modocache (Brian Gesiak)
User

Projects

User does not belong to any projects.

User Details

User Since
Feb 16 2014, 1:10 AM (187 w, 6 d)

Recent Activity

Tue, Sep 19

modocache accepted D38033: cmake: fix .git-dir detection with using git-rev-parse (regression since 4.0).

Thanks for the improvement!

Tue, Sep 19, 11:14 AM · Restricted Project

Wed, Sep 13

modocache committed rL313164: [CFG] Fix typo in docblock: blocsk/blocks.
[CFG] Fix typo in docblock: blocsk/blocks
Wed, Sep 13, 11:03 AM

Wed, Aug 30

modocache committed rL312164: [ARM] Use Swift error registers on non-Darwin targets.
[ARM] Use Swift error registers on non-Darwin targets
Wed, Aug 30, 1:05 PM
modocache closed D35835: [ARM] Use Swift error registers on non-Darwin targets.
Wed, Aug 30, 1:05 PM
modocache updated the diff for D35835: [ARM] Use Swift error registers on non-Darwin targets.

Added some tests to test/CodeGen/ARM/swifterror.ll, also courtesy of John Holdsworth.

Wed, Aug 30, 11:43 AM

Aug 23 2017

modocache updated the diff for D36847: [Support] Add reentrant start/stop Timer methods.

Replace Running member with StartCount.

Aug 23 2017, 8:09 AM

Aug 22 2017

modocache updated the diff for D36848: [CodeGen] Use reentrant methods to time IR gen.

Use startReentrantTimer and stopReentrantTimer

Aug 22 2017, 9:33 AM
modocache added a comment to D36847: [Support] Add reentrant start/stop Timer methods.

One downside of separate startReentrantTimer and stopReentrantTimer methods is that we may need to add a bool IsReentrant parameter to the initializer of TimeRegion, in order to support the way it's used in https://reviews.llvm.org/D36946.

Aug 22 2017, 8:51 AM
modocache updated the diff for D36847: [Support] Add reentrant start/stop Timer methods.

Instead of using virtual methods, add Timer::startReentrantTimer and Timer::stopReentrantTimer. @MatzeB, let me know if this isn't what you had in mind -- thanks!

Aug 22 2017, 8:40 AM
modocache planned changes to D36847: [Support] Add reentrant start/stop Timer methods.

Thanks for the suggestions, @MatzeB! I'll try to get rid of the added vtable.

Aug 22 2017, 7:22 AM

Aug 21 2017

modocache added a comment to D36847: [Support] Add reentrant start/stop Timer methods.

Friendly ping -- does anyone have any thoughts here?

Aug 21 2017, 9:53 AM

Aug 19 2017

modocache added inline comments to D36492: [time-report] Add preprocessor timer.
Aug 19 2017, 1:15 PM

Aug 18 2017

modocache committed rL311168: [Lexicon] Add "GEP".
[Lexicon] Add "GEP"
Aug 18 2017, 8:38 AM

Aug 17 2017

modocache created D36848: [CodeGen] Use reentrant methods to time IR gen.
Aug 17 2017, 2:00 PM
modocache created D36847: [Support] Add reentrant start/stop Timer methods.
Aug 17 2017, 1:58 PM

Aug 16 2017

modocache planned changes to D36492: [time-report] Add preprocessor timer.
Aug 16 2017, 2:02 PM
modocache added a comment to D36492: [time-report] Add preprocessor timer.

Friendly ping! I think this is ready to be reviewed. It adds an additional row, Preprocessing, to the Miscellaneous Ungrouped Timers section of the clang -ftime-report output:

Aug 16 2017, 11:57 AM

Aug 15 2017

modocache updated subscribers of D36492: [time-report] Add preprocessor timer.

Oops, sorry, didn't mean to remove the subscribers. arc diff --verbatim strikes again.

Aug 15 2017, 9:57 AM
modocache updated the diff for D36492: [time-report] Add preprocessor timer.

Add PreprocessorOptions::getTimer, and move the timer to the top of Preprocessor::Lex().

