[PATCH] D111952: [clang] [MinGW] Guess the right ix86 arch name spelling as sysroot

mstorsjo added a comment. I guess a downside of this solution, is that if an `i686` sysroot exists next to the clang binary, it becomes practically impossible to test codegen differences when you run it with various `-target iX86-w64-mingw32` options, as they'd all be corrected back to `i686`.

[PATCH] D111579: [clang] Fix DIFile directory root on Windows

mstorsjo added inline comments. Comment at: clang/test/CodeGen/debug-prefix-map.c:24 +// CHECK-NO-MAIN-FILE-NAME: !DIFile(filename: "{{/|C:}}UNLIKELY_PATH{{/|}}empty{{/|}}{{.*}}", +// CHECK-NO-MAIN-FILE-NAME-SAME:directory: "") +// CHECK-NO-MAIN-FILE-NAME:

[PATCH] D111579: [clang] Fix DIFile directory root on Windows

mstorsjo added inline comments. Comment at: clang/lib/CodeGen/CGDebugInfo.cpp:439 +if (llvm::sys::path::root_path(DirBuf) == DirBuf) { + // Don't strip the common prefix if it is only the root ("/" or "C:\") // since that would make LLVM diagnostic locations

[PATCH] D111457: [clang][test] Add lit helper for windows paths

mstorsjo added a comment. Wouldn't this one also be solved pretty much the same, but differently, by changing `if (llvm::sys::path::is_absolute(RemappedFile)) {` into `is_absolute_gnu`? Since we're remapping debug paths, it's plausible that the target path can be a different style (when cross

[PATCH] D105169: [Clang/Test]: Rename enable_noundef_analysis to disable-noundef-analysis and turn it off by default

mstorsjo added a comment. I've bisected a crash in generated code down to this commit. The code that crashes is clean when run in ubsan. The misbehaviour happens across 4 tested architectures (i686, x86_64, armv7, aarch64). The misbehaviour occurs in this preprocessed source,

[PATCH] D105169: [Clang/Test]: Rename enable_noundef_analysis to disable-noundef-analysis and turn it off by default

mstorsjo added a comment. In D105169#3069220 , @aqjune wrote: > It seems the original code has a use of an uninitialized variable. > Line 4420 at seek-preproc.c (function `ff_seek_frame_binary`): > >int64_t pos_min=pos_min, pos_max=pos_max, pos,

[PATCH] D111952: [clang] [MinGW] Guess the right ix86 arch name spelling as sysroot

mstorsjo created this revision. mstorsjo added reviewers: phosek, MaskRay. Herald added subscribers: pengfei, kristof.beyls. mstorsjo requested review of this revision. Herald added a project: clang. For x86, most contempory mingw toolchains use i686 as 32 bit x86 arch target. As long as the

[PATCH] D105169: [Clang/Test]: Rename enable_noundef_analysis to disable-noundef-analysis and turn it off by default

mstorsjo added a comment. In D105169#3069555 , @aqjune wrote: > I see, the crash is happening at the line because SimplifyCFG removes the > `sti->index_entries` null pointer check (which seems valid to me). > If `stl->index_entries` was null, it would

[PATCH] D111081: [clang] [MinGW] Fix paths on Gentoo

mstorsjo added a comment. Looks ok to me - WDYT @mgorny? CHANGES SINCE LAST ACTION https://reviews.llvm.org/D111081/new/ https://reviews.llvm.org/D111081 ___ cfe-commits mailing list cfe-commits@lists.llvm.org

[PATCH] D111707: [clang] [Windows] Mark PIC as implicitly enabled for aarch64, just like for x86_64

mstorsjo added inline comments. Comment at: clang/test/Driver/pic.c:322 // RUN: | FileCheck %s --check-prefix=CHECK-PIC2 +// RUN: %clang -c %s -target aarch64-windows-msvc -### 2>&1 \ +// RUN: | FileCheck %s --check-prefix=CHECK-PIC2 MaskRay wrote: >

[PATCH] D111707: [clang] [Windows] Mark PIC as implicitly enabled for aarch64, just like for x86_64

This revision was automatically updated to reflect the committed changes. Closed by commit rGb541845ea082: [clang] [Windows] Mark PIC as implicitly enabled for aarch64, just like for… (authored by mstorsjo). Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION

[PATCH] D111195: [clang][Tooling] Use Windows command lines on all Windows, except Cygwin

This revision was landed with ongoing or failed builds. This revision was automatically updated to reflect the committed changes. Closed by commit rGd9b9a7f42870: [clang][Tooling] Use Windows command lines on all Windows, except Cygwin (authored by jeremyd2019, committed by mstorsjo).

