Correct test.
- rLLD LLVM Linker
- Tue, Feb 19, 2:36 AM
- Subversion
- rC Clang
- 76,435 Commits
- ·
- Restricted Project
- Restricted Project
- Mon, Feb 18, 3:12 PM
- Subversion
Today
Today
Don't use zero register for 'J' constraint.
Tue, Feb 19, 2:55 AM · Restricted Project
Anastasia added inline comments to D58346: [Sema] Change addr space diagnostics in casts to follow C++ style.
atanasyan committed rGfae2a509fa5a: [MIPS] Handle cross-mode (regular <-> microMIPS) jumps (authored by atanasyan).
[MIPS] Handle cross-mode (regular <-> microMIPS) jumps
[MIPS] Handle cross-mode (regular <-> microMIPS) jumps
[MIPS] Handle cross-mode (regular <-> microMIPS) jumps
[ARM GlobalISel] Support G_PHI for Thumb2
[ARM GlobalISel] Support G_PHI for Thumb2
evgeny777 added a comment to D58173: [llvm-objcopy] Add --set-start, --change-start, --adjust-start.
--adjust-start causes an underflow. I think this should be an error.
evgeny777 updated the diff for D58173: [llvm-objcopy] Add --set-start, --change-start, --adjust-start.
Addressed
evgeny777 added a comment to D58173: [llvm-objcopy] Add --set-start, --change-start, --adjust-start.
I don't personally have a use case for this, but I could imagine a program with multiple possible start locations, and these switches allow switching between them. I'd like to here the concrete use case though.
@craig.topper Any comments?
Tue, Feb 19, 2:20 AM · Restricted Project
baloghadamsoftware accepted D57883: [clang-tidy] refactor ExceptionAnalyzer further to give ternary answer.
In D57883#1402049, @lebedev.ri wrote:It's for D57108, i'we guessed that such ternary answer will be required there.
Tue, Feb 19, 2:20 AM · Restricted Project
jhenderson added inline comments to D58280: [yaml2obj][obj2yaml] - Support SHT_GNU_versym (.gnu.version) section..
Herald added a project to D57108: [clang-tidy] diagnose possibiltiy to add 'noexcept' in modernize-use-noexcept: Restricted Project.
Sorry for the reviews, i'm really stalling it seems..
Tue, Feb 19, 2:16 AM · Restricted Project
jhenderson accepted D58119: [obj2yaml][yaml2obj] - Add support of parsing/dumping of the .gnu.version_r section..
LGTM.
grimar added inline comments to D58280: [yaml2obj][obj2yaml] - Support SHT_GNU_versym (.gnu.version) section..
LGTM, with two small comments.
grimar added inline comments to D58119: [obj2yaml][yaml2obj] - Add support of parsing/dumping of the .gnu.version_r section..
grimar updated the diff for D58119: [obj2yaml][yaml2obj] - Add support of parsing/dumping of the .gnu.version_r section..
- Addressed review comments.
LGTM.
baloghadamsoftware added a comment to D57883: [clang-tidy] refactor ExceptionAnalyzer further to give ternary answer.
If I understand it correctly, this is more of an infrastructure improvement than check enhancement. It looks like a nice and clean code. Where do we expect to use this new behavior? In the current check or in the upcoming "modernize" check?
Tue, Feb 19, 2:04 AM · Restricted Project
jhenderson accepted D58174: [yaml2obj] - Do not skip zeroes blocks if there are relocations against them..
LGTM.
jhenderson added inline comments to D57680: [llvm-objdump] Implement `-Mreg-names-raw`/`-std` options..
Tue, Feb 19, 1:54 AM · Restricted Project
- guard with NetBSD
<strings.h> should be included unconditionally.
BrandonTJones added a comment to D57375: X86AsmParser AVX-512: Return error instead of hitting assert.
ping
Tue, Feb 19, 1:44 AM · Restricted Project
In D58379#1401986, @krytarowski wrote:Please enable the tests for NetBSD as well.
<strings.h> should be included unconditionally.
jhenderson added inline comments to D58119: [obj2yaml][yaml2obj] - Add support of parsing/dumping of the .gnu.version_r section..
- Unconditionally test bcmp() in weak_hook_test
jhenderson added a comment to D58173: [llvm-objcopy] Add --set-start, --change-start, --adjust-start.
In D58173#1401364, @jakehehrlich wrote:
- What's the use case for this?
- What's the interaction for program headers?
For relocatable files I'd expect addresses to be ignored and for program headers I'd like to put more thought into this. In general, I'd like to take a stance that we not implement flags that we don't have a use case for.