Author: jvesely
Date: Sun Oct 22 07:21:56 2017
New Revision: 316290
URL: http://llvm.org/viewvc/llvm-project?rev=316290&view=rev
Log:
shared: Implement aligned vector loads (vloada_half)
Passes newly posted piglits on turks and carrizo
v2: add scalar vloada_half
v3: fix typo
Reviewer: Aaron Watr
Author: jvesely
Date: Sun Oct 22 07:21:59 2017
New Revision: 316291
URL: http://llvm.org/viewvc/llvm-project?rev=316291&view=rev
Log:
shared: Implement aligned vector stores (vstorea_half)
Float version passes newly posted piglit tests on turks, float and double pass
on carrizo.
v2: scalar vstor
Author: faisalv
Date: Sun Oct 22 07:45:08 2017
New Revision: 316292
URL: http://llvm.org/viewvc/llvm-project?rev=316292&view=rev
Log:
[C++17] Fix PR34970 - tweak overload resolution for class template
deduction-guides in line with WG21's p0620r0.
In order to identify the copy deduction candidate
erichkeane added inline comments.
Comment at: lib/Sema/SemaTemplate.cpp:4811
+ TemplateParameterList *Params =
+ cast(Template->getMostRecentDecl())
+ ->getTemplateParameters();
mstorsjo wrote:
> How does this work if there's another forward declar
mstorsjo added inline comments.
Comment at: lib/Sema/SemaTemplate.cpp:4811
+ TemplateParameterList *Params =
+ cast(Template->getMostRecentDecl())
+ ->getTemplateParameters();
erichkeane wrote:
> mstorsjo wrote:
> > How does this work if there's an
Author: mstorsjo
Date: Sun Oct 22 12:39:26 2017
New Revision: 316300
URL: http://llvm.org/viewvc/llvm-project?rev=316300&view=rev
Log:
Make HIDDEN_DIRECTIVE a function-like macro. NFCI.
This avoids a hack for making it a no-op for windows.
Also explicitly check for _WIN32 instead of assuming it.
This revision was automatically updated to reflect the committed changes.
Closed by commit rL316300: Make HIDDEN_DIRECTIVE a function-like macro. NFCI.
(authored by mstorsjo).
Changed prior to commit:
https://reviews.llvm.org/D39156?vs=119740&id=119792#toc
Repository:
rL LLVM
https://review
faisalv created this revision.
faisalv added a project: clang.
Herald added a subscriber: eraman.
I'd like to harden my patch here: https://reviews.llvm.org/rL316292 by adding
some assertions.
But since the assertions in Decl,.h (FunctionDecl) require knowledge from
DeclCXX.h (CXXDeductionGuide
Author: d0k
Date: Sun Oct 22 13:16:28 2017
New Revision: 316301
URL: http://llvm.org/viewvc/llvm-project?rev=316301&view=rev
Log:
Create fewer copies of StringMaps. No functionality change intended.
Modified:
cfe/trunk/lib/Frontend/PrecompiledPreamble.cpp
cfe/trunk/lib/Parse/ParseStmtAsm.
lichray added a comment.
Isn't it already an UB if someone set `WillHaveBody` and but later
`IsCopyDeductionCandidate` being read, vice versa?
Repository:
rL LLVM
https://reviews.llvm.org/D39166
___
cfe-commits mailing list
cfe-commits@lists.llv
faisalv added a comment.
In https://reviews.llvm.org/D39166#903131, @lichray wrote:
> Isn't it already an UB if someone set `WillHaveBody` and but later
> `IsCopyDeductionCandidate` being read, vice versa?
Yes that would be UB - but I'm not sure I see how that would happen w the
current imple
zturner requested changes to this revision.
zturner added a comment.
This revision now requires changes to proceed.
Please don't throw an exception here. Instead, write this as:
shlibpath_var = None
if platform.system() in ['Linux', 'FreeBSD', 'NetBSD']:
shilbpath = 'LD_LIBRARY_PATH'
lichray marked an inline comment as done.
lichray added a comment.
In https://reviews.llvm.org/D39162#903179, @zturner wrote:
> Please don't throw an exception here. Instead, write this as:
>
> lit_config.warning('Unable to determine shared library path variable for
> platform {}'.format(plat
joerg added a comment.
I think we should special case Darwin and Windows and fall-back to
LD_LIBRARY_PATH for the rest. Can't remember if there is a UNIX-like platform
left where it doesn't work.
https://reviews.llvm.org/D39162
___
cfe-commits mai
lichray added a comment.
In https://reviews.llvm.org/D39162#903186, @joerg wrote:
> I think we should special case Darwin and Windows and fall-back to
> LD_LIBRARY_PATH for the rest. Can't remember if there is a UNIX-like platform
> left where it doesn't work.
If those developers come to us,
krytarowski added a comment.
In https://reviews.llvm.org/D39162#903186, @joerg wrote:
> I think we should special case Darwin and Windows and fall-back to
> LD_LIBRARY_PATH for the rest. Can't remember if there is a UNIX-like platform
> left where it doesn't work.
SHLIB_PATH in HP/UX
https:
lichray updated this revision to Diff 119800.
https://reviews.llvm.org/D39162
Files:
test/Unit/lit.cfg.py
Index: test/Unit/lit.cfg.py
===
--- test/Unit/lit.cfg.py
+++ test/Unit/lit.cfg.py
@@ -35,13 +35,15 @@
if symbolizer in
joerg added a comment.
