- User Since
- Sep 11 2015, 9:59 AM (205 w, 4 d)
Tue, Aug 6
Jul 13 2019
Jul 5 2019
Jun 30 2019
Jun 22 2019
Jun 17 2019
Jun 7 2019
Jun 2 2019
Jan 1 2019
Improve PointerUnion expansion to show what type it holds
Jun 12 2016
No adverse (or otherwise) comments received, so committing as revision 272525
Jun 10 2016
Mar 28 2016
Closing diff after commit
Commited as revision 264603
Committed as revisions 264601 and 264602
This is still showing as "Needs review" Are there any further changes required or can someone accept it? I installed both VS2015 and VS2013 on my laptop and saw the correct solution files built by cmake.
Mar 27 2016
Adding missing period in comment!
Incorporating ariccio's comments and fixing some html
The file was moved to a more appropriate directory (Otherwise, the project files were stepping on each other).
svn mv utils/llvm.natvis utils/LLVMVisualizers/llvm.natvis
Now that we are generating projects for native visualizers, have it properly reflected in the directory structure
Add missing newline
Since this is now creating a project, we should do some housekeeping to move the files to different directories so that the build tree has an LLVMVisualizers project directory.
I understand your concerns, but on balance, I don't see a better course of action than what I've done. If I have some time, I'll submit a CMake change so we can eventually migrate to having CMake generate a Natvis tag, but that will take a long time to propagate, so I don't want to delay this operationally semantically equivalent change. Sound ok?
No. Testing shows that it works fine in the project with the <None> tag. VS2015 empirically looks at the extension.
I'm not sure I understand. I'm not installing anything for VS2013. You still have to manually copy into Documents\Visual Studio 2013\Visualizers. The purpose of this change is to eliminate this manual step for VS2015 by leveraging the new Natvis project support. Am I missing something?
What do you mean by "making both paths the same"?
Hmmm, I just tried adding set_source_files_properties(utils/clang.natvis LANGUAGE natvis) similar to what you suggest in http://lists.llvm.org/pipermail/llvm-dev/2016-January/093887.html, but it still produces the None tag in the .vcxproj file.
Apply whitespace comments from D18497 mutatis mutandis to this change
Missed one (the missing period in llvm.natvis).
Accepting Aaron's suggestions
Correctly handle non-Visual Studio case (e.g., Linux)
Mar 20 2016
Closing accepted revision
Committed as revision 263915.
Closing committed revision
Per Aaron, I want to close the committed diff, but I seem to need to accept it first. Further discussion as needed on mailing list
Closing committed revision. Further discussion as needed on mailing list
Accepting revision based on Aaron's LGTM email to reviews.llvm.org.
Mar 19 2016
Based on Aaron's agreement to conducting post-commit reviews, I've committed this as revision 263890. Please let me know if I should not have done so. (This change is a lot smaller and simpler than I feared when I commented on my roadmap in diff D18270, so I think this is a good inaugural post-commit review for me).
Mar 18 2016
Committed as revision 263821.
Thanks, Aaron. Committed as revision 263794
I think the code is correct as written, see response in comment
Mar 17 2016
Mar 16 2016
Forgot to close the revision
Committed as revision 263638. At Chandler's suggestion, I am going to send an announcement to cfe-dev letting people know of all of the recent improvements to native visualizers
Added reference types and a more accurate AttributedType enum. I am working on additional items for the future (class members, function types, etc.) but those will take longer, and I think it is worthwhile to get this revision committed unless there are defects in it. Agree?
Mar 14 2016
Committed as revision 262933
Mar 11 2016
Committed as svn revision 263270
Applying same logic to PointerUnion3 and PointerUnion4 as suggested by Aaron
Good catch. Give me two minutes....
Mar 10 2016
Mar 8 2016
Mar 7 2016
Based on Eric Feiveson's comment that I can suppress the numeric value of an enum visualization with ", en" I have uploaded a new diff. By suppressing the numeric values, instead of the verbose TemplateArgumentLoc visualization
Updated in response to review comments to fix issues around MSVC versions
Hmm, testing reveals that the preexisting clang.natvis does not work in VS2013!
I believe that it will all work on VS2013 (and have kicked off a build to confirm). I guess I would rather drop the number so it doesn't get out of sync whenever the base compiler version changes.
Mar 5 2016
Sep 11 2015
Thanks, Aaron! Hal mentioned that I might need to ask you to commit on my behalf as I don't have commit access as of now.