This made me think, is there a way to automatically find out redundant
deserializations? Eg. switch to text mode, build dependency graph and
compare against the dependency graph in -fmodules mode. Or even a more
brute force: store the states of the key objects (AST, identifiers,
lookups,
On Mar 10, 2015, at 1:10 AM, Vassil Vassilev vvasi...@cern.ch
mailto:vvasi...@cern.ch wrote:
ping... I really want to close that annoying bugzilla ticket ;)
On 18/09/14 20:08, Vassil Vassilev wrote:
On 18/09/14 18:04, Ben Langmuir wrote:
On Sep 18, 2014, at 3:47 AM, Vassil Vassilev
On 10/03/15 15:42, Ben Langmuir wrote:
Committed as r231789. Feel free to close the PR, or I’ll do it later.
Thanks for the patch!
Thank you Ben! PR closed.
Vassil
On Mar 10, 2015, at 6:50 AM, Vassil Vassilev vvasi...@cern.ch
mailto:vvasi...@cern.ch wrote:
On 10/03/15 14:21, Ben Langmuir
On 10/03/15 03:03, Richard Smith wrote:
On Mon, Mar 9, 2015 at 5:22 AM, Vassil Vassilev vvasi...@cern.ch
mailto:vvasi...@cern.ch wrote:
On 13/01/15 23:16, Richard Smith wrote:
On Thu, Nov 27, 2014 at 2:50 PM, Vassil Vassilev
vvasi...@cern.ch mailto:vvasi...@cern.ch wrote
ping... I really want to close that annoying bugzilla ticket ;)
On 18/09/14 20:08, Vassil Vassilev wrote:
On 18/09/14 18:04, Ben Langmuir wrote:
On Sep 18, 2014, at 3:47 AM, Vassil Vassilev
vasil.georgiev.vasi...@cern.ch wrote:
On 09/17/2014 09:51 PM, Ben Langmuir wrote:
Index: docs
On 13/01/15 23:16, Richard Smith wrote:
On Thu, Nov 27, 2014 at 2:50 PM, Vassil Vassilev vvasi...@cern.ch
mailto:vvasi...@cern.ch wrote:
On 25/11/14 00:52, Richard Smith wrote:
+ } + else if
(!isUnresolvedExceptionSpec(FPT-getExceptionSpecType())
Per the LLVM coding style
On 25/02/15 21:07, Sterling Augustine wrote:
Circling back to this incremental parsing.
On Tue, Feb 17, 2015 at 1:28 AM, Vassil Vassilev vvasi...@cern.ch wrote:
The reason we went for IncrementalProcessing is that it does more than just
parsing, it allows clang to be able to continue working
On 17/02/15 02:02, Sterling Augustine wrote:
Hello,
I'm working on a change that will enable a client to build an AST from
an input file, and then, based on the contents of that AST, add
additional tokens to be parsed. Richard Smith recommended adding a
feature to the parser to do this, but it
, this base dtor must be emitted as an alias. Could you help me out?
Vassil
On Mon, Nov 24, 2014 at 11:09 AM, Vassil Vassilev vvasi...@cern.ch
mailto:vvasi...@cern.ch wrote:
Sorry for the delay. Attaching the new version.
Vassil
On 14/11/14 02:47, Richard Smith wrote
.
The patch needs rebasing (we have a new helper function in ASTContext
to update the exception specification of a declaration), but looks
like the right direction.
On Thu, Nov 6, 2014 at 4:23 AM, Vassil Vassilev
vasil.georgiev.vasi...@cern.ch
mailto:vasil.georgiev.vasi...@cern.ch wrote:
Hi
be okay with just
using FileCheck:
// CHECK: {{Inputs[/\\]declare-use[/\\]module.map}}:30:10: note: header file...
Jordan
On Nov 3, 2014, at 2:10 , Vassil Vassilev vvasi...@cern.ch wrote:
Hi guys,
I am working on http://llvm.org/bugs/show_bug.cgi?id=20507 Now the diagnostic
gets issued
Hi Richard,
I am attaching the patch we discussed at the dev meeting. Still
haven't found small reproducer...
The issue appears to stem from the fact that module A contains only a
forward declaration of a function and it exception spec cannot be
computed. In module B is the definition with
Hi guys,
I am working on http://llvm.org/bugs/show_bug.cgi?id=20507 Now the
diagnostic gets issued for:
Clang :: Modules/declare-use1.cpp
Clang :: Modules/declare-use2.cpp
Clang :: Modules/declare-use3.cpp
Clang :: Modules/strict-decluse.cpp
It says smth like:
On 03/11/14 16:49, Ben Langmuir wrote:
On Nov 3, 2014, at 2:10 AM, Vassil Vassilev vvasi...@cern.ch wrote:
Hi guys,
I am working on http://llvm.org/bugs/show_bug.cgi?id=20507 Now the diagnostic
gets issued for:
Clang :: Modules/declare-use1.cpp
Clang :: Modules/declare-use2.cpp
Thanks!
