jordan_rose (Jordan Rose)
User

Projects

User does not belong to any projects.

User Details

User Since
Aug 29 2012, 11:21 AM (264 w, 2 d)

Recent Activity

Tue, Sep 19

jordan_rose updated the diff for D38010: lit.py: Allow configs and local configs to have a setup_script entry.

Move setup code from execute_tests_in_pool to execute_tests.

Tue, Sep 19, 5:33 PM
jordan_rose added a comment to D38010: lit.py: Allow configs and local configs to have a setup_script entry.

Yep, that's not bad either. I'm not super happy with the idea of lit.local.cfg parsing relying on running in the same process as lit itself, but maybe that ship has sailed.

Tue, Sep 19, 4:46 PM
jordan_rose added a comment to D38010: lit.py: Allow configs and local configs to have a setup_script entry.

Oh, no, I definitely need an arbitrarily complex script. Swift wants to prebuild mock libraries shared by several test cases. Today we're either doing it once per test case or using a terrible hack mode that can parse source on the fly.

Tue, Sep 19, 4:12 PM
jordan_rose added a comment to D38010: lit.py: Allow configs and local configs to have a setup_script entry.

You mentioned you want to make this work with the multiprocessing pool. That worries me a little bit because different processes are going to be using the same shared output dir (which isn't created up front by the infrastructure), if I understand correctly. Like if you have A and A/B and A/C, then we will run the setup script 3 times in parallel all using the same shared output folder. Is this going to be a source of flakiness?

Tue, Sep 19, 3:49 PM
jordan_rose added a comment to D37756: [lit] Force site configs to be run before source-tree configs.

Previously I was just able to do this:

Tue, Sep 19, 3:45 PM
jordan_rose updated the diff for D38010: lit.py: Allow configs and local configs to have a setup_script entry.

Make setup_script refer to a Python script rather than an arbitrary executable, as suggested by @zturner.

Tue, Sep 19, 3:29 PM
jordan_rose added a comment to D37756: [lit] Force site configs to be run before source-tree configs.

I'm unhappy that I can't run lit's own tests without having to configure, though. That seems like a regression.

Tue, Sep 19, 3:12 PM
jordan_rose added inline comments to D38010: lit.py: Allow configs and local configs to have a setup_script entry.
Tue, Sep 19, 11:48 AM
jordan_rose added inline comments to D38010: lit.py: Allow configs and local configs to have a setup_script entry.
Tue, Sep 19, 11:34 AM
jordan_rose added a comment to D38010: lit.py: Allow configs and local configs to have a setup_script entry.

Actually, I guess I'm a bit confused. What is this solving that you can't already do in a local config? Couldn't you just embed the stuff you want to execute in the local config?

Tue, Sep 19, 11:33 AM

Mon, Sep 18

jordan_rose added a reviewer for D37756: [lit] Force site configs to be run before source-tree configs: ddunbar.
  • Run lit.py from the source tree against a directory in the build tree.

    Not aware of anyone who does this, but in theory it should work.
Mon, Sep 18, 8:11 PM
jordan_rose created D38010: lit.py: Allow configs and local configs to have a setup_script entry.
Mon, Sep 18, 5:01 PM

Jul 19 2017

jordan_rose updated subscribers of D35533: [Basic] Update CMakeLists.txt to handle Repo with git.

The addition of "${path}/.git/HEAD" at line 37 is used only once when repo is initially synced "for the first time". This is because ${git_path}/logs/HEAD file at line 34 for Git or Git submodule does not exist with initial repo sync[.]

Jul 19 2017, 9:37 AM

Jul 18 2017

jordan_rose added a comment to D35138: [cmake] Remove deprecated GetSVN.cmake parameters.

Rather than delete these outright, it's probably polite to error on the presence of FIRST_SOURCE_DIR. There might be other out-of-tree projects that need to update.

Jul 18 2017, 2:41 PM
jordan_rose added a comment to D35132: [cmake] GetSVN.cmake takes a list of arguments.

It looks fine to me too. Ideally @beanz would be around to weigh in but it's not such a big change. I think it's okay to land.

Jul 18 2017, 2:41 PM
jordan_rose added a comment to D35533: [Basic] Update CMakeLists.txt to handle Repo with git.

As commented at D34955#798369, this isn't sufficient because it doesn't force a rebuild when new changes are merged in from upstream. It's not harmful to stick with the old version information, but if you can find something better that would be, um, better.