[PATCH] D111195: [clang][Tooling] Use Windows command lines on all Windows, except Cygwin

mstorsjo accepted this revision. mstorsjo added a comment. This revision is now accepted and ready to land. LGTM CHANGES SINCE LAST ACTION https://reviews.llvm.org/D95/new/ https://reviews.llvm.org/D95 ___ cfe-commits mailing list

[PATCH] D111081: [clang] [MinGW] Fix paths on Gentoo

mstorsjo added a reviewer: mstorsjo. mstorsjo added a subscriber: cfe-commits. mstorsjo added a comment. (I’ll have a look and comment on the patch later.) CHANGES SINCE LAST ACTION https://reviews.llvm.org/D111081/new/ https://reviews.llvm.org/D111081

[PATCH] D111081: [clang] [MinGW] Fix paths on Gentoo

mstorsjo added a comment. The change looks ok to me, but I think it'd be good to have a testcase for finding these paths (IIRC there are a bunch of simulated sysroots under clang/test/Driver). CHANGES SINCE LAST ACTION https://reviews.llvm.org/D111081/new/ https://reviews.llvm.org/D111081

[PATCH] D111195: [clang][Tooling] Use Windows command lines on all Windows, except Cygwin

mstorsjo added a subscriber: rnk. mstorsjo added a comment. This looks ok to me I guess. Technically in this case, this is pretty much equivalent to `#ifdef _WIN32`, i.e. any form when running on windows, where the process considers itself windows (cygwin doesn't define `_WIN32`), but I guess

[PATCH] D42225: libcxx: Provide overloads for basic_filebuf::open() et al that take wchar_t* filenames on Windows.

mstorsjo added a comment. In D42225#3020688 , @CaseyCarter wrote: > In D42225#2963216 , @ldionne wrote: > >> In D42225#2963190 , @mstorsjo wrote: >> >>> In D42225#2962348

[PATCH] D111952: [clang] [MinGW] Guess the right ix86 arch name spelling as sysroot

mstorsjo updated this revision to Diff 381795. mstorsjo added a comment. Moved the fixup into `computeTargetTriple`, so it's only done after applying the `-m32`/`-m64` options. Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D111952/new/

[PATCH] D111952: [clang] [MinGW] Guess the right ix86 arch name spelling as sysroot

mstorsjo added a comment. In D111952#3068680 , @mstorsjo wrote: > I guess a downside of this solution, is that if an `i686` sysroot exists next > to the clang binary, it becomes practically impossible to test codegen > differences when you run it with

[PATCH] D112387: [clang] [MinGW] Rename the 'Arch' member to 'SysrootName'. NFC.

mstorsjo created this revision. mstorsjo added a reviewer: MaskRay. mstorsjo requested review of this revision. Herald added a project: clang. This string isn't a plain architecture name, but contains the whole subdir name used for the sysroot, which often is equal to the target triple.

[PATCH] D114651: [clang-cl] Expose -Wall to clang-cl by unaliasing -Wall, keeping /Wall as alias to -Weverything

mstorsjo added a comment. In D114651#3158807 , @rnk wrote: > I will add that multiple users have run into this problem, and I think it > might be more practical to consider unaliasing Wall altogether. Clang doesn't > emit the same set of warnings as

[PATCH] D113254: [clang] Fix a misadjusted path style comparison in a unittest

mstorsjo added a comment. Ping @dexonsmith Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D113254/new/ https://reviews.llvm.org/D113254 ___ cfe-commits mailing list cfe-commits@lists.llvm.org

[PATCH] D114064: [clang] [MinGW] Pass --no-demangle through to the mingw linker

mstorsjo added a reviewer: MaskRay. mstorsjo added a comment. Adding more potential reviewers. For context, see https://github.com/llvm/llvm-project/blob/llvmorg-14-init/clang/lib/Driver/Driver.cpp#L335-L353 where the original option is filtered out. Repository: rG LLVM Github Monorepo