Vassil
On 31/10/14 17:55, Argyrios Kyrtzidis wrote:
Committed slightly modified version in r220965.
On Oct 28, 2014, at 3:04 PM, Vassil Vassilev vvasi...@cern.ch
mailto:vvasi...@cern.ch wrote:
ping...
On 15/10/14 12:51, Vassil Vassilev wrote:
Forgot to attach the patch I came up
ping...
On 15/10/14 12:51, Vassil Vassilev wrote:
Forgot to attach the patch I came up with.
Vassil
On 15/10/14 21:41, Vassil Vassilev wrote:
Hi folks,
If ENABLE_CLANG_ARCMT=0 it seems we don't build libclang, so the
unittests for libclang should be disabled too. Please review.
Cheers
On 14/10/14 20:10, Brad King wrote:
On 09/24/2014 03:05 PM, Brad King wrote: On 09/15/2014 02:48 PM, Brad King
wrote:
On 09/15/2014 02:42 PM, Reid Kleckner wrote:
lgtm
Thanks. I do not have commit access. Would someone apply, please?
Ping. I've confirmed that the patches still work on
Hi folks,
If ENABLE_CLANG_ARCMT=0 it seems we don't build libclang, so the
unittests for libclang should be disabled too. Please review.
Cheers,
Vassil
___
cfe-commits mailing list
cfe-commits@cs.uiuc.edu
Forgot to attach the patch I came up with.
Vassil
On 15/10/14 21:41, Vassil Vassilev wrote:
Hi folks,
If ENABLE_CLANG_ARCMT=0 it seems we don't build libclang, so the
unittests for libclang should be disabled too. Please review.
Cheers,
Vassil
--- tools/clang/unittests/Makefile
+++ tools
Could you add cast_or_null also in the
ASTStmtReader::VisitPredefinedExpr(PredefinedExpr *E), otherwise the
deserialization would be out of sync.
Vassil
On 10/10/14 20:58, Alexey Bataev wrote:
Author: abataev
Date: Fri Oct 10 13:58:13 2014
New Revision: 219525
URL:
On 10/02/2014 01:27 AM, David Blaikie wrote:
On Wed, Oct 1, 2014 at 3:01 AM, Vassil Vassilev
vasil.georgiev.vasi...@cern.ch
mailto:vasil.georgiev.vasi...@cern.ch wrote:
We tried to trim down the reproducer, we basically touched each
line to try to simplify it but unsuccessfully
On 2.10.2014 г., at 17:21, David Blaikie dblai...@gmail.com wrote:
On Thu, Oct 2, 2014 at 12:34 AM, Vassil Vassilev
vasil.georgiev.vasi...@cern.ch wrote:
On 10/02/2014 01:27 AM, David Blaikie wrote:
On Wed, Oct 1, 2014 at 3:01 AM, Vassil Vassilev
vasil.georgiev.vasi...@cern.ch
not assert when adding an implicit member coming from
a module at
writing time.
Patch by Vassil Vassilev!
Added:
cfe/trunk/test/Modules/Inputs/PR20399/
cfe/trunk/test/Modules/Inputs/PR20399/FirstHeader.h
cfe/trunk/test/Modules/Inputs/PR20399/SecondHeader.h
mailto:rich...@metafoo.co.uk wrote:
I think we'll be ready to do this for Clang 3.6. I don't have
strong feelings about whether we should do this now or wait until
we've had more time for testing and bug fixes, though.
On Tue, Sep 30, 2014 at 1:16 AM, Vassil Vassilev
That is pretty cool! Does this mean we could merge fcxx-modules into
fmodules flags?
Vassil
On 09/29/2014 07:46 PM, Richard Smith wrote:
Author: rsmith
Date: Mon Sep 29 12:46:41 2014
New Revision: 218614
URL: http://llvm.org/viewvc/llvm-project?rev=218614view=rev
Log:
Update modules
Thanks! Could you commit, please (I don't have commit access)?
Vassil
On 29/09/14 20:03, Richard Smith wrote:
LGTM, sorry for the delay. Thanks!
On Tue, Sep 23, 2014 at 1:58 PM, Vassil Vassilev vvasi...@cern.ch
mailto:vvasi...@cern.ch wrote:
ping...
