Page MenuHomePhabricator

ldionne (Louis Dionne)
User

Projects

User does not belong to any projects.

User Details

User Since
Feb 11 2015, 3:26 PM (219 w, 2 h)

Recent Activity

Today

ldionne committed rPSTL359134: [pstl] Make the default backend be the serial backend and always provide….
[pstl] Make the default backend be the serial backend and always provide…
Wed, Apr 24, 1:28 PM
ldionne committed rG65a422c81c99: [pstl] Make the default backend be the serial backend and always provide… (authored by ldionne).
[pstl] Make the default backend be the serial backend and always provide…
Wed, Apr 24, 1:15 PM
ldionne committed rL359134: [pstl] Make the default backend be the serial backend and always provide….
[pstl] Make the default backend be the serial backend and always provide…
Wed, Apr 24, 1:10 PM
ldionne closed D59792: [pstl] Make the default backend be the serial backend and always provide parallel policies.
Wed, Apr 24, 1:10 PM · Restricted Project, Restricted Project

Yesterday

ldionne accepted D61039: [libcxx] Use relative path for libc++ library when generating script.
Tue, Apr 23, 2:41 PM
ldionne added a comment to D61032: [clang] Avoid defining duplicate header search paths for libc++ on Darwin.

I will build a significant codebase with the change and report on the results, since we lack good tests for the removal of /usr/include/c++/v1.

Tue, Apr 23, 11:29 AM · Restricted Project
ldionne created D61032: [clang] Avoid defining duplicate header search paths for libc++ on Darwin.
Tue, Apr 23, 11:24 AM · Restricted Project
ldionne added a comment to D59792: [pstl] Make the default backend be the serial backend and always provide parallel policies.

@MikeDvorskiy ping!

Tue, Apr 23, 9:36 AM · Restricted Project, Restricted Project
ldionne committed rGd524fe5b3a09: [libc++] Remove redundant conditionals for Apple platforms (authored by ldionne).
[libc++] Remove redundant conditionals for Apple platforms
Tue, Apr 23, 7:04 AM
ldionne committed rCXX358988: [libc++] Remove redundant conditionals for Apple platforms.
[libc++] Remove redundant conditionals for Apple platforms
Tue, Apr 23, 7:03 AM
ldionne committed rL358988: [libc++] Remove redundant conditionals for Apple platforms.
[libc++] Remove redundant conditionals for Apple platforms
Tue, Apr 23, 7:03 AM
ldionne closed D60842: [libc++] Remove redundant conditionals for Apple platforms.
Tue, Apr 23, 7:03 AM · Restricted Project, Restricted Project
ldionne accepted D60842: [libc++] Remove redundant conditionals for Apple platforms.

I'll try this out and back out if it breaks something -- I tested locally but we never know.

Tue, Apr 23, 7:02 AM · Restricted Project, Restricted Project

Mon, Apr 22

ldionne committed rCXXA358937: [libc++abi] Don't use a .sh.cpp test for uncaught_exception.
[libc++abi] Don't use a .sh.cpp test for uncaught_exception
Mon, Apr 22, 5:24 PM
ldionne committed rGc86011f5bc35: [libc++abi] Don't use a .sh.cpp test for uncaught_exception (authored by ldionne).
[libc++abi] Don't use a .sh.cpp test for uncaught_exception
Mon, Apr 22, 5:02 PM
ldionne committed rL358937: [libc++abi] Don't use a .sh.cpp test for uncaught_exception.
[libc++abi] Don't use a .sh.cpp test for uncaught_exception
Mon, Apr 22, 5:01 PM
ldionne created D60969: [libc++][CMake] Refactor how we link against system libraries.
Mon, Apr 22, 10:17 AM · Restricted Project
ldionne added inline comments to D60969: [libc++][CMake] Refactor how we link against system libraries.
Mon, Apr 22, 10:17 AM · Restricted Project
ldionne committed rUNW358896: [NFC] Fix typo in debug log.
[NFC] Fix typo in debug log
Mon, Apr 22, 8:51 AM
ldionne committed rG4256cf1b0401: [NFC] Fix typo in debug log (authored by ldionne).
[NFC] Fix typo in debug log
Mon, Apr 22, 8:39 AM
ldionne committed rL358896: [NFC] Fix typo in debug log.
[NFC] Fix typo in debug log
Mon, Apr 22, 8:38 AM

Thu, Apr 18

ldionne committed rG4664916017ef: [libc++] Link against libc++abi in the libc++abi tests (authored by ldionne).
[libc++] Link against libc++abi in the libc++abi tests
Thu, Apr 18, 1:59 PM
ldionne committed rL358712: [libc++] Link against libc++abi in the libc++abi tests.
[libc++] Link against libc++abi in the libc++abi tests
Thu, Apr 18, 1:58 PM
ldionne committed rCXX358712: [libc++] Link against libc++abi in the libc++abi tests.
[libc++] Link against libc++abi in the libc++abi tests
Thu, Apr 18, 1:58 PM
ldionne added a comment to D59792: [pstl] Make the default backend be the serial backend and always provide parallel policies.