Jul 18 2017, 10:26 AM

Jul 17 2017

jordan_rose accepted D34955: [Basic] Detect Git submodule version in CMake.

Whoops, yes, this new version looks good!

Jul 17 2017, 10:31 AM

Jul 10 2017

jordan_rose added inline comments to D34955: [Basic] Detect Git submodule version in CMake.
Jul 10 2017, 2:17 PM
jordan_rose edited reviewers for D35186: [analyzer] Add annotation for functions taking user-facing strings, added: dcoughlin; removed: jordan_rose.
Jul 10 2017, 10:15 AM

Jul 5 2017

jordan_rose added a comment to D34955: [Basic] Detect Git submodule version in CMake.

I'm not sure why we would care if a commit is made on a branch. The important information here is what commit is checked out, and that's what HEAD refers to.

Jul 5 2017, 1:42 PM

Jul 3 2017

jordan_rose requested changes to D34955: [Basic] Detect Git submodule version in CMake.

If I'm remembering correctly from when I set this up, this isn't just about detecting which version control system you're using; it's about finding a file that will change on every commit. Otherwise, the generated Version.cpp won't be rebuilt after you update.

Jul 3 2017, 9:20 AM

Apr 10 2017

jordan_rose resigned from D31861: [Lexicon] Add BDCE.
Apr 10 2017, 11:07 AM

Mar 28 2017

jordan_rose added a comment to D26108: Add -Wnullability-completeness-on-arrays..

Yep, that's the correct syntax. It's not wonderful, but it's already where the C standard lets you put const, so we're just following established practice. (Learn something new and awful about C every day!)

Mar 28 2017, 10:39 AM
jordan_rose added a comment to D26108: Add -Wnullability-completeness-on-arrays..

No, the arrays themselves need to be marked as nullable or non-nullable. Remember that in C array parameters are passed as pointers.

Mar 28 2017, 10:33 AM

Mar 27 2017

jordan_rose updated subscribers of D31372: Support Microsoft mangling of swift calling convention methods.
Mar 27 2017, 9:54 AM
jordan_rose edited reviewers for D31372: Support Microsoft mangling of swift calling convention methods, added: rjmccall, aschwaighofer; removed: jordan_rose.
Mar 27 2017, 9:54 AM

Mar 22 2017

jordan_rose added a comment to D22391: [Sema] Add warning for implicitly casting a null constant to a non null pointer type .

This looks like it's only coming up for declarations. What about assignments?

Mar 22 2017, 6:14 PM

Mar 15 2017

jordan_rose added inline comments to D31005: [Objective-C] Fix "repeated use of weak" warning with -fobjc-weak.
Mar 15 2017, 5:05 PM
jordan_rose added a reviewer for D31005: [Objective-C] Fix "repeated use of weak" warning with -fobjc-weak: rjmccall.

The warning-related parts look reasonable to me.

Mar 15 2017, 3:42 PM

Mar 10 2017

jordan_rose closed D30855: [unittest] Explicitly specify alignment when using BumpPtrAllocator..

Committed in rL297540.

Mar 10 2017, 5:37 PM
jordan_rose edited reviewers for D30855: [unittest] Explicitly specify alignment when using BumpPtrAllocator., added: bogner; removed: dexonsmith.
Mar 10 2017, 5:04 PM
jordan_rose created D30855: [unittest] Explicitly specify alignment when using BumpPtrAllocator..
Mar 10 2017, 5:03 PM

Mar 8 2017

jordan_rose closed D30723: Add red zones to BumpPtrAllocator under ASan.

Committed in rL297310.

Mar 8 2017, 2:05 PM
jordan_rose added a comment to D30723: Add red zones to BumpPtrAllocator under ASan.

Thanks, Pete! Will land soon, then.

Mar 8 2017, 11:42 AM

Mar 7 2017

jordan_rose created D30723: Add red zones to BumpPtrAllocator under ASan.
Mar 7 2017, 3:50 PM

Feb 28 2017

jordan_rose accepted D30482: [Analyzer] Fix crash in ObjCPropertyChecker on protocol property.

Looks good to me. Good catch.

Feb 28 2017, 5:17 PM

Feb 7 2017

jordan_rose edited reviewers for D29181: Make TrailingObjects private inheritance of TrailingObjectsImpl protected, added: jyknight, rsmith; removed: llvm-commits.
Feb 7 2017, 7:26 PM

Jan 24 2017

jordan_rose added inline comments to D22391: [Sema] Add warning for implicitly casting a null constant to a non null pointer type .
Jan 24 2017, 4:39 PM

Jan 19 2017

jordan_rose closed D28924: [AST Printer] Print attributes on enum constants.