Aug 15 2017, 9:54 AM

Aug 14 2017

modocache committed rL310838: [Parse] Fix typo in header docs (NFC).
[Parse] Fix typo in header docs (NFC)
Aug 14 2017, 7:30 AM

Aug 13 2017

modocache committed rL310810: [opt-viewer] Listify `dict_items` for Py3 indexing.
[opt-viewer] Listify `dict_items` for Py3 indexing
Aug 13 2017, 9:20 PM
modocache closed D36630: [opt-viewer] Listify `dict_items` for Py3 indexing.
Aug 13 2017, 9:20 PM
modocache added a comment to D36630: [opt-viewer] Listify `dict_items` for Py3 indexing.

Great, thanks!

Aug 13 2017, 9:20 PM

Aug 11 2017

modocache created D36630: [opt-viewer] Listify `dict_items` for Py3 indexing.
Aug 11 2017, 11:47 AM
modocache committed rL310741: [opt-viewer] Decode HTML bytes for Python 3.
[opt-viewer] Decode HTML bytes for Python 3
Aug 11 2017, 11:07 AM
modocache closed D36624: [opt-viewer] Decode HTML bytes for Python 3.
Aug 11 2017, 11:07 AM
modocache added a comment to D36624: [opt-viewer] Decode HTML bytes for Python 3.

I double-checked this diff worked when run on a directory containing a single YAML file, using Python 2, and it appeared to work. Let me know if it breaks anything for any of your workflows!

Aug 11 2017, 11:05 AM
modocache committed rL310740: [opt-viewer] Use Python 3-compatible iteritems.
[opt-viewer] Use Python 3-compatible iteritems
Aug 11 2017, 11:03 AM
modocache closed D36623: [opt-viewer] Use Python 3-compatible iteritems.
Aug 11 2017, 11:02 AM
modocache added a comment to D36623: [opt-viewer] Use Python 3-compatible iteritems.

Hey, no worries! I almost never catch Python 2/3 incompatibilities ahead of time. As I mentioned on PR34129, it'd be great to get some automated testing in place that could catch Python 2/3 errors as well, but until then I'm happy to fix these as I notice them.

Aug 11 2017, 11:02 AM
modocache committed rL310739: [opt-viewer] Use Python 3-compatible `intern()`.
[opt-viewer] Use Python 3-compatible `intern()`
Aug 11 2017, 10:59 AM
modocache closed D36622: [opt-viewer] Use Python 3-compatible `intern()`.
Aug 11 2017, 10:59 AM
modocache added a comment to D36622: [opt-viewer] Use Python 3-compatible `intern()`.

Yup, I've confirmed this still works with Python 2. Thanks, @anemet!

Aug 11 2017, 10:59 AM
modocache created D36624: [opt-viewer] Decode HTML bytes for Python 3.
Aug 11 2017, 10:42 AM
modocache created D36623: [opt-viewer] Use Python 3-compatible iteritems.
Aug 11 2017, 10:21 AM
modocache created D36622: [opt-viewer] Use Python 3-compatible `intern()`.
Aug 11 2017, 10:15 AM

Aug 10 2017

modocache committed rL310627: [Parse] Document Parser::SkipFunctionBodies.
[Parse] Document Parser::SkipFunctionBodies
Aug 10 2017, 10:03 AM
modocache closed D36531: [Parse] Document Parser::SkipFunctionBodies by committing rL310627: [Parse] Document Parser::SkipFunctionBodies.
Aug 10 2017, 10:03 AM
modocache committed rL310609: [Parse] Document PrintStats, SkipFunctionBodies.
[Parse] Document PrintStats, SkipFunctionBodies
Aug 10 2017, 8:31 AM
modocache closed D36530: [Parse] Document PrintStats, SkipFunctionBodies by committing rL310609: [Parse] Document PrintStats, SkipFunctionBodies.
Aug 10 2017, 8:31 AM
modocache added a comment to D36530: [Parse] Document PrintStats, SkipFunctionBodies.

Oops, sorry. I couldn't find anyone recent in the commit history. I hope no one minds if I just go ahead and commit this.

