From cfe-users at lists.llvm.org  Tue Dec  3 10:32:55 2019
From: cfe-users at lists.llvm.org (Robert Ankeney via cfe-users)
Date: Tue, 3 Dec 2019 10:32:55 -0800
Subject: [cfe-users] DeclRefExpr question
Message-ID: <CAJL6GZpXSV8J2OTWU2VDPRNVYqvi19k56C7bn-VFxuq4uZJNOQ@mail.gmail.com>

Let's assume I have a DeclRefExpr that is referenceing some variable that
is declared as "static const int var;". How can I determine the variable is
static? There is no isStaticLocal() or getStorageClass() call as there
would be for the VarDecl. Is there something within the QualType that gets
me that info?

Many thanks,
Robert
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/cfe-users/attachments/20191203/b68ab167/attachment.html>

From cfe-users at lists.llvm.org  Tue Dec  3 10:43:04 2019
From: cfe-users at lists.llvm.org (David Blaikie via cfe-users)
Date: Tue, 3 Dec 2019 13:43:04 -0500
Subject: [cfe-users] DeclRefExpr question
In-Reply-To: <CAJL6GZpXSV8J2OTWU2VDPRNVYqvi19k56C7bn-VFxuq4uZJNOQ@mail.gmail.com>
References: <CAJL6GZpXSV8J2OTWU2VDPRNVYqvi19k56C7bn-VFxuq4uZJNOQ@mail.gmail.com>
Message-ID: <CAENS6EuzTKxRPmpCXLHmJWppn0h-7U-FRPo1Yc3VcEyj1Mjo6Q@mail.gmail.com>

I'd guess you'd call "getDecl" to get the decl the DeclRefExpr is
referencing, then dyn_cast that (the decl is a ValueDecl*) to VarDecl to
see if it's a variable that's being referenced (could be a function, etc,
etc) then do whatever you want to do with that VarDecl.

On Tue, Dec 3, 2019 at 1:33 PM Robert Ankeney via cfe-users <
cfe-users at lists.llvm.org> wrote:

> Let's assume I have a DeclRefExpr that is referenceing some variable that
> is declared as "static const int var;". How can I determine the variable is
> static? There is no isStaticLocal() or getStorageClass() call as there
> would be for the VarDecl. Is there something within the QualType that gets
> me that info?
>
> Many thanks,
> Robert
>
> _______________________________________________
> cfe-users mailing list
> cfe-users at lists.llvm.org
> https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-users
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/cfe-users/attachments/20191203/00175857/attachment.html>

From cfe-users at lists.llvm.org  Tue Dec  3 11:11:17 2019
From: cfe-users at lists.llvm.org (Robert Ankeney via cfe-users)
Date: Tue, 3 Dec 2019 11:11:17 -0800
Subject: [cfe-users] DeclRefExpr question
In-Reply-To: <CAENS6EuzTKxRPmpCXLHmJWppn0h-7U-FRPo1Yc3VcEyj1Mjo6Q@mail.gmail.com>
References: <CAJL6GZpXSV8J2OTWU2VDPRNVYqvi19k56C7bn-VFxuq4uZJNOQ@mail.gmail.com>
 <CAENS6EuzTKxRPmpCXLHmJWppn0h-7U-FRPo1Yc3VcEyj1Mjo6Q@mail.gmail.com>
Message-ID: <CAJL6GZqVZzdkQs9xZ9PFFMAYrHH+QMfNf5eHsHLGhCzNV1uWOQ@mail.gmail.com>

Worked great! Many thanks!

Robert


On Tue, Dec 3, 2019 at 10:43 AM David Blaikie <dblaikie at gmail.com> wrote:

> I'd guess you'd call "getDecl" to get the decl the DeclRefExpr is
> referencing, then dyn_cast that (the decl is a ValueDecl*) to VarDecl to
> see if it's a variable that's being referenced (could be a function, etc,
> etc) then do whatever you want to do with that VarDecl.
>
> On Tue, Dec 3, 2019 at 1:33 PM Robert Ankeney via cfe-users <
> cfe-users at lists.llvm.org> wrote:
>
>> Let's assume I have a DeclRefExpr that is referenceing some variable that
>> is declared as "static const int var;". How can I determine the variable is
>> static? There is no isStaticLocal() or getStorageClass() call as there
>> would be for the VarDecl. Is there something within the QualType that gets
>> me that info?
>>
>> Many thanks,
>> Robert
>>
>> _______________________________________________
>> cfe-users mailing list
>> cfe-users at lists.llvm.org
>> https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-users
>>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/cfe-users/attachments/20191203/12efc8be/attachment.html>