[PATCH] D100810: [llvm] Use `GNUInstallDirs` to support custom installation dirs

mstorsjo added inline comments. Comment at: llvm/CMakeLists.txt:75 set(LLVM_ENABLE_PROJECTS "" CACHE STRING - "Semicolon-separated list of projects to build (${LLVM_KNOWN_PROJECTS}), or \"all\".") + "Semicolon-separated list of projects to build

[PATCH] D114064: [clang] [MinGW] Pass --no-demangle through to the mingw linker

This revision was landed with ongoing or failed builds. This revision was automatically updated to reflect the committed changes. Closed by commit rGa8877c5ccc0e: [clang] [MinGW] Pass --no-demangle through to the mingw linker (authored by mstorsjo). Repository: rG LLVM Github Monorepo CHANGES

[PATCH] D116503: [clang] Add arguments for silencing unused argument warnings for some but not all arguments

mstorsjo created this revision. mstorsjo added reviewers: aaron.ballman, sepavloff, phosek, MaskRay. Herald added a subscriber: dang. mstorsjo requested review of this revision. Herald added a project: clang. When passing a set of flags to configure defaults for a specific target (similar to the

[PATCH] D116503: [clang] Add arguments for silencing unused argument warnings for some but not all arguments

mstorsjo updated this revision to Diff 397476. mstorsjo added a comment. Regenerated ClangCommandLineReference.rst (with the relevant changes), removed NoXarchOption. Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D116503/new/

[PATCH] D116503: [clang] Add arguments for silencing unused argument warnings for some but not all arguments

mstorsjo added a comment. In D116503#3221083 , @MaskRay wrote: > A more searchable subject (mentioning the new option in the subject is useful > for archaeology) may be `Add > --start-no-unused-arguments/--end-no-unused-arguments to silence some unused

[PATCH] D116020: [clang][#52782] Bail on incomplete parameter type in stdcall name mangling

mstorsjo accepted this revision. mstorsjo added a comment. This revision is now accepted and ready to land. LGTM, the change seems sensible to me. But I’d prefer if you could hold off pushing it for another couple days, as others who might want to comment might not be following during the

[PATCH] D114064: [clang] [MinGW] Pass --no-demangle through to the mingw linker

mstorsjo added a comment. In D114064#3216147 , @Ericson2314 wrote: > Not sure what the exact division of labor is between the GNU and MinGW > backends, but assuming there is no way to share this between them the patch > looks good. Yeah. Also at this

[PATCH] D116485: [clang][MinGW] Explicitly ignore `-fPIC` & friends

mstorsjo accepted this revision. mstorsjo added a comment. This revision is now accepted and ready to land. LGTM, thanks. The fact that GCC used to emit a non-fatal warning for this option, but no longer does, probably is the key deciding factor here. Repository: rG LLVM Github Monorepo

[PATCH] D112430: [ARM][libunwind] add PACBTI-M support for libunwind

mstorsjo added a comment. Looks overall ok, without diving in very deep, except for this one change that seemed unrelated. Comment at: libunwind/src/Unwind-EHABI.cpp:312 uint8_t registers = getByte(data, offset++); - if (registers & 0xf0 ||

[PATCH] D109856: [libunwind][ARM] Handle end of stack during unwind

mstorsjo accepted this revision. mstorsjo added a comment. This revision is now accepted and ready to land. In D109856#3155235 , @danielkiss wrote: > In D109856#3154763 , @mstorsjo > wrote: > >> Looks

[PATCH] D114576: [PR52549][clang-cl] Predefine _MSVC_EXECUTION_CHARACTER_SET

mstorsjo added a comment. Would it be possible to move setting of this flag to somewhere else (e.g. somewhere in Driver/ToolChains/MSVC.cpp), so that it would be picked up also when building with the gcc-style driver with `--target=*-windows-msvc`? Overall in earlier discussions, there's been

[PATCH] D114576: [PR52549][clang-cl] Predefine _MSVC_EXECUTION_CHARACTER_SET

mstorsjo accepted this revision. mstorsjo added a comment. This revision is now accepted and ready to land. Thanks! This LGTM - but maybe wait a while if someone else has something to say. (Although I guess americans are on holiday today, so there's probably not many that will see and care

[PATCH] D114651: [clang-cl] Expose -Wall to clang-cl by unaliasing -Wall, keeping /Wall as alias to -Weverything

mstorsjo added a comment. In D114651#3156374 , @zero9178 wrote: > Is the deviation from MSVC behaviour here intentional? MSVC flags allow both > using a `/` as well as `-` as prefix. That means Both `-Wall` and `/Wall` are > accepted by MSVC as well as