On 19/09/14 14:37, Vassil
ping...
On 19/09/14 14:37, Vassil Vassilev wrote:
Hi,
I am attaching a patch addressing bug report
http://llvm.org/bugs/show_bug.cgi?id=20399
I believe this bug could be reproduced also by a chaining PCHs but I
haven't had the time to do it.
Vassil
! It seems I got it all wrong and I lost my reproducer :( I will
try to come up with a reasonable patch next week.
Vassil
Ben
On Sep 17, 2014, at 1:04 AM, Vassil Vassilev vasil.georgiev.vasi...@cern.ch
wrote:
Hi,
I am attaching a patch addressing llvm.org/bugs/show_bug.cgi?id=20467
Vassil
Hi,
I am attaching a patch addressing bug report
http://llvm.org/bugs/show_bug.cgi?id=20399
I believe this bug could be reproduced also by a chaining PCHs but I
haven't had the time to do it.
Vassil
Index: test/Modules/Inputs/PR20399/stl_map.h
but Modules in general (I decided to put them
into a separate section). Is it any better?
Vassil
Ben
On Sep 17, 2014, at 12:59 AM, Vassil Vassilev vasil.georgiev.vasi...@cern.ch
wrote:
Hi,
I am attaching a patch addressing http://llvm.org/bugs/show_bug.cgi?id=18985
I wasn't sure whether I had
On 09/17/2014 09:58 PM, Aaron Ballman wrote:
On Wed, Sep 17, 2014 at 4:04 AM, Vassil Vassilev
vasil.georgiev.vasi...@cern.ch wrote:
Hi,
I am attaching a patch addressing llvm.org/bugs/show_bug.cgi?id=20467
Vassil
Index: lib/Frontend/CompilerInstance.cpp
On 18/09/14 18:04, Ben Langmuir wrote:
On Sep 18, 2014, at 3:47 AM, Vassil Vassilev vasil.georgiev.vasi...@cern.ch
wrote:
On 09/17/2014 09:51 PM, Ben Langmuir wrote:
Index: docs/LanguageExtensions.rst
===
--- docs
Hi,
I am attaching a patch addressing
http://llvm.org/bugs/show_bug.cgi?id=18985
I wasn't sure whether I had to say 'experimental C++ modules'.
Vassil
Index: docs/LanguageExtensions.rst
===
--- docs/LanguageExtensions.rst
Hi,
I am attaching a patch addressing llvm.org/bugs/show_bug.cgi?id=20467
Vassil
Index: include/clang/Basic/DiagnosticFrontendKinds.td
===
--- include/clang/Basic/DiagnosticFrontendKinds.td (revision 217432)
+++
bump...
On 20/06/14 21:03, Vassil Vassilev wrote:
ping...
On 15/06/14 17:44, Vassil Vassilev wrote:
Hi,
This patch gives me access to the list of consumers, which I find
very handy in plugin development. For example, this allows my plugin
to put itself in a preferred position among the rest
some memory reduction in some cases for clang, too.
Is there any chance to keep the ReleaseMacroInfo upstream?
Vassil
On 08/04/2014 01:50 AM, Richard Smith wrote:
Fixed in a much more simple way in r214675. Thanks for reporting!
On Sun, Aug 3, 2014 at 11:52 AM, Vassil Vassilev vvasi...@cern.ch
)
[cling] MyClass m; m.do(); // Here do and MyClass may have completely
different implementation.
Vassil
Yaron
2014-08-04 10:47 GMT+03:00 Vassil Vassilev
vasil.georgiev.vasi...@cern.ch mailto:vasil.georgiev.vasi...@cern.ch:
Hi Richard,
Thanks for the fix!
Unfortunately
-04 12:26 GMT+03:00 Vassil Vassilev
vasil.georgiev.vasi...@cern.ch mailto:vasil.georgiev.vasi...@cern.ch:
Hi Yaron,
On 08/04/2014 10:31 AM, Yaron Keren wrote:
Hi Vassil,
If you decide to keep the last code posted as a cling patch, it
could do with 'I' only instead of 'I
-08-04 13:42 GMT+03:00 Vassil Vassilev
vasil.georgiev.vasi...@cern.ch mailto:vasil.georgiev.vasi...@cern.ch:
Hi Yaron,
Thanks for raising those very interesting points!
On 08/04/2014 12:12 PM, Yaron Keren wrote:
Hi,
You should not destruct SmallVector explicitly as its
) {
MIChainHead = I-Next;
I-~MacroInfoChain();
}
or something else?