From cfe-users at lists.llvm.org  Tue Dec 10 22:45:08 2019
From: cfe-users at lists.llvm.org (henry ding via cfe-users)
Date: Wed, 11 Dec 2019 14:45:08 +0800
Subject: [cfe-users] got wrong parameter type using clang
Message-ID: <CAOS0VuE6Yh6e34SbZ=KbCsPXijC0SV71zCqoVu+BSg-1rfEEkQ@mail.gmail.com>

simple test.cpp:
void test(std::string xx,string bb,int aa){}

clang usage:
int getFuncInfo(Rewriter* TheRewrite,clang::ASTContext *Context, const
FunctionDecl *func,FuncNode &node){
    clang::LangOptions LangOpts;
    LangOpts.CPlusPlus = true;
    Policy.FullyQualifiedName = 1;
   .............................................................
  . ..........................................................
   for(unsigned int i=0; i<func->getNumParams(); i++)
           {
                 string t =
QualType::getAsString(func->getParamDecl(i)->getType().split(),Policy);
                string n =
func->getParamDecl(i)->getName();//func->parameters()[i]->getQualifiedNameAsString()
                 cout << t + "#" + n << endl;
         }
}

But,i got:
int#xx  //should string#xx
int#bb //should string#bb
int#aa

Has anyone encountered this situation ？

From cfe-users at lists.llvm.org  Wed Dec 11 10:13:14 2019
From: cfe-users at lists.llvm.org (Richard Smith via cfe-users)
Date: Wed, 11 Dec 2019 10:13:14 -0800
Subject: [cfe-users] got wrong parameter type using clang
In-Reply-To: <CAOS0VuE6Yh6e34SbZ=KbCsPXijC0SV71zCqoVu+BSg-1rfEEkQ@mail.gmail.com>
References: <CAOS0VuE6Yh6e34SbZ=KbCsPXijC0SV71zCqoVu+BSg-1rfEEkQ@mail.gmail.com>
Message-ID: <CAOfiQqniCNnOG1Rr10NZ9i6N6W2iKj0M1GwHZ=w1zBw7cSO2BQ@mail.gmail.com>

On Tue, 10 Dec 2019, 22:45 henry ding via cfe-users, <
cfe-users at lists.llvm.org> wrote:

> simple test.cpp:
> void test(std::string xx,string bb,int aa){}
>
> clang usage:
> int getFuncInfo(Rewriter* TheRewrite,clang::ASTContext *Context, const
> FunctionDecl *func,FuncNode &node){
>     clang::LangOptions LangOpts;
>     LangOpts.CPlusPlus = true;
>     Policy.FullyQualifiedName = 1;
>    .............................................................
>   . ..........................................................
>    for(unsigned int i=0; i<func->getNumParams(); i++)
>            {
>                  string t =
> QualType::getAsString(func->getParamDecl(i)->getType().split(),Policy);
>                 string n =
>
> func->getParamDecl(i)->getName();//func->parameters()[i]->getQualifiedNameAsString()
>                  cout << t + "#" + n << endl;
>          }
> }
>
> But,i got:
> int#xx  //should string#xx
> int#bb //should string#bb
> int#aa
>
> Has anyone encountered this situation ?


The bug is probably somewhere else. Make sure the proper language options
are set up before parsing the source file, and that you're printing out any
diagnostics. My guess would be that the source is failing to compile (wrong
options, or can't find some standard header, or something), and the
diagnostics are being discarded, and clang is replacing the parameter type
with 'int' in error recovery.

_______________________________________________
> cfe-users mailing list
> cfe-users at lists.llvm.org
> https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-users
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/cfe-users/attachments/20191211/019948b8/attachment.html>

From cfe-users at lists.llvm.org  Thu Dec 12 20:18:04 2019
From: cfe-users at lists.llvm.org (henry ding via cfe-users)
Date: Fri, 13 Dec 2019 12:18:04 +0800
Subject: [cfe-users] got wrong parameter type using clang
In-Reply-To: <CAOfiQqniCNnOG1Rr10NZ9i6N6W2iKj0M1GwHZ=w1zBw7cSO2BQ@mail.gmail.com>
References: <CAOS0VuE6Yh6e34SbZ=KbCsPXijC0SV71zCqoVu+BSg-1rfEEkQ@mail.gmail.com>
 <CAOfiQqniCNnOG1Rr10NZ9i6N6W2iKj0M1GwHZ=w1zBw7cSO2BQ@mail.gmail.com>
Message-ID: <CAOS0VuG6KVT0s-TSeMQFfApWWr1R=kvGXjFjQLk73=r6wyZ+XA@mail.gmail.com>