[PATCH] D113268: [clangd] Fix tests with forward slash as separator on windows

mstorsjo added a comment. Sorry for taking so long to get back to you on this matter... In D113268#3112026 , @sammccall wrote: >> For someone unfamiliar with the tool, is there a simple "smoke test >> procedure" I could try out with it to kick the

[PATCH] D113268: [clangd] Fix tests with forward slash as separator on windows

mstorsjo updated this revision to Diff 388857. mstorsjo added a comment. Added a comment about the surprising change where the order of append and native is switched. Made the modification to `testRoot()` less hacky. Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION

[PATCH] D110216: [clang] retain type sugar in auto / template argument deduction

mstorsjo added a comment. This broke building ANGLE as part of Qt 5.15 for a mingw target, with the following error: ../../../3rdparty/angle/src/libANGLE/renderer/d3d/d3d11/ResourceManager11.cpp:532:38: error: explicit instantiation of 'allocate' does not refer to a function template,

[PATCH] D110216: [clang] retain type sugar in auto / template argument deduction

mstorsjo added a comment. In D110216#3130193 , @mizvekov wrote: > In D110216#3130184 , @mstorsjo > wrote: > >> Do you happen to know what's going on here? The files mentioned are >>

[PATCH] D114064: [clang] [MinGW] Pass --no-demangle through to the mingw linker

mstorsjo created this revision. mstorsjo added a reviewer: rnk. mstorsjo requested review of this revision. Herald added a project: clang. Clang has custom handling of --no-demangle, where it is removed from the input -Wl and -Xlinker options, and readded specifically by the drivers where it's

[PATCH] D112430: [ARM][libunwind] add PACBTI-M support for libunwind

mstorsjo accepted this revision. mstorsjo added a comment. This revision is now accepted and ready to land. I guess this is ok then. I can't say I've followed every bit in detail, but it looks sensible overall. Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION

[PATCH] D109856: [libunwind][ARM] Handle end of stack during unwind

mstorsjo added a comment. Looks reasonable I think. Is this a deficiency in the EHABI implementation only, i.e. this aspect works as it should in the regular dwarf implementation? CHANGES SINCE LAST ACTION https://reviews.llvm.org/D109856/new/ https://reviews.llvm.org/D109856

[PATCH] D112387: [clang] [MinGW] Rename the 'Arch' member to 'SysrootName'. NFC.

mstorsjo updated this revision to Diff 382798. mstorsjo added a comment. Renamed the variable to `SubdirName`. Will push tomorrow. Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D112387/new/ https://reviews.llvm.org/D112387 Files:

[PATCH] D111952: [clang] [MinGW] Guess the right ix86 arch name spelling as sysroot

mstorsjo updated this revision to Diff 382801. mstorsjo added a comment. Rebased after renaming `SysrootName` to `SubdirName`. Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D111952/new/ https://reviews.llvm.org/D111952 Files:

[PATCH] D112387: [clang] [MinGW] Rename the 'Arch' member to 'SysrootName'. NFC.

mstorsjo added a comment. In D112387#3090707 , @MaskRay wrote: > How is this different from `clang::driver::Driver::SysRoot`? > > From `SysrootName + "/sys-root"` looks like the `sys-root` directory is the > sysroot? Perhaps there is a better way

[PATCH] D100756: [llvm-rc] [4/4] Add a GNU windres-like frontend to llvm-rc

mstorsjo added a comment. In D100756#3088035 , @nickdesaulniers wrote: > thanks for this patch @mstorsjo ! This is help AOSP LLVM move our windows > builds of LLVM away from MinGW to LLVM! >

[PATCH] D111952: [clang] [MinGW] Guess the right ix86 arch name spelling as sysroot

mstorsjo updated this revision to Diff 382946. mstorsjo marked an inline comment as done. mstorsjo added a comment. Fixed the naming of the new function, using `--target=` in the newly added test lines, fixed a case of missed clang-format. Repository: rG LLVM Github Monorepo CHANGES SINCE

[PATCH] D112387: [clang] [MinGW] Rename the 'Arch' member to 'SysrootName'. NFC.

