hintonda added a comment.
ping...
https://reviews.llvm.org/D16971
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
hintonda created this revision.
This option helps locate the origin of a diagnostic message
by outputing the enum name and index associated with a specific
DiagID, allowing users to grep the code for the enum name directly
without having to find it in the td files first.
Additional ideas:
1. add
hintonda added a comment.
In https://reviews.llvm.org/D35175#803496, @arphaman wrote:
> Thanks, that's pretty cool!
>
> How bigger did the clang binary get after you've added all these strings?
Currently looks like around 200k (4534 @ 33 byte avg length + ptr). If this is
too much, we could m
hintonda updated this revision to Diff 105961.
hintonda added a comment.
- Make option cc1 only. Rename function, and add test.
https://reviews.llvm.org/D35175
Files:
include/clang/Basic/DiagnosticIDs.h
include/clang/Basic/DiagnosticOptions.def
include/clang/Driver/CC1Options.td
lib/Ba
hintonda updated this revision to Diff 106094.
hintonda added a comment.
- Only maintain enum names in debug builds. Current cost of maintaining enum
names is 176k.
https://reviews.llvm.org/D35175
Files:
include/clang/Basic/DiagnosticIDs.h
include/clang/Basic/DiagnosticOptions.def
inclu
hintonda added a comment.
It's just an effort to make the code a bit more accessible, especially for new
users (or ones not used to running find/grep).
Steve had suggested adding an option that took the entire message and matched
it when it was produced. However, that won't work very well sinc
hintonda abandoned this revision.
hintonda added a comment.
Okay, sounds good. Look forward to seeing Alex's script...
https://reviews.llvm.org/D35175
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
http://lists.llvm.org/cgi-bin/mailman/lis
hintonda added a comment.
Not sure how to do this all in a script, but perhaps we could enhance diagtool
to generate this mapping for you. It currently only lists warnings, but I
don't think it would be difficult to extend it and generate a mapping you could
use in your script.
https://revi
hintonda added a comment.
I'd be happy to do that if it would help. If so, should I do it here create a
new diff?
Perhaps we might even make sense add the ability to pass in a message and find
the matching name/index.
https://reviews.llvm.org/D35175
___
hintonda added a comment.
Great, thanks...
https://reviews.llvm.org/D35175
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
hintonda accepted this revision.
hintonda added a comment.
This revision is now accepted and ready to land.
LGTM.
Category ids' might be useful, but I'd wait until someone actually needs them.
Repository:
rL LLVM
https://reviews.llvm.org/D35306
hintonda added inline comments.
Comment at: lib/Basic/CMakeLists.txt:17
macro(find_first_existing_vc_file out_var path)
set(git_path "${path}/.git")
LLVM already has a version of find_first_existing_vc_file in
llvm/include/llvm/Support/CMakelists.txt.
Wou
hintonda created this revision.
hintonda added reviewers: compnerd, bryant, beanz.
Herald added a subscriber: mgorny.
Herald added a reviewer: EricWF.
Herald added a reviewer: EricWF.
Don't print, possibly erroneous, warning if
LIBCXXABI_INCLUDE_TESTS is false.
This patch fixes a problem introduc
hintonda created this revision.
hintonda added reviewers: compnerd, beanz.
Herald added a subscriber: mgorny.
If CMAKE_ASM_COMPILER is set, cmake assumes
CMAKE_ASM_COMPILER_ID was also set, and doesn't try to set it.
Without CMAKE_ASM_COMPILER_ID, cmake can't set
CMAKE_ASM_COMPILER_OPTIONS_TARGET
hintonda updated this revision to Diff 130367.
hintonda added a comment.
Refactor to eliminate need for second stage and toolchain files.
Repository:
rC Clang
https://reviews.llvm.org/D41829
Files:
cmake/caches/Linux.cmake
Index: cmake/caches/Linux.cmake
==
hintonda updated this revision to Diff 130638.
hintonda edited the summary of this revision.
hintonda added a comment.
- Also pass LLVM_CONFIG_EXE.
- Add llvm-objcopy, and make sure that is stage2 cache file is passed,
Repository:
rC Clang
https://reviews.llvm.org/D41829
Files:
cmake/cache
hintonda added a comment.
Unless there's any further comment, I plan to commit these patches tomorrow.
Repository:
rCXXA libc++abi
https://reviews.llvm.org/D41623
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
http://lists.llvm.org/cgi-b
hintonda added a comment.
In https://reviews.llvm.org/D42232#982043, @compnerd wrote:
> Took a bit of reading to realize that the compiler is being set to clang, and
> therefore `CMAKE_ASM_COMPILER_ID` should be `Clang`.
Thanks, I'll add that to the commit message.
Repository:
rC Clang
ht
hintonda added a comment.
In https://reviews.llvm.org/D41623#982189, @phosek wrote:
> Isn't `NO_DEFAULT_PATH` more appropriate here? That's what libc++ uses as
> well.