Is there still a reason to maintain the _PSTL_USE_PAR_POLICIES configuration? If not, we should move forward with this patch. I know @rodgert has been bitten by this at least a couple of times.

Thu, Apr 18, 12:02 PM · Restricted Project, Restricted Project
ldionne updated the diff for D59792: [pstl] Make the default backend be the serial backend and always provide parallel policies.

Rebase onto master

Thu, Apr 18, 12:02 PM · Restricted Project, Restricted Project
ldionne added a comment to D60880: [libc++] Use COPYFILE_CLONE from the macOS copyfile(3) API to support APFS clones.

According to the documentation COPYFILE_CLONE will clone a symlink rather than its contents, whereas COPY_FILE_DATA will follow the link. That's why I'm checking for symlinks in https://reviews.llvm.org/D60802. I don't know whether copy_file_impl_copyfile makes any guarantees.

Thu, Apr 18, 11:33 AM · Restricted Project
ldionne committed rPSTL358700: [pstl] Add a serial backend for the PSTL.
[pstl] Add a serial backend for the PSTL
Thu, Apr 18, 11:25 AM
ldionne committed rGdcf4b9aee06a: [pstl] Add a serial backend for the PSTL (authored by ldionne).
[pstl] Add a serial backend for the PSTL
Thu, Apr 18, 11:19 AM
ldionne committed rL358700: [pstl] Add a serial backend for the PSTL.
[pstl] Add a serial backend for the PSTL
Thu, Apr 18, 11:19 AM
ldionne closed D59791: [pstl] Add a serial backend for the PSTL.
Thu, Apr 18, 11:19 AM · Restricted Project, Restricted Project
ldionne added a comment to D60880: [libc++] Use COPYFILE_CLONE from the macOS copyfile(3) API to support APFS clones.

I'm not sure this change is valid because now we're also copying things like attributes, ACL and even the stat of the file. But I wanted to have that discussion in public here for documentation.

Thu, Apr 18, 11:04 AM · Restricted Project
ldionne created D60880: [libc++] Use COPYFILE_CLONE from the macOS copyfile(3) API to support APFS clones.
Thu, Apr 18, 11:03 AM · Restricted Project
ldionne committed rCXXA358690: [libc++] Make sure we re-export some missing libc++abi symbols from libc++.
[libc++] Make sure we re-export some missing libc++abi symbols from libc++
Thu, Apr 18, 10:34 AM
ldionne committed rG549048f39006: [libc++] Make sure we re-export some missing libc++abi symbols from libc++ (authored by ldionne).
[libc++] Make sure we re-export some missing libc++abi symbols from libc++
Thu, Apr 18, 10:18 AM
ldionne committed rCXX358690: [libc++] Make sure we re-export some missing libc++abi symbols from libc++.
[libc++] Make sure we re-export some missing libc++abi symbols from libc++
Thu, Apr 18, 10:17 AM
ldionne committed rL358690: [libc++] Make sure we re-export some missing libc++abi symbols from libc++.
[libc++] Make sure we re-export some missing libc++abi symbols from libc++
Thu, Apr 18, 10:17 AM
ldionne closed D60424: [libc++] Make sure we re-export some missing libc++abi symbols from libc++.
Thu, Apr 18, 10:17 AM · Restricted Project, Restricted Project
ldionne added a comment to D59791: [pstl] Add a serial backend for the PSTL.

Is there anything to fix in this patch? If not, and if that implementation is the correct one given today's backend API, I'd like to check this in so we can do followup cleanup. It'll also help me as I try to see if a more general backend API is possible because I'll have both the TBB and the serial (trivial) examples to work off of.

Thu, Apr 18, 10:04 AM · Restricted Project, Restricted Project
ldionne added a comment to D60537: [pstl] Do not inject execution policies into namespace ::pstl in the tests.
I don't think the pstl should include headers that are useless to all but one specific shipping vehicle of the PSTL (the standalone version)

Why cannot it be lie onto high "include\pstl" (or even "pstl") level of "LLVM upstream" as a special folder, which will not be taken into libcxx and libcstd++ libraries? (BTW, and as well the other extra functionality which is not going to be released with the next libcxx/libcstd release?, for e[ample alternative back-ends, experimental features, etc)
Just ignore it "extra" folders when the code are integrating into libcxx and libcstd repositories.
What do you think here?