Landed as-is in rL292571.

Jan 19 2017, 7:45 PM
jordan_rose added a comment to D28924: [AST Printer] Print attributes on enum constants.

Interestingly, this case doesn't actually work yet; we unconditionally print enum attributes after the close-brace even though that's not valid for C++ attributes. Do you think I should change that as well, or just land this patch as is?

Jan 19 2017, 5:25 PM
jordan_rose added a comment to D28924: [AST Printer] Print attributes on enum constants.

Good idea, will do.

Jan 19 2017, 4:15 PM
jordan_rose created D28924: [AST Printer] Print attributes on enum constants.
Jan 19 2017, 4:01 PM

Dec 19 2016

jordan_rose closed D27837: Add fix-it notes to the nullability consistency warning.
Dec 19 2016, 1:12 PM
jordan_rose accepted D27837: Add fix-it notes to the nullability consistency warning.

Doug was okay with the idea and I trust Alex and Akira would have caught any major problems. Committed in r290132.

Dec 19 2016, 1:12 PM

Dec 16 2016

jordan_rose updated the diff for D27837: Add fix-it notes to the nullability consistency warning.

Updated names to match LLVM style.

Dec 16 2016, 2:26 PM
jordan_rose added inline comments to D27837: Add fix-it notes to the nullability consistency warning.
Dec 16 2016, 8:53 AM

Dec 15 2016

jordan_rose retitled D27837: Add fix-it notes to the nullability consistency warning from to Add fix-it notes to the nullability consistency warning.
Dec 15 2016, 6:02 PM

Nov 14 2016

jordan_rose added a comment to D26568: [ADT] Delete Twine constructors which accept StringRef rvalues (NFC).

I'm inclined to agree with @dblaikie's email comment: assigning a Twine is pretty much never safe anyway, so we should just delete operator= altogether. (I'd love to get rid of the copy-constructor too and just allow moves, but I think it still makes sense to copy-construct when combining a Twine argument with multiple different suffixes.)

Nov 14 2016, 11:15 AM

Nov 10 2016

jordan_rose added a comment to D26227: Don't require nullability on 'va_list'..

One more fix-up in rL286542. Hopefully the last one.

Nov 10 2016, 5:39 PM
jordan_rose closed D26227: Don't require nullability on 'va_list'..

Committed in rL286522, with a fix-up to make the check for va_lists more conservative in rL286531.

Nov 10 2016, 4:50 PM
jordan_rose closed D26109: Warn when 'assume_nonnull' infers nullability within an array..

Committed as rL286521.

Nov 10 2016, 4:49 PM
jordan_rose closed D26108: Add -Wnullability-completeness-on-arrays..

Committed as rL286520, with a slight fix-up for MSVC in rL286525.

Nov 10 2016, 4:49 PM
jordan_rose closed D25850: Accept nullability annotations (_Nullable) on array parameters.

Committed as rL286519.

Nov 10 2016, 4:48 PM

Nov 9 2016

jordan_rose added inline comments to D26108: Add -Wnullability-completeness-on-arrays..
Nov 9 2016, 5:21 PM
jordan_rose added a comment to D26108: Add -Wnullability-completeness-on-arrays..

Ah, my apologies, @ahatanak. I was testing against the Swift branch of Clang.

Nov 9 2016, 4:50 PM

Nov 8 2016

jordan_rose added a dependent revision for D26108: Add -Wnullability-completeness-on-arrays.: D26109: Warn when 'assume_nonnull' infers nullability within an array..
Nov 8 2016, 6:10 PM
jordan_rose added a dependency for D26109: Warn when 'assume_nonnull' infers nullability within an array.: D26108: Add -Wnullability-completeness-on-arrays..

Depends on D26108 too, for the tests to apply cleanly.

Nov 8 2016, 6:10 PM
jordan_rose added a dependent revision for D26108: Add -Wnullability-completeness-on-arrays.: D26227: Don't require nullability on 'va_list'..
Nov 8 2016, 6:10 PM
jordan_rose added a dependency for D26227: Don't require nullability on 'va_list'.: D26108: Add -Wnullability-completeness-on-arrays..

Depends on D26108 too (for the tests to apply cleanly).