Unfortunately no.
When cross compiling, it's important to look for headers in CMAKE_SYSROOT
and/or CMAKE_FIND_ROOT_PATH, or
hintonda added a comment.
In https://reviews.llvm.org/D41623#982262, @hintonda wrote:
> In https://reviews.llvm.org/D41623#982189, @phosek wrote:
>
> > Isn't `NO_DEFAULT_PATH` more appropriate here? That's what libc++ uses as
> > well.
>
>
> Unfortunately no.
>
> When cross compiling, it's impor
hintonda updated this revision to Diff 130748.
hintonda added a comment.
- Use LLVM_CONFIG_PATH instead of LLVM_CONFIG_EXE.
Repository:
rC Clang
https://reviews.llvm.org/D41829
Files:
cmake/caches/Linux.cmake
Index: cmake/caches/Linux.cmake
hintonda updated this revision to Diff 130820.
hintonda added a comment.
- Save CMAKE_REQUIRED_FLAGS for another patch.
Repository:
rCXX libc++
https://reviews.llvm.org/D41622
Files:
CMakeLists.txt
Index: CMakeLists.txt
===
hintonda created this revision.
hintonda added reviewers: beanz, compnerd, phosek, EricWF.
Herald added subscribers: cfe-commits, mgorny.
Clang and llvm already use llvm_setup_rpath(), so this change will
help standarize rpath usage across all projects.
Repository:
rCXX libc++
https://reviews
hintonda created this revision.
hintonda added reviewers: beanz, compnerd, phosek, EricWF.
Herald added subscribers: cfe-commits, mgorny.
Clang and llvm already use llvm_setup_rpath(), so this change will
help standarize rpath usage across all projects.
Repository:
rCXXA libc++abi
https://rev
hintonda created this revision.
hintonda added reviewers: beanz, compnerd, phosek, EricWF.
Herald added subscribers: cfe-commits, mgorny.
Set cmake policy CMP0068=NEW, if available -- depends on
https://reviews.llvm.org/D42463 which
also adds target property "BUILD_WITH_INSTALL_NAME_DIR On" to ma
hintonda added a comment.
It turns out that policies don't survive a call to `cmake_minimum_required()`,
so we need to rethink how llvm, et al, handles policies, e.g.:
- move `cmake_minimum_required()` inside the `if( CMAKE_SOURCE_DIR STREQUAL
CMAKE_CURRENT_SOURCE_DIR )` branch.
- move all `cma
hintonda reopened this revision.
hintonda added a comment.
This revision is now accepted and ready to land.
Reopen after rollback.
@smeenai pointed out LLVM may not be available in standalone builds
Repository:
rCXX libc++
https://reviews.llvm.org/D42459
__
hintonda reopened this revision.
hintonda added a comment.
This revision is now accepted and ready to land.
Reopen after rollback.
@smeenai pointed out LLVM may not be available in standalone builds
Repository:
rL LLVM
https://reviews.llvm.org/D42460
__
hintonda requested review of this revision.
hintonda added a comment.
I probably should have put this in the summary, but the reason for this patch
isn't just to clean up the cmake files, it's to fix a cmake configuration
failure when cross-compiling for Linux on Darwin:
CMake Error at
/Users/
hintonda updated this revision to Diff 131506.
hintonda added a comment.
Only call llvm_setup_rpath() if LLVM_FOUND is true.
Repository:
rCXX libc++
https://reviews.llvm.org/D42459
Files:
lib/CMakeLists.txt
Index: lib/CMakeLists.txt
===
hintonda updated this revision to Diff 131528.
hintonda added a comment.
Only call llvm_setup_rpath() if LLVM_FOUND is true.
Repository:
rCXXA libc++abi
https://reviews.llvm.org/D42460
Files:
src/CMakeLists.txt
Index: src/CMakeLists.txt
===
hintonda updated this revision to Diff 131537.
hintonda added a comment.
Only call llvm_setup_rpath() if LLVM_FOUND is true.
Repository:
rUNW libunwind
https://reviews.llvm.org/D42461
Files:
src/CMakeLists.txt
Index: src/CMakeLists.txt
hintonda updated this revision to Diff 131742.
hintonda edited the summary of this revision.
hintonda added a comment.
- Libfuzzers won't build, so turn off.
- Fix typos, and set cache.
Repository:
rC Clang
https://reviews.llvm.org/D41829
Files:
cmake/caches/Linux.cmake
Index: cmake/cache
hintonda added a comment.
In https://reviews.llvm.org/D42490#994154, @beanz wrote:
> Historically we've duplicated `cmake_policy` calls on a per-project basis
> which we needed to support standalone builds. That said, it would be nice if
> we had LLVM vend some CMake modules that encapsulated t
301 - 334 of 334 matches
Mail list logo