Thu, Apr 18, 8:43 AM · Restricted Project
ldionne committed rGebaf1d5e2b87: [libc++] Unconditionally enable the __pad_and_output optimization (authored by ldionne).
[libc++] Unconditionally enable the __pad_and_output optimization
Thu, Apr 18, 8:18 AM
ldionne committed rCXX358677: [libc++] Unconditionally enable the __pad_and_output optimization.
[libc++] Unconditionally enable the __pad_and_output optimization
Thu, Apr 18, 8:17 AM
ldionne committed rL358677: [libc++] Unconditionally enable the __pad_and_output optimization.
[libc++] Unconditionally enable the __pad_and_output optimization
Thu, Apr 18, 8:17 AM
ldionne added a comment to D60537: [pstl] Do not inject execution policies into namespace ::pstl in the tests.

Louis,

I have an issue regarding the dummy stdlib headers.
Proposed location is valid only for the test. (/test/support/stdlib)

The tests is not necessary part of standalone PSTL because is not a part of implementation of PSTL.
For example, If i have an application which use standalone pstl include "pstl/test/support/stdlib" would look strange at least.

So, I suggest moving the dummy stdlib headers into ..pstl\internal\stdlib or into ..pstl\stdlib

Thu, Apr 18, 7:58 AM · Restricted Project
ldionne committed rCXX358671: [libc++][CMake] Always provide new/delete in libc++ unless specified otherwise.
[libc++][CMake] Always provide new/delete in libc++ unless specified otherwise
Thu, Apr 18, 7:47 AM
ldionne committed rG17a75f7e55bb: [libc++][CMake] Always provide new/delete in libc++ unless specified otherwise (authored by ldionne).
[libc++][CMake] Always provide new/delete in libc++ unless specified otherwise
Thu, Apr 18, 7:47 AM
ldionne committed rL358671: [libc++][CMake] Always provide new/delete in libc++ unless specified otherwise.
[libc++][CMake] Always provide new/delete in libc++ unless specified otherwise
Thu, Apr 18, 7:46 AM
ldionne closed D60797: [libc++][CMake] Always provide new/delete in libc++ unless specified otherwise.
Thu, Apr 18, 7:46 AM · Restricted Project, Restricted Project

Wed, Apr 17

ldionne created D60842: [libc++] Remove redundant conditionals for Apple platforms.
Wed, Apr 17, 3:25 PM · Restricted Project, Restricted Project
ldionne committed rGa4bd340bf501: [libc++][CMake] Remove unnecessary conditional for defining new handlers (authored by ldionne).
[libc++][CMake] Remove unnecessary conditional for defining new handlers
Wed, Apr 17, 2:56 PM
ldionne committed rL358616: [libc++][CMake] Remove unnecessary conditional for defining new handlers.
[libc++][CMake] Remove unnecessary conditional for defining new handlers
Wed, Apr 17, 2:56 PM
ldionne committed rCXX358616: [libc++][CMake] Remove unnecessary conditional for defining new handlers.
[libc++][CMake] Remove unnecessary conditional for defining new handlers
Wed, Apr 17, 2:56 PM
ldionne accepted D60417: [libunwind] Add support for ARMv7-M architecture which uses the Thumb 2 ISA (unified syntax).

I'm relaying approval to @mstorsjo, who seems okay with the change.

Wed, Apr 17, 1:40 PM · Restricted Project
ldionne committed rG09ef420d6254: [libc++] (Take 2) Add a test that uses the debug database from multiple threads (authored by ldionne).
[libc++] (Take 2) Add a test that uses the debug database from multiple threads
Wed, Apr 17, 1:07 PM
ldionne committed rCXX358610: [libc++] (Take 2) Add a test that uses the debug database from multiple threads.
[libc++] (Take 2) Add a test that uses the debug database from multiple threads
Wed, Apr 17, 1:07 PM
ldionne committed rL358610: [libc++] (Take 2) Add a test that uses the debug database from multiple threads.
[libc++] (Take 2) Add a test that uses the debug database from multiple threads
Wed, Apr 17, 1:07 PM
ldionne committed rG7a5eae15a146: [libc++][CMake] Allow building neither the shared nor the static library (authored by ldionne).
[libc++][CMake] Allow building neither the shared nor the static library
Wed, Apr 17, 12:50 PM
ldionne committed rCXX358608: [libc++][CMake] Allow building neither the shared nor the static library.
[libc++][CMake] Allow building neither the shared nor the static library
Wed, Apr 17, 12:50 PM
ldionne committed rL358608: [libc++][CMake] Allow building neither the shared nor the static library.
[libc++][CMake] Allow building neither the shared nor the static library
Wed, Apr 17, 12:50 PM
ldionne accepted D57872: [CMake] Split linked libraries for shared and static libc++.