l found  there were a typedecl which typedef  string as int, this may
be the reason

std::string|int|/usr/include/c++/4.8.5/bits/stringfwd.h


2019-12-12 2:13 GMT+08:00, Richard Smith <richard at metafoo.co.uk>:
> On Tue, 10 Dec 2019, 22:45 henry ding via cfe-users, <
> cfe-users at lists.llvm.org> wrote:
>
>> simple test.cpp:
>> void test(std::string xx,string bb,int aa){}
>>
>> clang usage:
>> int getFuncInfo(Rewriter* TheRewrite,clang::ASTContext *Context, const
>> FunctionDecl *func,FuncNode &node){
>>     clang::LangOptions LangOpts;
>>     LangOpts.CPlusPlus = true;
>>     Policy.FullyQualifiedName = 1;
>>    .............................................................
>>   . ..........................................................
>>    for(unsigned int i=0; i<func->getNumParams(); i++)
>>            {
>>                  string t =
>> QualType::getAsString(func->getParamDecl(i)->getType().split(),Policy);
>>                 string n =
>>
>> func->getParamDecl(i)->getName();//func->parameters()[i]->getQualifiedNameAsString()
>>                  cout << t + "#" + n << endl;
>>          }
>> }
>>
>> But,i got:
>> int#xx  //should string#xx
>> int#bb //should string#bb
>> int#aa
>>
>> Has anyone encountered this situation ?
>
>
> The bug is probably somewhere else. Make sure the proper language options
> are set up before parsing the source file, and that you're printing out any
> diagnostics. My guess would be that the source is failing to compile (wrong
> options, or can't find some standard header, or something), and the
> diagnostics are being discarded, and clang is replacing the parameter type
> with 'int' in error recovery.
>
> _______________________________________________
>> cfe-users mailing list
>> cfe-users at lists.llvm.org
>> https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-users
>>
>

From cfe-users at lists.llvm.org  Sat Dec 14 09:55:04 2019
From: cfe-users at lists.llvm.org (Xuanyi Zhou via cfe-users)
Date: Sat, 14 Dec 2019 12:55:04 -0500
Subject: [cfe-users] Checking if a class can be copy constructed
Message-ID: <5df5222b.1c69fb81.dfbaa.79ab@mx.google.com>

An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/cfe-users/attachments/20191214/e4b0ff07/attachment.html>

From cfe-users at lists.llvm.org  Sat Dec 14 16:31:26 2019
From: cfe-users at lists.llvm.org (Luke Zhou via cfe-users)
Date: Sat, 14 Dec 2019 19:31:26 -0500
Subject: [cfe-users] Checking if a class can be copy constructed
Message-ID: <CALxpuRKCMU4RhT+wYH4y9VJ9E4cR++cmPRA7aiNDFhjfDOXFQQ@mail.gmail.com>

Hello again! It seems that the system failed to extract text from my
previous post. This is my first time using a mailing list, so please excuse
me for my mistakes. Here are the original contents of that post:

Hello,
>
> I’ve been trying to do some code generation using clang (not libclang),
> and want to find out if a class can be copy constructed.
>
> I’ve tried testing if the copy constructor is deleted, but this does not
> work well with templates. For example, I have a class that contains a
> std::vector<std::unique_ptr>, whose copy constructor will apparently fail
> to instantiate, but doesn’t cause the copy constructors of std::vector or
> my class to be marked as deleted. Also, isInvalidDecl() returns false,
> which (I think) means that the compiler didn’t even attempt to instantiate
> the constructor as it has never been used.
>
> I believe that my best bet is to try to manually check if the constructor
> can instantiate. I’ve looked at clang::Sema, but was overwhelmed by the
> large amount of functions. I also noticed that a lot of these functions
> seem to have side effects, which is undesired since I do not want this
> check to cause other parts of the code to not compile.
>
> Am I on the right path? If so, which functions in clang::Sema should I
> use, and how? If not, is there a better way?
>
> Thank you very much.
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/cfe-users/attachments/20191214/8718ad31/attachment.html>

From cfe-users at lists.llvm.org  Wed Dec 11 12:47:28 2019
From: cfe-users at lists.llvm.org (Luke Zhou via cfe-users)
Date: Wed, 11 Dec 2019 15:47:28 -0500
Subject: [cfe-users] Checking if a class can be copy constructed
Message-ID: <CALxpuR+xWuC3mUaD6z1EBdCTTJjxLdaoWVyPoft3015FvTK8Pw@mail.gmail.com>

Hello,


I’ve been trying to do some code generation using clang (not libclang), and
want to find out if a class can be copy constructed.