This revision was automatically updated to reflect the committed changes. Closed by commit rG897c86dec5af: [clang] [MinGW] Rename the Arch member to SubdirName. NFC. (authored by mstorsjo). Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D112387/new/

[PATCH] D112289: Support: Use sys::path::is_style_{posix,windows}() in a few places

mstorsjo accepted this revision. mstorsjo added a comment. This revision is now accepted and ready to land. LGTM, too. D112786 does a couple more things that do pretty much the same - I don't mind if you want to fold them into this, or keep it as-is, with the

[PATCH] D111952: [clang] [MinGW] Guess the right ix86 arch name spelling as sysroot

mstorsjo added inline comments. Comment at: clang/lib/Driver/ToolChains/MinGW.h:63 + static void FixTripleArch(const Driver , llvm::Triple , +const llvm::opt::ArgList ); MaskRay wrote: > Use `functionName` for newer functions.

[PATCH] D113269: [clang-move] Fix unit tests with forward slash as separator on windows

mstorsjo added inline comments. Comment at: clang-tools-extra/unittests/clang-move/ClangMoveTests.cpp:235 - // std::string IncludeArg = Twine("-I" + WorkingDir; + // std::string IncludeArg = Twine("-I" + Dir; tooling::runToolOnCodeWithArgs( aaron.ballman

[PATCH] D113264: [clang] [DirectoryWatcher] Remove leading \\?\ from GetFinalPathNameByHandleW

This revision was landed with ongoing or failed builds. This revision was automatically updated to reflect the committed changes. Closed by commit rG2ca6fc34fc08: [clang] [DirectoryWatcher] Remove leading \\?\ from GetFinalPathNameByHandleW (authored by mstorsjo). Changed prior to commit:

[PATCH] D113269: [clang-move] Fix unit tests with forward slash as separator on windows

This revision was landed with ongoing or failed builds. This revision was automatically updated to reflect the committed changes. Closed by commit rG98f0bf74ca6d: [clang-move] Fix unit tests with forward slash as separator on windows (authored by mstorsjo). Changed prior to commit:

[PATCH] D113268: [clangd] Fix tests with forward slash as separator on windows

mstorsjo added a subscriber: rnk. mstorsjo added a comment. FWIW, this change is not about mingw, it's about making the windows-with-forward-slashes configuration usable. The windows-with-forward-slashes configuration works just as fine in MSVC configurations, and that's where I've tested it.

[PATCH] D113268: [clangd] Fix tests with forward slash as separator on windows

mstorsjo added a comment. I forgot to add - I agree with the general sentiment though, that it’s not worth bending over backwards to make tests pass in a setup that’s known not to work in real use though. Then again, if someone were to want to actually step up to make it work in mingw setups,

[PATCH] D113268: [clangd] Fix tests with forward slash as separator on windows

mstorsjo added a comment. In D113268#3111798 , @sammccall wrote: > In D113268#3111709 , @mstorsjo > wrote: > >> FWIW, this change is not about mingw, it's about making the >> windows-with-forward-slashes

[PATCH] D113268: [clangd] Fix tests with forward slash as separator on windows

mstorsjo added a comment. Oh, I realized that issues relating to “mingw” probably are from msys (which often is used together, but is an entirely different thing) - yeah I can see that there’d be lots of unfixable issues with that, and I’m not signing up for looking into that… Msys is the

[PATCH] D111199: [Clang][LLVM][Attr] support btf_type_tag attribute

mstorsjo added a comment. I went ahead and reverted this, as it caused crashes when compiling a number of projects. The most reduced testcase is this: $ cat reduced.c void a(*); void a() {} $ clang -c reduced.c -O2 -g A full case (which reduces into this) is this,

[PATCH] D113254: [clang] Fix a misadjusted path style comparison in a unittest

mstorsjo created this revision. mstorsjo added a reviewer: dexonsmith. mstorsjo requested review of this revision. Herald added a project: clang. This was changed incorrectly by accident in 99023627010bbfefb71e25a2b4d056de1cbd354e

[PATCH] D113268: [clangd] Fix tests with forward slash as separator on windows

mstorsjo created this revision. mstorsjo added reviewers: sammccall, aaron.ballman. Herald added subscribers: usaxena95, kadircet, arphaman. mstorsjo requested review of this revision. Herald added subscribers: MaskRay, ilya-biryukov. Herald added a project: clang-tools-extra. If running on

