h new functions parameters
Cool! That's actually quite a small change with big outcome!
Jun 7 2018
Address review comments
Jun 6 2018
Jun 4 2018
Use fixture in unit-test
fix indentation in unit-test
I could not properly test getMainBufferWithPrecompiledPreamble because it's private and requires some extra context to be called.
Jun 1 2018
May 31 2018
This is the proper fix. When we get a buffer for main file we should use the UserFilesAreVolatile flag to specify if memory mapping needs to occur or not.
May 30 2018
Memory mapping locks the file on Windows. That means that in the addressed case it locks main file which we are editing. After that happens many tools struggle to do something with this file, for example git.
It's the quickest solution, probably not the best. Better one would introduce isVolatile parameter in FileSystem::getBufferForFile which should be set to true somewhere (I didn't have time to search for the specific location yet)
May 29 2018
Address review comments
The precise path which leads here is yet unclear for me (it requires much more time to research) but changing this specific default value to true fixes main file from being memory mapped and therefore being blocked (on Windows)
I assume it comes from FileManager::getBufferForFile(StringRef Filename) but not sure about prior events causing it.
May 25 2018
Update according to c++ part changes
May 24 2018
May 23 2018
Output range of token being replaced instead of it's text