Nov 8 2016, 6:10 PM
jordan_rose added a comment to D26108: Add -Wnullability-completeness-on-arrays..

It works fine for me, though note the "depends on D25850". The other patches in the series do seem to have been thrown off by D26226 landing first, though, so I'll update those.

Nov 8 2016, 5:59 PM

Nov 7 2016

jordan_rose closed D26319: Disallow StringRef assignment from temporary std::strings..

Committed in r286139, accidentally omitting the tests. Tests added in r286141.

Nov 7 2016, 12:50 PM
jordan_rose accepted D26319: Disallow StringRef assignment from temporary std::strings..

Approved by Duncan via email.

Nov 7 2016, 12:46 PM
jordan_rose closed D26317: Fix use-of-temporary with StringRef in code coverage.

Committed as r286122.

Nov 7 2016, 9:37 AM

Nov 4 2016

jordan_rose added a reviewer for D26319: Disallow StringRef assignment from temporary std::strings.: dexonsmith.
Nov 4 2016, 5:18 PM
jordan_rose added a dependent revision for D26317: Fix use-of-temporary with StringRef in code coverage: D26319: Disallow StringRef assignment from temporary std::strings..
Nov 4 2016, 5:18 PM
jordan_rose retitled D26317: Fix use-of-temporary with StringRef in code coverage from to Fix use-of-temporary with StringRef in code coverage.
Nov 4 2016, 4:59 PM

Nov 2 2016

jordan_rose closed D26226: Don't require nullability on template parameters in typedefs..

Committed in r285856.

Nov 2 2016, 1:56 PM

Nov 1 2016

jordan_rose added reviewers for D26227: Don't require nullability on 'va_list'.: doug.gregor, rsmith.
Nov 1 2016, 6:30 PM
jordan_rose added a dependent revision for D25850: Accept nullability annotations (_Nullable) on array parameters: D26227: Don't require nullability on 'va_list'..
Nov 1 2016, 6:30 PM
jordan_rose added a dependent revision for D26226: Don't require nullability on template parameters in typedefs.: D26227: Don't require nullability on 'va_list'..
Nov 1 2016, 6:30 PM
jordan_rose retitled D26226: Don't require nullability on template parameters in typedefs. from to Don't require nullability on template parameters in typedefs..
Nov 1 2016, 6:27 PM
jordan_rose added a comment to D26219: Bitcode: Change reader interface to take memory buffers..

Sounds great to me, but it would be kind of nice to do it in two commits: one to add the new API and one to remove the old one.

Nov 1 2016, 3:59 PM

Oct 28 2016

jordan_rose added a reviewer for D26109: Warn when 'assume_nonnull' infers nullability within an array.: doug.gregor.
Oct 28 2016, 5:45 PM
jordan_rose added a dependent revision for D25850: Accept nullability annotations (_Nullable) on array parameters: D26109: Warn when 'assume_nonnull' infers nullability within an array..
Oct 28 2016, 5:45 PM
jordan_rose added a reviewer for D26108: Add -Wnullability-completeness-on-arrays.: doug.gregor.
Oct 28 2016, 5:44 PM

Oct 27 2016

jordan_rose added a comment to D26016: BitcodeReader: Require clients to read the block info block at most once..

That's how I read it too, but that's reasonable behavior if you have several data blobs of the same kind. Your file can be { MAGIC_NUMBER, { BLOCKINFO, DATA, DATA }, { BLOCKINFO, DATA, DATA }, …} instead of { MAGIC_NUMBER, BLOCKINFO, { DATA, DATA }, { DATA, DATA}, …}. Why is the former better for concatenation than the latter? Because you don't need to know the size of the BLOCKINFO.

Oct 27 2016, 1:37 PM
jordan_rose added a comment to D26016: BitcodeReader: Require clients to read the block info block at most once..

I don't feel too strongly about this, but I do think it's an important point to understand.

Oct 27 2016, 1:23 PM
jordan_rose added a comment to D26016: BitcodeReader: Require clients to read the block info block at most once..

BLOCKINFO is part of the container format, not part of an LLVM module. I'm objecting to the idea that it's better to write one BLOCKINFO for several "data blobs" (LLVM module, Swift module, whatever) than to just stick the BLOCKINFO in each data blob and be done with it. Right now we have a completely concatenative format, as long as you separate out the fixed-size magic number at the start; this would break that.

Oct 27 2016, 1:23 PM
jordan_rose added a comment to D26016: BitcodeReader: Require clients to read the block info block at most once..