Aug 10 2017, 8:22 AM
modocache planned changes to D36492: [time-report] Add preprocessor timer.

Thanks for the feedback, @vsk, I really appreciate it! I have some other work done for this on my local checkout, but I was going a little bonkers working on it without knowing whether people would want it merged or not. I'll update this with your feedback and upload the rest for review as well.

Aug 10 2017, 7:44 AM

Aug 9 2017

modocache created D36531: [Parse] Document Parser::SkipFunctionBodies.
Aug 9 2017, 10:12 AM
modocache created D36530: [Parse] Document PrintStats, SkipFunctionBodies.
Aug 9 2017, 10:05 AM

Aug 8 2017

modocache created D36492: [time-report] Add preprocessor timer.
Aug 8 2017, 3:25 PM

Aug 3 2017

modocache added a comment to D36026: [lit] Avoid copying llvm/utils/lit/tests/Inputs with lit site configs.

Thanks!

Aug 3 2017, 10:40 AM
modocache accepted D36263: [lit] check-lit compatibility with multi-config CMake generators.
Aug 3 2017, 8:06 AM
modocache abandoned D36212: [lit] Use Visual Studio build config when testing.

Abandoning in favor of D36212. Thanks @gbedwell!

Aug 3 2017, 8:06 AM

Aug 2 2017

modocache added a comment to D36026: [lit] Avoid copying llvm/utils/lit/tests/Inputs with lit site configs.

@rnk After this patch, the check-lit target leaves untracked files in the source tree:

Aug 2 2017, 9:05 AM
modocache added a comment to D36212: [lit] Use Visual Studio build config when testing.

Yup, I've tried reverting D36026 locally, and after doing so check-lit no longer dirties my source tree. I'll comment there with next steps. IMHO reverting and then applying the changes you suggested might be a good way to go!

Aug 2 2017, 9:02 AM
modocache added a comment to D36212: [lit] Use Visual Studio build config when testing.

I'm happy either way! I was going to put the patch up for review but I noticed that after running some lit tests there appear to be some artifacts being created and left over outside of the build directory.

Aug 2 2017, 8:57 AM
modocache added a comment to D36212: [lit] Use Visual Studio build config when testing.

Apologies for jumping in on your patch here.

Aug 2 2017, 6:38 AM

Aug 1 2017

modocache updated subscribers of D36212: [lit] Use Visual Studio build config when testing.
Aug 1 2017, 11:53 PM
modocache created D36212: [lit] Use Visual Studio build config when testing.
Aug 1 2017, 11:30 PM

Jul 29 2017

modocache accepted D36026: [lit] Avoid copying llvm/utils/lit/tests/Inputs with lit site configs.

LGTM, thanks!

Jul 29 2017, 10:49 AM

Jul 27 2017

modocache added a comment to D35880: Un-revert "Teach the CMake build system to run lit's test suite. These can be run".

This Windows buildbot failure could potentially be related: http://lab.llvm.org:8011/builders/llvm-clang-x86_64-expensive-checks-win/builds/4046. The build failed due to a timeout. That being said, the bot *is* named "expensive-checks", so maybe a timeout is to be expected...? I'll keep monitoring builds for now.

Jul 27 2017, 2:44 PM
modocache accepted D35909: [lit] Port googletest lit tests to Windows.
Jul 27 2017, 12:33 PM
modocache added a comment to D35909: [lit] Port googletest lit tests to Windows.

Yup! Sorry for the wait. This works for me! :)

Jul 27 2017, 12:33 PM
modocache added a comment to D27746: [lit] Fix TestRunner unit test on Windows.

Thanks!

Jul 27 2017, 12:28 PM
modocache committed rL309312: [lit] Fix TestRunner unit test on Windows.
[lit] Fix TestRunner unit test on Windows
Jul 27 2017, 12:28 PM
modocache closed D27746: [lit] Fix TestRunner unit test on Windows.
Jul 27 2017, 12:27 PM
modocache added a reverting commit for rL257268: Revert "Teach the CMake build system to run lit's test suite. These can be run": rL309310: Un-revert "Teach the CMake build system to run lit's test suite. These can be….
Jul 27 2017, 12:21 PM
modocache committed rL309310: Un-revert "Teach the CMake build system to run lit's test suite. These can be….
Un-revert "Teach the CMake build system to run lit's test suite. These can be…
Jul 27 2017, 12:21 PM
modocache closed D35880: Un-revert "Teach the CMake build system to run lit's test suite. These can be run".
Jul 27 2017, 12:21 PM
modocache added a comment to D35880: Un-revert "Teach the CMake build system to run lit's test suite. These can be run".