Yaron
2014-08-02 23:05 GMT+03:00 Vassil Vassilev vvasi...@cern.ch
mailto:vvasi...@cern.ch:
Hi,
In cases where ReleaseMacroInfo gets called and it doesn't
cleanup the Preprocessor's MIChainHead can lead
. In all other cases it should not
change.
As it is now, the loop will always terminate with MIChainHead == nullptr.
Yaron
2014-08-03 21:10 GMT+03:00 Vassil Vassilev vvasi...@cern.ch
mailto:vvasi...@cern.ch:
Hi Yaron,
Yes I meant double destruction.
Vassil
On 03/08/14 20:08
(MacroInfoChain *I = current) {
+if ((I-MI) == MI) {
+ I-Next = (I-Next) ? I-Next-Next : 0;
+ if (I == MIChainHead)
+MIChainHead = I-Next;
+ break;
+}
+current = I-Next;
+ }
+
MI-~MacroInfo();
}
On 03/08/14 20:47, Vassil Vassilev wrote:
Hi,
Sorry overlooked
Hi,
In cases where ReleaseMacroInfo gets called and it doesn't cleanup
the Preprocessor's MIChainHead can lead to double deletion. I am sending
the patch that fixes the problem for me.
Vassil
diff --git a/lib/Lex/PPDirectives.cpp b/lib/Lex/PPDirectives.cpp
index 5f38387..1a9b5eb 100644
---
ping...
On 15/06/14 17:44, Vassil Vassilev wrote:
Hi,
This patch gives me access to the list of consumers, which I find
very handy in plugin development. For example, this allows my plugin
to put itself in a preferred position among the rest of the consumers.
It'd be great if it becomes
Hi,
This patch gives me access to the list of consumers, which I find
very handy in plugin development. For example, this allows my plugin to
put itself in a preferred position among the rest of the consumers.
It'd be great if it becomes a part of the mainline.
Many thanks,
Vassil
Index:
...@gmail.com
wrote:
On 22 May 2014 05:04, Vassil Vassilev vasil.georgiev.vasi...@cern.ch
wrote:
Hi Rafael,
Do you think we could rely for this case on post commit review?
According to policy no.
Richard, can you help us with this 2 month old code review? A rebased
patch is attached.
Did I
Hi Rafael,
Do you think we could rely for this case on post commit review?
Cheers,
Vassil
On 04/24/2014 08:52 PM, Rafael Espíndola wrote:
ping 3
On 24 March 2014 00:01, Rafael Espíndola rafael.espind...@gmail.com wrote:
ping 2
On 17 March 2014 06:59, Vassil Vassilev vasil.georgiev.vasi
ping...
On 03/11/2014 01:05 PM, Rafael Espíndola wrote:
I am not sure I understand the intended semantics of the UnnamedAddr argument
to GetOrCreateLLVMGlobal. Is it “this use does not care about the specific
address”? Because, if so, the (existing) logic is completely backwards: we
need to
On 03/10/2014 04:40 PM, Rafael Espíndola wrote:
bccing cfe-dev, moving the cfe-commits
On 10 March 2014 09:58, Vassil Vassilev vasil.georgiev.vasi...@cern.ch wrote:
Hi,
I noticed that CodeGenModule.cpp (around line number 1514):
if (UnnamedAddr)
Entry-setUnnamedAddr(true
On 10/18/12 7:04 PM, Douglas Gregor wrote:
On Oct 11, 2012, at 8:02 AM, Vassil Vassilev vasil.georgiev.vasi...@cern.ch
wrote:
Hi,
We need to attach multiple ExternalSources to Sema, to provide modules /and/
last resort lookup. Right now that's a pain. The attached patch makes this much
Ping...
On 10/11/2012 05:02 PM, Vassil Vassilev wrote:
Hi,
We need to attach multiple ExternalSources to Sema, to provide modules
/and/ last resort lookup. Right now that's a pain. The attached patch
makes this much easier, at (almost?) no runtime cost for the compiler.
Okay to commit
Hi,
We need to attach multiple ExternalSources to Sema, to provide modules
/and/ last resort lookup. Right now that's a pain. The attached patch
makes this much easier, at (almost?) no runtime cost for the compiler.
Okay to commit?
Vassil
Index: tools/clang/include/clang/Sema/Sema.h
if we decide to keep it...
Vassil
Probably you should wait for Doug or someone to review this before committing.
Jordan
On Oct 11, 2012, at 8:02 , Vassil Vassilev vasil.georgiev.vasi...@cern.ch
wrote:
Hi,
We need to attach multiple ExternalSources to Sema, to provide modules /and/
last
On 8/18/12 4:47 AM, Jordan Rose wrote:
I made a few small tweaks, but before I commit I want to confirm again that
this patch works for cling. Vassil?