[PATCH] D113264: [clang] [DirectoryWatcher] Remove leading \\?\ from GetFinalPathNameByHandleW

mstorsjo created this revision. mstorsjo added a reviewer: aaron.ballman. mstorsjo requested review of this revision. Herald added a project: clang. This is needed for the paths to work when using forward slashes; this fixes the DirectoryWatcherTests unit tests. Also allocate missing space for

[PATCH] D113269: [clang-move] Fix unit tests with forward slash as separator on windows

mstorsjo created this revision. mstorsjo added reviewers: aaron.ballman, sammccall. mstorsjo requested review of this revision. Herald added a project: clang-tools-extra. Repository: rG LLVM Github Monorepo https://reviews.llvm.org/D113269 Files:

[PATCH] D113268: [clangd] Fix tests with forward slash as separator on windows

mstorsjo added inline comments. Comment at: clang-tools-extra/clangd/tool/ClangdMain.cpp:546 fs::make_absolute(TestScheme::TestDir, Path); +path::native(Path); return std::string(Path); mstorsjo wrote: > sammccall wrote: > > This change is an

[PATCH] D111457: [clang][test] Add lit helper for windows paths

mstorsjo added a subscriber: rnk. mstorsjo added a comment. In D111457#3085273 , @keith wrote: > Yep this makes sense. I would rather not scope that into this if that's ok. > Since this test was already invalid and broken (actual fix in >

[PATCH] D111457: [clang][test] Add lit helper for windows paths

mstorsjo added a comment. In D111457#3073726 , @keith wrote: > In D111457#3066508 , @mstorsjo > wrote: > >> Wouldn't this one also be solved pretty much the same, but differently, by >> changing `if

[PATCH] D111952: [clang] [MinGW] Guess the right ix86 arch name spelling as sysroot

This revision was automatically updated to reflect the committed changes. Closed by commit rGd758069f5e0d: [clang] [MinGW] Guess the right ix86 arch name spelling as sysroot (authored by mstorsjo). Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION

[PATCH] D113254: [clang] Fix a misadjusted path style comparison in a unittest

This revision was automatically updated to reflect the committed changes. Closed by commit rG120d44d1a00b: [clang] Fix a misadjusted path style comparison in a unittest (authored by mstorsjo). Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION

[PATCH] D113254: [clang] Fix a misadjusted path style comparison in a unittest

mstorsjo added a comment. Yeah I guess I should have just committed this as trivial. Anyway, it doesn't seem to have had any notable impact on anything - luckily. Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D113254/new/

[PATCH] D115441: [X86][MS] Add 80bit long double support for Windows

mstorsjo added a comment. In D115441#3187009 , @rnk wrote: > I seem to recall assuming that Windows `long double` was 64-bits in many, > many places. Unfortunately, I have no idea where that could've happened. Nothing comes to mind for me about that -

[PATCH] D115441: [X86][MS] Add 80bit long double support for Windows

mstorsjo added a comment. In D115441#3188172 , @pengfei wrote: >> In GCC on Windows (and clang in mingw mode), long double is always 80 bit on >> x86. (On i386, sizeof(long double) == 12, while on x86_64 it's 16.) > > How about the alignment? I can see

[PATCH] D115045: [Clang] Ignore CLANG_DEFAULT_LINKER for custom-linker toolchains

mstorsjo added a comment. In D115045#3175648 , @phosek wrote: > Rather, we shouldn't consider `-DCLANG_DEFAULT_LINKER=lld` as a valid value. > Instead AMDGPU bot should use `-DCLANG_DEFAULT_LINKER=ld.lld`. I think this would be a kinda disruptive

[PATCH] D106721: [AArch64] Implemnt MSVC __mulh and __umulh builtins and corresponding IR level intrinsics

mstorsjo created this revision. mstorsjo added reviewers: rnk, STL_MSFT, efriedma, DavidSpickett. Herald added subscribers: danielkiss, hiraditya, kristof.beyls. mstorsjo requested review of this revision. Herald added projects: clang, LLVM. This seems to work, but would it need more testing for

[PATCH] D104601: [Preprocessor] Implement -fminimize-whitespace.

mstorsjo added a comment. This broke some cases for me, where the output is missing some newlines, breaking the output badly. I’ll try to provide a repro… Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D104601/new/ https://reviews.llvm.org/D104601

[PATCH] D104601: [Preprocessor] Implement -fminimize-whitespace.

mstorsjo added a comment. In D104601#2906588 , @mstorsjo wrote: > This broke some cases for me, where the output is missing some newlines, > breaking the output badly. I’ll try to provide a repro… https://martin.st/temp/repro.tar.xz clang -target