Love the cleanup!

Wed, Apr 17, 11:38 AM · Restricted Project
ldionne committed rG81875a67b0d1: [libc++] Use the no_destroy attribute to avoid destroying debug DB statics (authored by ldionne).
[libc++] Use the no_destroy attribute to avoid destroying debug DB statics
Wed, Apr 17, 11:21 AM
ldionne committed rL358602: [libc++] Use the no_destroy attribute to avoid destroying debug DB statics.
[libc++] Use the no_destroy attribute to avoid destroying debug DB statics
Wed, Apr 17, 11:21 AM
ldionne committed rCXX358602: [libc++] Use the no_destroy attribute to avoid destroying debug DB statics.
[libc++] Use the no_destroy attribute to avoid destroying debug DB statics
Wed, Apr 17, 11:21 AM
ldionne closed D60830: [libc++] Use the no_destroy attribute to avoid destroying debug DB statics.
Wed, Apr 17, 11:21 AM · Restricted Project
ldionne added a comment to D60830: [libc++] Use the no_destroy attribute to avoid destroying debug DB statics.

I am concerned that we're going to start flagging leaks because we don't destroy those -- do you share the concern? I mean I can also just push and see what our various sanitizers say.

Wed, Apr 17, 10:10 AM · Restricted Project
ldionne created D60830: [libc++] Use the no_destroy attribute to avoid destroying debug DB statics.
Wed, Apr 17, 10:00 AM · Restricted Project
ldionne committed rG840197e5edc2: Revert "[libc++] Add a test that uses the debug database from multiple threads" (authored by ldionne).
Revert "[libc++] Add a test that uses the debug database from multiple threads"
Wed, Apr 17, 9:42 AM
ldionne committed rL358593: Revert "[libc++] Add a test that uses the debug database from multiple threads".
Revert "[libc++] Add a test that uses the debug database from multiple threads"
Wed, Apr 17, 9:41 AM
ldionne committed rCXX358593: Revert "[libc++] Add a test that uses the debug database from multiple threads".
Revert "[libc++] Add a test that uses the debug database from multiple threads"
Wed, Apr 17, 9:41 AM
ldionne committed rG00f2ab1c2f01: [libc++] Add a test that uses the debug database from multiple threads (authored by ldionne).
[libc++] Add a test that uses the debug database from multiple threads
Wed, Apr 17, 9:25 AM
ldionne committed rL358591: [libc++] Add a test that uses the debug database from multiple threads.
[libc++] Add a test that uses the debug database from multiple threads
Wed, Apr 17, 9:25 AM
ldionne committed rCXX358591: [libc++] Add a test that uses the debug database from multiple threads.
[libc++] Add a test that uses the debug database from multiple threads
Wed, Apr 17, 9:25 AM
ldionne committed rGcb843f5b5502: [libc++][NFC] Make size of allocation more self-documenting (authored by ldionne).
[libc++][NFC] Make size of allocation more self-documenting
Wed, Apr 17, 9:11 AM
ldionne committed rCXX358588: [libc++][NFC] Make size of allocation more self-documenting.
[libc++][NFC] Make size of allocation more self-documenting
Wed, Apr 17, 9:10 AM
ldionne committed rL358588: [libc++][NFC] Make size of allocation more self-documenting.
[libc++][NFC] Make size of allocation more self-documenting
Wed, Apr 17, 9:10 AM
ldionne committed rG491ff04a49a6: Explicitly say we don't define new/delete in libc++ during Apple stage1… (authored by ldionne).
Explicitly say we don't define new/delete in libc++ during Apple stage1…
Wed, Apr 17, 7:58 AM
ldionne committed rL358582: Explicitly say we don't define new/delete in libc++ during Apple stage1….
Explicitly say we don't define new/delete in libc++ during Apple stage1…
Wed, Apr 17, 7:57 AM
ldionne committed rC358582: Explicitly say we don't define new/delete in libc++ during Apple stage1….
Explicitly say we don't define new/delete in libc++ during Apple stage1…
Wed, Apr 17, 7:57 AM

Tue, Apr 16

ldionne created D60797: [libc++][CMake] Always provide new/delete in libc++ unless specified otherwise.
Tue, Apr 16, 3:08 PM · Restricted Project, Restricted Project
ldionne accepted D60743: Allow building without threads in standalone builds.