I’ve tried testing if the copy constructor is deleted, but this does not
work well with templates. For example, I have a class that contains a
std::vector<std::unique_ptr>, whose copy constructor will apparently fail
to instantiate, but doesn’t cause the copy constructors of std::vector or
my class to be marked as deleted. Also, isInvalidDecl() returns false,
which (I think) means that the compiler didn’t even attempt to instantiate
the constructor as it has never been used.


I believe that my best bet is to try to manually check if the constructor
can instantiate. I’ve looked at clang::Sema, but was overwhelmed by the
large amount of functions. I also noticed that a lot of these functions
seem to have side effects, which is undesired since I do not want this
check to cause other parts of the code to not compile.


Am I on the right path? If so, which functions in clang::Sema should I use,
and how? If not, is there a better way?


Thank you very much.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/cfe-users/attachments/20191211/ffbd2275/attachment.html>

From cfe-users at lists.llvm.org  Sun Dec  8 08:35:50 2019
From: cfe-users at lists.llvm.org (Victor via cfe-users)
Date: Sun, 8 Dec 2019 21:35:50 +0500
Subject: [cfe-users] Clang and CDO, CDQ x64 instructions
Message-ID: <CAGc6Bq-qW6D9tme64Ln9VDiE7-U=_qz0qjXX++Fcn24Rcazg-g@mail.gmail.com>

Hello Clang users and developers!

I have problem compiling some generated assembly code with clang:
mc.s:35739:2: error: invalid instruction mnemonic 'cdo'

Does Clang support x64 CDO and CDQ instructions?
The code is generated by MLRISC library. A bit older versions used CQTO/CLTD
and that worked.

Invocation options:
clang -g -O2 -D_GNU_SOURCE -pedantic -m64
 -Wno-unused-command-line-argument  mc.s

clang --version:
clang version 9.0.1-+20191205081203+432bf48c08d-1~exp1~20191205191740.102
Target: x86_64-pc-linux-gnu
Thread model: posix
InstalledDir: /usr/bin

Best wishes
Victor
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/cfe-users/attachments/20191208/3cec8017/attachment.html>

From cfe-users at lists.llvm.org  Thu Dec 19 00:19:00 2019
From: cfe-users at lists.llvm.org (Guofeng Zhang via cfe-users)
Date: Thu, 19 Dec 2019 16:19:00 +0800
Subject: [cfe-users] Clang optimization options with out dead code
	elimination?
Message-ID: <CAMpUGn32vag2pVZOA5dmSxH6NOV_Ojcpq6XmT4U2x+i8L4Wm=g@mail.gmail.com>

Hi,

We use clang 8+. Does clang have options (not -O0) that do not remove dead
code when clang optimize the code?

We have a old project, now migrating to Clang. It works well with -O0, but
failed when use -O1 to -O3.

Thanks,

Guofeng
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/cfe-users/attachments/20191219/ad8f28d9/attachment.html>

From cfe-users at lists.llvm.org  Thu Dec 19 11:12:21 2019
From: cfe-users at lists.llvm.org (David Blaikie via cfe-users)
Date: Thu, 19 Dec 2019 11:12:21 -0800
Subject: [cfe-users] Clang optimization options with out dead code
	elimination?
In-Reply-To: <CAMpUGn32vag2pVZOA5dmSxH6NOV_Ojcpq6XmT4U2x+i8L4Wm=g@mail.gmail.com>
References: <CAMpUGn32vag2pVZOA5dmSxH6NOV_Ojcpq6XmT4U2x+i8L4Wm=g@mail.gmail.com>
Message-ID: <CAENS6EvCSSZfDQELcRxYv_DWV8ESKSLnXHcApJU9UN5ORaOkCA@mail.gmail.com>

There are lots of reasons a program might fail to do what's desired when
optimized - dead code is only one of them, any reason to believe that's the
issue at hand? But no, LLVM doesn't really have any support for that.

You might try running your program using the various sanitizersn (address,
memory, undefined behavior sanitizer) to help identify if there are bugs
that could be fixed.

On Thu, Dec 19, 2019 at 12:19 AM Guofeng Zhang via cfe-users <
cfe-users at lists.llvm.org> wrote:

> Hi,
>
> We use clang 8+. Does clang have options (not -O0) that do not remove dead
> code when clang optimize the code?
>
> We have a old project, now migrating to Clang. It works well with -O0, but
> failed when use -O1 to -O3.
>
> Thanks,
>
> Guofeng
>
> _______________________________________________
> cfe-users mailing list
> cfe-users at lists.llvm.org
> https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-users
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/cfe-users/attachments/20191219/d683662f/attachment.html>