[PATCH] D106924: [Preprocessor] -E -P: Ensure newline after 8 skipped lines.

mstorsjo added a comment. Thanks! I can confirm that this fixes the issue I ran into. Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D106924/new/ https://reviews.llvm.org/D106924 ___ cfe-commits mailing

[PATCH] D116503: [clang] Add --start-no-unused-arguments/--end-no-unused-arguments to silence some unused argument warnings

mstorsjo updated this revision to Diff 397719. mstorsjo marked 2 inline comments as done. mstorsjo added a comment. Add testcases for using the new option with clang-cl, use `--target=` everywhere. Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION

[PATCH] D116503: [clang] Add --start-no-unused-arguments/--end-no-unused-arguments to silence some unused argument warnings

mstorsjo added a comment. In D116503#3223094 , @MaskRay wrote: > But make sure to wait a bit to see what others think. Do you think it's ok to go ahead and land this change now, or should I wait a bit more? (I'd like to have it landed with enough

[PATCH] D116919: [AST] Add RParen loc for decltype AutoTypeloc.

mstorsjo added a comment. Ok, now I managed to make a standalone reproducer: https://martin.st/temp/pch-preproc.hxx and https://martin.st/temp/main-preproc.cpp $ bin/clang -target x86_64-w64-mingw32 -w -x c++-header -c pch-preproc.hxx -o pch-preproc.hxx.pch -std=c++17 -O3 $ bin/clang

[PATCH] D116919: [AST] Add RParen loc for decltype AutoTypeloc.

mstorsjo added a comment. The relanded version in rG41fbdfa4d5601cccbcdc0ded8ef35190d502f7f3 seems to be breaking some builds with PCH for me, failing asserts like this: clang:

[PATCH] D116503: [clang] Add --start-no-unused-arguments/--end-no-unused-arguments to silence some unused argument warnings

This revision was landed with ongoing or failed builds. This revision was automatically updated to reflect the committed changes. Closed by commit rG50ec1306d060: [clang] Add --start-no-unused-arguments/--end-no-unused-arguments to silence… (authored by mstorsjo). Repository: rG LLVM Github

[PATCH] D116919: [AST] Add RParen loc for decltype AutoTypeloc.

mstorsjo added a comment. In D116919#3247887 , @hokein wrote: > Thanks, everyone! > > I manage to figure out the cause, the crash was caused by an arbitrary > RPareLoc -- we missed to set the RPareLoc in > `TreeTransform::TransformAutoType`. I reland

[PATCH] D121687: [clang-tidy] Don't try to build CTTestTidyModule for Windows with dylibs

mstorsjo created this revision. mstorsjo added a reviewer: rnk. Herald added subscribers: xazax.hun, mgorny. Herald added a project: All. mstorsjo requested review of this revision. Herald added a project: clang-tools-extra. In MinGW mode, it's possible to build LLVM/Clang with

[PATCH] D120727: [libc++] Overhaul how we select the ABI library

mstorsjo added subscribers: phosek, mstorsjo. mstorsjo added a comment. FWIW I think D116689 interacts with this somewhat. I think D116689 is useful for the fixes I want to do wrt libcxxabi integration on Windows (but I

[PATCH] D122298: [clang-cl] Ignore /Wv and /Wv:17 flags

mstorsjo accepted this revision. mstorsjo added a comment. This revision is now accepted and ready to land. The patch itself seems sensible to me. Comment at: clang/include/clang/Driver/Options.td:2496 Group, Flags<[CC1Option, NoArgumentUnused, HelpHidden]>; -def

[PATCH] D121873: [clang][extract-api] Add enum support

mstorsjo added a comment. This (or some closely related commit?) causes a huge amount of warnings when building with GCC: warning: ‘clang::extractapi::EnumRecord’ has a field ‘clang::extractapi::EnumRecord::Constants’ whose type uses the anonymous namespace [-Wsubobject-linkage] 167 |

[PATCH] D121687: [clang-tidy] Don't try to build CTTestTidyModule for Windows with dylibs

mstorsjo added inline comments. Comment at: clang-tools-extra/test/CMakeLists.txt:84 if (NOT LLVM_INSTALL_TOOLCHAIN_ONLY) - llvm_add_library( - CTTestTidyModule - MODULE clang-tidy/CTTestTidyModule.cpp - PLUGIN_TOOL clang-tidy - DEPENDS clang-tidy-headers)