I meant homogeneous data. With heterogeneous data you just couldn't reuse the cursor either way.

Oct 27 2016, 1:08 PM
jordan_rose added a comment to D26016: BitcodeReader: Require clients to read the block info block at most once..

To elaborate a bit, the bitstream format is used to hold things other than LLVM bitcode (we currently use it for Swift "module" interface files—thanks for CCing me), and changing the requirements of that format seems unfortunate. (I do understand that it was never promised to be a stable container format.)

Oct 27 2016, 12:57 PM
jordan_rose added a comment to D26016: BitcodeReader: Require clients to read the block info block at most once..

This seems a little unfortunate, because it means that you can't just serialize two full blobs of data one after the other and reuse the same cursor on both of them.

Oct 27 2016, 12:54 PM

Oct 26 2016

jordan_rose added a comment to D25993: [Objective-C] Add objc_subclassing_restricted attribute.

Looks good from my end.

Oct 26 2016, 10:02 AM

Oct 25 2016

jordan_rose added a comment to D25850: Accept nullability annotations (_Nullable) on array parameters.

Oops. Ignore the API notes file, which is only in Swift's branch of Clang right now.

Oct 25 2016, 5:36 PM
jordan_rose updated the diff for D25850: Accept nullability annotations (_Nullable) on array parameters.

Updated based on feedback from Richard, fixed typedef support, added tests.

Oct 25 2016, 5:07 PM
jordan_rose added a comment to D25850: Accept nullability annotations (_Nullable) on array parameters.

_Nonnull in this position seems very similar to static (which typically also implies non-nullness).

Oct 25 2016, 4:07 PM
jordan_rose added inline comments to D25850: Accept nullability annotations (_Nullable) on array parameters.
Oct 25 2016, 11:41 AM

Oct 20 2016

jordan_rose updated D25850: Accept nullability annotations (_Nullable) on array parameters.
Oct 20 2016, 5:13 PM
jordan_rose retitled D25850: Accept nullability annotations (_Nullable) on array parameters from to [WIP] Accept nullability annotations (_Nullable) on array parameters.
Oct 20 2016, 5:09 PM

Oct 10 2016

jordan_rose closed D25446: Disallow ArrayRef assignment from temporaries.
Oct 10 2016, 2:07 PM
jordan_rose accepted D25446: Disallow ArrayRef assignment from temporaries.

Approved by Duncan via email. Committed as r283798.

Oct 10 2016, 2:06 PM
jordan_rose added a comment to D25446: Disallow ArrayRef assignment from temporaries.

LGTM (with or without the name) if you add std::is_assignable-based static_asserts to unittests/ADT/ArrayRefTest.cpp.

Oct 10 2016, 1:48 PM
jordan_rose added a reviewer for D25446: Disallow ArrayRef assignment from temporaries: stephant.
Oct 10 2016, 12:38 PM
jordan_rose added a comment to D25446: Disallow ArrayRef assignment from temporaries.
Oct 10 2016, 12:37 PM
jordan_rose added a comment to D25446: Disallow ArrayRef assignment from temporaries.

FWIW, LLVM, Clang, and clang-tools-extra had no issues in this area. Swift had a few. (Unfortunately Swift also has had a few with the constructor involved, so it would be great if we could figure out how to make that safer as well.)

Oct 10 2016, 12:36 PM
jordan_rose updated D25446: Disallow ArrayRef assignment from temporaries.
Oct 10 2016, 12:35 PM
jordan_rose retitled D25446: Disallow ArrayRef assignment from temporaries from to Disallow ArrayRef assignment from temporaries.
Oct 10 2016, 12:35 PM

Jun 27 2016

jordan_rose closed D21111: Avoid accessing an invalid PresumedLoc.

Committed as r273976.

Jun 27 2016, 6:09 PM

Jun 22 2016

jordan_rose added a comment to D21111: Avoid accessing an invalid PresumedLoc.

I'm happy to have someone else review this (or "LGTM" this), but it's so small that I'm not sure who else to ask. I'd rather not just commit it cause it's been a while since I've touched Clang.

Jun 22 2016, 1:20 PM

Jun 7 2016

jordan_rose updated D21111: Avoid accessing an invalid PresumedLoc.
Jun 7 2016, 5:49 PM
jordan_rose retitled D21111: Avoid accessing an invalid PresumedLoc from to Avoid accessing an invalid PresumedLoc.
Jun 7 2016, 5:48 PM