Looks good. Thanks a lot to both of you Jordan and Andy!
Vassil
___
cfe-commits mailing list
Hi Andy,
Thanks for the both patches! It's amazing how quickly you find your
way in cling!
The second patch is what we exactly need.
Thanks a lot for your time,
Vassil
On 8/14/12 7:02 PM, Andy Gibbs wrote:
Hi,
Attached is an improvement to the backup checking algorithm that better
Hi,
On 8/11/12 9:52 PM, Andy Gibbs wrote:
On Saturday, August 11, 2012 6:03 PM, Vassil Vassilev wrote:
On 8/10/12 8:22 PM, Andy Gibbs wrote:
[...snip...]
I need to look into this for you. Certainly this is a use-case I
didn't
consider -- my apologies. Would you be willing to post some
Hi,
Thanks for the quick response!
On 8/10/12 8:22 PM, Andy Gibbs wrote:
On Friday, August 10, 2012 3:08 PM, Vassil Vassilev wrote:
Hi,
If I understand correctly:
+ if (E (FilesParsedForDirectives.count(E)
+ || HS.findModuleForHeader(E)))
+continue
Hi,
If I understand correctly:
+ if (E (FilesParsedForDirectives.count(E)
+ || HS.findModuleForHeader(E)))
+continue;
This doesn't handle the case where the files come from memory
buffers and have only FileIDs. It breaks cling's use case.
Is there any
Hi Sean,
Thanks for your comments!
On 07/18/2012 11:32 PM, Sean Silva wrote:
Why not
std::equal(LHS.begin(), LHS.end(), RHS.begin(), RHS.end())
?
Yes I can use std::equal, however I don't think I can gain much.
You may also want factor in if they are the same kind of
DeclGroupRef (i.e.
ping...
Original Message
Subject:Compare two DeclGroupRefs
Date: Mon, 16 Jul 2012 10:01:22 +0200
From: Vassil Vassilev vasil.georgiev.vasi...@cern.ch
To: cfe-...@cs.uiuc.edu cfe-...@cs.uiuc.edu
Hello devs,
I need to compare two clang::DeclGroupRefs. Could
Hi,
I am working on the following example:
[cling]$ #include iostream
[cling]$ #include fstream
[cling]$ std::ofstream myfile;
[cling]$ myfile.open(TmpClassDef.h);
[cling]$ myfile class MyClass{};\n
[cling]$ myfile error_here;;
[cling]$ myfile.close();
[cling]$ #include TmpClassDef.h ---
:43:55 -0800
From: Douglas Gregor dgre...@apple.com
To: Vassil Vassilev vasil.georgiev.vasi...@cern.ch
CC: cfe-...@cs.uiuc.edu cfe-...@cs.uiuc.edu
On Dec 2, 2011, at 7:02 AM, Vassil Vassilev wrote:
Hi,
In cling we have line-by-line input that comes from a terminal-like prompt.
We
ping
On 04.12.2011 г. 11:34, Vassil Vassilev wrote:
Hi Doug,
Sorry I didn't give the use case. I thought that this will be only
benefit for the compiler.
So in our use case (in cling) we want to allow stmts on the global
scope like:
[cling$] int i =0;
[cling
wrote:
On Nov 30, 2011, at 6:22 AM, Vassil Vassilev wrote:
Hi,
The attached patch tries to split the semantic checking from the AST node
creation.
I think it is useful one to be able to check already created AST node. Do
you think that
makes sense? Can that patch end up in clang's
Bump...
On 11/30/2011 03:22 PM, Vassil Vassilev wrote:
Hi,
The attached patch tries to split the semantic checking from the AST
node creation.
I think it is useful one to be able to check already created AST
node. Do you think that
makes sense? Can that patch end up in clang's mainline
Hi,
The attached patch tries to split the semantic checking from the AST
node creation.
I think it is useful one to be able to check already created AST
node. Do you think that
makes sense? Can that patch end up in clang's mainline?
If it gets in, later on I can improve the
Hi,
Is the website menu position intentionally fixed. My screen is not
big enough and I don't see section The Code, Quick Links and Clang
Events. If it is not intentional the attached patch should fix it.
Cheers,
Vassil
Index: tools/clang/www/menu.css
Hi,
This is not patch actually and I don't think it needs long time for
revising, but it helps the generation of doxygen from DeclCXX.
Vassil
Original Message
Subject:Re: [cfe-dev] Error in Clang Doxygen?
Date: Fri, 29 Jul 2011 18:45:02 +0200
From: Vassil
68 matches
Mail list logo