[PATCH] D122766: [clang] Use forward slash as the path separator for Windows in __FILE__, __builtin_FILE(), and std::source_location

mstorsjo added a comment. > When targeting Windows, the path separator used when targeting Windows > depends on the build environment when Clang _itself_ is built. This leads to > inconsistencies in Chrome builds where Clang running on non-Windows > environments uses the forward slash (/) path

[PATCH] D122766: [clang] Use forward slash as the path separator for Windows in __FILE__, __builtin_FILE(), and std::source_location

mstorsjo added a comment. In D122766#3420925 , @thakis wrote: > Windows can handle slashes, but several tools can't. I worry that if we do > something different than cl, some random corner case might break (dbghelp, or > some source server thing or

[PATCH] D121687: [clang-tidy] Don't try to build CTTestTidyModule for Windows with dylibs

mstorsjo added a comment. @rnk Can you have a look at this? Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D121687/new/ https://reviews.llvm.org/D121687 ___ cfe-commits mailing list

[PATCH] D121687: [clang-tidy] Don't try to build CTTestTidyModule for Windows with dylibs

This revision was automatically updated to reflect the committed changes. Closed by commit rG3e7a8aab759a: [clang-tidy] Dont try to build CTTestTidyModule for Windows with dylibs (authored by mstorsjo). Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION

[PATCH] D118980: Don't dllexport reference temporaries

mstorsjo accepted this revision. mstorsjo added a comment. This revision is now accepted and ready to land. LGTM, seems reasonable! Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D118980/new/ https://reviews.llvm.org/D118980

[PATCH] D106674: Runtime for Interop directive

mstorsjo added inline comments. Comment at: openmp/runtime/src/dllexports:553 +omp_get_interop_ptr 761 +omp_get_interop_str 762 jdoerfert wrote: > Those values are taken by now, you need new values that are not

[PATCH] D106674: Runtime for Interop directive

mstorsjo added inline comments. Comment at: openmp/runtime/src/dllexports:553 +omp_get_interop_ptr 761 +omp_get_interop_str 762 mstorsjo wrote: > jdoerfert wrote: > > Those values are taken by now, you need new

[PATCH] D119234: [clang] [MinGW] Recognize -lcrtdll as a library replacing -lmsvcrt

mstorsjo created this revision. mstorsjo added reviewers: mati865, rnk. mstorsjo requested review of this revision. Herald added a project: clang. Repository: rG LLVM Github Monorepo https://reviews.llvm.org/D119234 Files: clang/lib/Driver/ToolChains/MinGW.cpp

[PATCH] D119234: [clang] [MinGW] Recognize -lcrtdll as a library replacing -lmsvcrt

This revision was landed with ongoing or failed builds. This revision was automatically updated to reflect the committed changes. Closed by commit rG079b6d02d1f5: [clang] [MinGW] Recognize -lcrtdll as a library replacing -lmsvcrt (authored by mstorsjo). Repository: rG LLVM Github Monorepo

[PATCH] D119326: [clang] [MinGW] Default to DWARF 4

mstorsjo created this revision. mstorsjo added reviewers: dblaikie, rnk, mati865. mstorsjo requested review of this revision. Herald added a project: clang. Neither LLDB nor GDB seem to work with DWARF 5 debug information on Windows right now. This applies the same change as in

[PATCH] D110485: Support [[no_unique_address]] for all targets.

mstorsjo added a comment. To pick up the thread here again, `[[no_unique_address]]` is done and settled in MSVC, with the slightly surprising semantics: `[[no_unique_address]]` is accepted, without any warning (in C++20 mode), but it has no effect. (This, not related to LLVM, but because they

[PATCH] D110485: Support [[no_unique_address]] for all targets.

mstorsjo added a comment. In D110485#3308854 , @mstorsjo wrote: > To pick up the thread here again, `[[no_unique_address]]` is done and settled > in MSVC, with the slightly surprising semantics: `[[no_unique_address]]` is > accepted, without any

[PATCH] D119326: [clang] [MinGW] Default to DWARF 4

This revision was automatically updated to reflect the committed changes. Closed by commit rG6cf64b2d2858: [clang] [MinGW] Default to DWARF 4 (authored by mstorsjo). Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D119326/new/

<    4   5   6   7   8   9   10   11   12   13   >