The fact that we have to add this manually is unfortunate.

Tue, Apr 16, 2:50 PM · Restricted Project
ldionne committed rG4eff3de99423: [NFC] Build libc++ verbosely in the macOS CI (authored by ldionne).
[NFC] Build libc++ verbosely in the macOS CI
Tue, Apr 16, 2:15 PM
ldionne committed rL358529: [NFC] Build libc++ verbosely in the macOS CI.
[NFC] Build libc++ verbosely in the macOS CI
Tue, Apr 16, 2:15 PM
ldionne committed rCXX358529: [NFC] Build libc++ verbosely in the macOS CI.
[NFC] Build libc++ verbosely in the macOS CI
Tue, Apr 16, 2:15 PM
ldionne committed rG4b9c19e42d67: [libc++] Make sure we use new/delete from libc++abi on CI for Apple platforms (authored by ldionne).
[libc++] Make sure we use new/delete from libc++abi on CI for Apple platforms
Tue, Apr 16, 1:45 PM
ldionne committed rL358524: [libc++] Make sure we use new/delete from libc++abi on CI for Apple platforms.
[libc++] Make sure we use new/delete from libc++abi on CI for Apple platforms
Tue, Apr 16, 1:45 PM
ldionne committed rCXX358524: [libc++] Make sure we use new/delete from libc++abi on CI for Apple platforms.
[libc++] Make sure we use new/delete from libc++abi on CI for Apple platforms
Tue, Apr 16, 1:45 PM
ldionne committed rG3084db3bb1a0: [libc++] Remove old workaround for buildit (authored by ldionne).
[libc++] Remove old workaround for buildit
Tue, Apr 16, 12:25 PM
ldionne committed rCXX358518: [libc++] Remove old workaround for buildit.
[libc++] Remove old workaround for buildit
Tue, Apr 16, 12:25 PM
ldionne committed rL358518: [libc++] Remove old workaround for buildit.
[libc++] Remove old workaround for buildit
Tue, Apr 16, 12:25 PM
ldionne closed D60717: [libc++] Remove old workaround for buildit.
Tue, Apr 16, 12:25 PM · Restricted Project, Restricted Project
ldionne accepted D60595: [pstl] Clean up parameter uglifications.

Please test your commit access with this.

Tue, Apr 16, 11:10 AM · Restricted Project
ldionne added a comment to D54410: [libc++] Add C++17 deduction guides for std::function.

Gentle ping

Tue, Apr 16, 7:53 AM
ldionne committed rPSTL358497: [pstl] Remove the stdlib headers from the PSTL and move them to the tests.
[pstl] Remove the stdlib headers from the PSTL and move them to the tests
Tue, Apr 16, 7:48 AM
ldionne committed rG4d88b17b3f28: [pstl] Remove the stdlib headers from the PSTL and move them to the tests (authored by ldionne).
[pstl] Remove the stdlib headers from the PSTL and move them to the tests
Tue, Apr 16, 7:39 AM
ldionne committed rL358497: [pstl] Remove the stdlib headers from the PSTL and move them to the tests.
[pstl] Remove the stdlib headers from the PSTL and move them to the tests
Tue, Apr 16, 7:38 AM
ldionne closed D60535: [pstl] Remove the stdlib headers from the PSTL and move them to the tests.
Tue, Apr 16, 7:38 AM · Restricted Project, Restricted Project
ldionne added a comment to D57872: [CMake] Split linked libraries for shared and static libc++.

I think LIBCXX_STATIC_LIBRARIES and LIBCXX_SHARED_LIBRARIES would be better renamed LIBCXX_STATIC_LIBRARY_DEPENDENCIES and LIBCXX_SHARED_LIBRARY_DEPENDENCIES , or something similar. What do you think?

Tue, Apr 16, 7:27 AM · Restricted Project
ldionne added a comment to D60743: Allow building without threads in standalone builds.

Can you please add context to the diff?

Tue, Apr 16, 7:24 AM · Restricted Project
ldionne added a comment to D60750: [libc++][CMake] Clean up logic for choosing which unwinder lib to link with tests.

I'd love more FIXMEs to be resolved before we move forward.

Tue, Apr 16, 7:24 AM · Restricted Project

Mon, Apr 15

ldionne added a comment to D60626: [clang] Aligned allocation is actually supported in macosx 10.13.

Reverted in r358437 because of unforeseen CI failures. The failures I was expecting are in libc++, not in compiler-rt. I'll investigate locally and re-apply.

Mon, Apr 15, 12:12 PM · Restricted Project, Restricted Project