I am definitely a fan of turning these on first, then optimizing to reduce the time it takes for them to run. For the record, with this patch applied locally, check-all takes 681.54 real, 3349.92 user, 614.90 sys, whereas before it took 667.35 real, 3293.70 user, 600.09 sys. I hope that's acceptable price to pay for keeping lit's own tests passing, at least until we can optimize it to run faster.

Jul 27 2017, 12:17 PM
modocache committed rL309292: [lit] Fix order of checks in shtest-shell.py test.
[lit] Fix order of checks in shtest-shell.py test
Jul 27 2017, 9:51 AM
modocache added a comment to D35947: [lit] Fix order of checks in shtest-shell.py test.

Thanks!

Jul 27 2017, 9:51 AM
modocache closed D35947: [lit] Fix order of checks in shtest-shell.py test.
Jul 27 2017, 9:51 AM
modocache updated the diff for D35947: [lit] Fix order of checks in shtest-shell.py test.

Update commit message

Jul 27 2017, 9:22 AM
modocache added a comment to D35880: Un-revert "Teach the CMake build system to run lit's test suite. These can be run".

I sent up D35947 to fix the new failures. Then I think I'll try to land this and get lit tests running continuously, unless anyone has any objections. :)

Jul 27 2017, 9:21 AM
modocache created D35947: [lit] Fix order of checks in shtest-shell.py test.
Jul 27 2017, 9:18 AM
modocache added a comment to D35880: Un-revert "Teach the CMake build system to run lit's test suite. These can be run".

Hmm, when run locally on my macOS, max-failures.py and shtest-shell.py are now failing, so those will definitely need to be fixed before this is landed. I'm on rL309282.

Jul 27 2017, 9:05 AM
modocache added a reviewer for D35880: Un-revert "Teach the CMake build system to run lit's test suite. These can be run": rnk.
Jul 27 2017, 7:56 AM
modocache updated subscribers of D35880: Un-revert "Teach the CMake build system to run lit's test suite. These can be run".

An easy way to reproduce:

./utils/lit/lit.py -v utils/lit/tests/max-failures.py utils/lit/tests/shtest-shell.py
Jul 27 2017, 7:56 AM
modocache added a reviewer for D27893: [lit] Fix shtest-format test on Windows: rnk.
Jul 27 2017, 7:52 AM
modocache added a reviewer for D27746: [lit] Fix TestRunner unit test on Windows: rnk.
Jul 27 2017, 7:51 AM
modocache abandoned D27908: [lit] Fix shtest-output-printing test on Windows.

The equivalent of this change was committed by @rnk in rL309198.

Jul 27 2017, 7:51 AM
modocache abandoned D27747: [lit] Fix shtest-shell tests on Windows.

The equivalent of this change was committed by @rnk in rL309200. I slightly prefer this change because the version of Python that is currently running is used to execute the shell tests, whereas rL309200 may use Python 3 to run the tests, but shell out to Python 2.

Jul 27 2017, 7:51 AM
modocache abandoned D27701: [lit] Fix discovery test on Windows.

The equivalent of this change was committed by @rnk in rL309194. I sort of prefer this because it verifies the same separator is used throughout, instead of accepting either. But either way works for me!

Jul 27 2017, 7:49 AM

Jul 26 2017

modocache requested changes to D35909: [lit] Port googletest lit tests to Windows.

LGTM, but I want to double-check this passes for me locally. Also, I think you may have left some debugging code in here.

Jul 26 2017, 12:19 PM
modocache added a comment to D35744: Make new PM honor -fdebug-info-for-profiling.