I've looked at this in some detail now. I'm not exactly sure yet why it is
broken. The patch seems quite wrong to me. DW_CFA_GNU_args_size should be
applied only when unwinding a call instruction and that regard, the commit
message of the original change is quite correct.
krytarowski added a comment.
In https://reviews.llvm.org/D39162#903190, @krytarowski wrote:
> In https://reviews.llvm.org/D39162#903186, @joerg wrote:
>
> > I think we should special case Darwin and Windows and fall-back to
> > LD_LIBRARY_PATH for the rest. Can't remember if there is a UNIX-like
compnerd added inline comments.
Comment at: src/UnwindRegistersRestore.S:98
+ # skip fs
+ # skip gs
+ movq 56(%rcx), %rsp # cut back rsp to new location
Doesn't Win64 ABI require some of the MMX registers be saved/restored too?
https://reviews.llvm.org/D38
mstorsjo added inline comments.
Comment at: src/UnwindRegistersRestore.S:98
+ # skip fs
+ # skip gs
+ movq 56(%rcx), %rsp # cut back rsp to new location
compnerd wrote:
> Doesn't Win64 ABI require some of the MMX registers be saved/restored too?
Right, yes,
Very much a +1 on having attributes map in via `__attribute__` and
`[[clang::]]` uniformly when appropriate.
On Sat, Oct 21, 2017 at 8:14 AM, Aaron Ballman via cfe-commits <
cfe-commits@lists.llvm.org> wrote:
> Attributes come with multiple spelling flavors, but when it comes to
> adding new attr
Author: faisalv
Date: Sun Oct 22 15:29:52 2017
New Revision: 316304
URL: http://llvm.org/viewvc/llvm-project?rev=316304&view=rev
Log:
[c++2a] Update cxx_status w __VA_OPT__ marked as completed in SVN.
Modified:
cfe/trunk/www/cxx_status.html
Modified: cfe/trunk/www/cxx_status.html
URL:
http:
Because this is not tge only part of ghe library search algorithm, and it’s
not necessarily an error if the variable isn’t set
On Sun, Oct 22, 2017 at 1:56 PM Zhihao Yuan via Phabricator <
revi...@reviews.llvm.org> wrote:
> lichray marked an inline comment as done.
> lichray added a comment.
>
>
erichkeane added inline comments.
Comment at: lib/Sema/SemaTemplate.cpp:4811
+ TemplateParameterList *Params =
+ cast(Template->getMostRecentDecl())
+ ->getTemplateParameters();
mstorsjo wrote:
> erichkeane wrote:
> > mstorsjo wrote:
> > > How does
lichray updated this revision to Diff 119803.
lichray edited the summary of this revision.
lichray added a comment.
Changed to an warning given @zturner 's comments and experiments.
https://reviews.llvm.org/D39162
Files:
test/Unit/lit.cfg.py
Index: test/Unit/lit.cfg.py
=
alexshap created this revision.
Herald added subscribers: szepet, xazax.hun.
In getLValueElement Base may represent the address of a label (as in the
newly-added test case),
in this case it's not a loc::MemRegionVal and Base.castAs()
triggers an assert.
Test plan: make check-all
Repository:
rjmccall added inline comments.
Comment at: clang/lib/CodeGen/CGObjCMac.cpp:5095
+fieldType = fieldType->getAsArrayTypeUnsafe()->getElementType();
+ }
+
You can't just use isa<> here; there can be typedefs of incomplete array type.
https://reviews.llvm.org
Author: rsmith
Date: Sun Oct 22 20:58:34 2017
New Revision: 316310
URL: http://llvm.org/viewvc/llvm-project?rev=316310&view=rev
Log:
For better compatibility with C++11 and C++14, emit a nondiscardable definition
of a static constexpr data member if it's defined 'constexpr' out of line, not
only i
phosek created this revision.
phosek added a project: clang.
Passing a flavor to LLD requires command line argument, but if these
are being passed through a response file, this will fail because LLD
needs to know which driver to use before processing the response file.
Use ld.lld directly instead
Author: ibiryukov
Date: Sun Oct 22 23:06:21 2017
New Revision: 316311
URL: http://llvm.org/viewvc/llvm-project?rev=316311&view=rev
Log:
[clangd] Report proper kinds for 'Keyword' and 'Snippet' completion items.
Reviewers: rwols, malaperle, krasimir, bkramer, sammccall
Reviewed By: rwols, sammcca
This revision was automatically updated to reflect the committed changes.
Closed by commit rL316311: [clangd] Report proper kinds for 'Keyword' and
'Snippet' completion items. (authored by ibiryukov).
Repository:
rL LLVM
https://reviews.llvm.org/D38720
Files:
clang-tools-extra/trunk/clangd/
ilya-biryukov updated this revision to Diff 119813.
ilya-biryukov marked an inline comment as done.
ilya-biryukov added a comment.
- Inlined makeFutureAPIFromCallback.
https://reviews.llvm.org/D38629
Files:
clangd/ClangdServer.cpp
clangd/ClangdServer.h
Index: clangd/ClangdServer.h
ilya-biryukov added inline comments.
Comment at: clangd/Trace.cpp:87
+
+static Tracer* T = nullptr;
+} // namespace
Maybe use `unique_ptr` (or even `llvm::Optional`)? Otherwise we
leak memory and don't flush the stream if users of tracing API forget to call
`st
ilya-biryukov added inline comments.
Comment at: clangd/ClangdServer.cpp:51
+template
+std::future makeFutureAPIFromCallback(
sammccall wrote:
> I'm not sure we need a template here rather than just writing the code
> directly.
Removed the template, inlined i
35 matches
Mail list logo