Just FYI, test/Other/new-pm-pgo.ll appears to be failing on several of the buildbots. One example: http://lab.llvm.org:8011/builders/clang-hexagon-elf/builds/10801

Jul 26 2017, 8:38 AM
modocache committed rL309123: [lit] Mark several of lit's tests XFAIL on Windows.
[lit] Mark several of lit's tests XFAIL on Windows
Jul 26 2017, 8:11 AM
modocache added a reverting commit for rL257221: Teach the CMake build system to run lit's test suite. These can be run: rL309123: [lit] Mark several of lit's tests XFAIL on Windows.
Jul 26 2017, 8:11 AM
modocache closed D35879: [lit] Mark several of lit's tests XFAIL on Windows.
Jul 26 2017, 8:11 AM
modocache committed rL309122: [lit] Fix type error for parallelism groups.
[lit] Fix type error for parallelism groups
Jul 26 2017, 8:03 AM
modocache added a comment to D35878: [lit] Fix type error for parallelism groups.

Thanks for all the reviews, @mgorny! @rnk, let me know if this is OK by you when you can.

Jul 26 2017, 8:03 AM
modocache closed D35878: [lit] Fix type error for parallelism groups.
Jul 26 2017, 8:03 AM
modocache added a comment to D35877: Revert "[lit] Remove dead code not referenced in the LLVM SVN repo.".

Great, thanks @mgorny!

Jul 26 2017, 8:00 AM
modocache committed rL309120: Revert "[lit] Remove dead code not referenced in the LLVM SVN repo.".
Revert "[lit] Remove dead code not referenced in the LLVM SVN repo."
Jul 26 2017, 8:00 AM
modocache added a reverting commit for rL306623: [lit] Remove dead code not referenced in the LLVM SVN repo.: rL309120: Revert "[lit] Remove dead code not referenced in the LLVM SVN repo.".
Jul 26 2017, 8:00 AM
modocache closed D35877: Revert "[lit] Remove dead code not referenced in the LLVM SVN repo.".
Jul 26 2017, 8:00 AM
modocache updated the diff for D35880: Un-revert "Teach the CMake build system to run lit's test suite. These can be run".

Use single ticks, not double

Jul 26 2017, 7:22 AM
modocache added a comment to D35880: Un-revert "Teach the CMake build system to run lit's test suite. These can be run".

That race condition may indeed exist, in rL257268 @delcypher mentions one reason for the revert is "the lit per test timeout tests fail." Perhaps related? Do you have more information about when the test fails?

Jul 26 2017, 7:17 AM

Jul 25 2017

modocache added a comment to D35857: [lit] Fix UnboundLocalError for invalid shtest redirects.

OK, @mgorny, I sent up diffs for the above! I wrote down the details as a comment on https://bugs.llvm.org/show_bug.cgi?id=33704. Thanks!

Jul 25 2017, 10:36 PM
modocache created D35880: Un-revert "Teach the CMake build system to run lit's test suite. These can be run".
Jul 25 2017, 10:31 PM
modocache abandoned D24995: [cmake][lit] Add check-lit CMake target.

Abandoning this in favor of D35880, which un-reverts @delcypher's previous work.

Jul 25 2017, 10:31 PM
modocache added a comment to D34792: [lit] Remove dead code not referenced in the LLVM SVN repo..

I submitted a revert for review: https://reviews.llvm.org/D35877. As described in my comments in https://bugs.llvm.org/show_bug.cgi?id=33704, the diff I submitted is one of three that should get the lit test suite passing again -- reviews greatly appreciated! :)

Jul 25 2017, 10:26 PM
modocache added a comment to D35879: [lit] Mark several of lit's tests XFAIL on Windows.

The change to utils/lit/tests/shtest-encoding.py is shown in Phabricator as a change to a binary file because that test file contains a non-UTF8 character. Like the other files, I just added the following lines:

Jul 25 2017, 10:21 PM
modocache created D35879: [lit] Mark several of lit's tests XFAIL on Windows.
Jul 25 2017, 10:19 PM
modocache created D35878: [lit] Fix type error for parallelism groups.
Jul 25 2017, 9:17 PM