gmane.comp.compilers.clang.devel

"extern void v"

Jordan Rose — 2013-05-24T16:45:20

PR16131 describes an analyzer assertion on taking the address of an extern void variable:

extern void v;
static void f(void) {
        &v;
}
Clearly the analyzer shouldn't crash, but I'm surprised that this is legal at all. Yet I can't find anything in the standard actually forbidding it—you can certainly do the same thing with any incomplete struct type. Can somebody confirm that this is at least grammatically correct?

(I suppose like an incomplete object it provides a way to name an address without allowing access.)

Jordan_______________________________________________
cfe-dev mailing list
cfe-dev-Tmj1lob9twqVc3sceRu5cw< at >public.gmane.org
http://lists.cs.uiuc.edu/mailman/listinfo/cfe-dev

Re: C99 index function

Rafael Espíndola — 2013-05-24T16:09:56

this reduces to just

--------------------
int index;
------------------

Looks like there two differences in behavior

* gcc doesn't define it with -std=c99, we do
* gcc only produces an warning when a builtin is declare as a
non-function, we produce an error.

It might be reasonable to change clang's behavior to mach gcc in the
-std=c99 case.  Can you open a bug report to track this?

Cheers,
Rafael

Can not find .a file from Clang scan-build

凌欢 — 2013-05-24T10:10:45

hi, all:
One of my iOS project throw the following errors when I execute the scan-build : 

clang: error: no such file or directory: '/Users/mqq/hudson/clang_browser_newcore_ipad/tools/WebPageTest/WebPageTest/build/Release-iphoneos/libWebPageTest.a'
clang: error: no such file or directory: '/Users/mqq/hudson/clang_browser_newcore_ipad/MttHD_shell/Frameworks/KIF/build/Release-iphoneos/libKIF.a'
clang: error: no such file or directory: '/Users/mqq/hudson/clang_browser_newcore_ipad/xcodebuild/Release-iphoneos/libhttp_server.a'
 
At the same time ,my another script can build successfully which is not added scan-build, so is it the bug for scan-build ?
Inclosed  is the scan-build log the scripts.
Can anynone give me tips?
 BRs
Howard Ling
 _______________________________________________
cfe-dev mailing list
cfe-dev-Tmj1lob9twqVc3sceRu5cw< at >public.gmane.org
http://lists.cs.uiuc.edu/mailman/listinfo/cfe-dev

C99 index function

Albert Netymk — 2013-05-24T09:33:17

Hello,

When I use clang to compile this file:
```
#include <stdio.h>

int index;
int main(int argc, char **argv)
{
    puts("hello");
    return 0;
}
```
I got the error: redefinition of 'index' as different kind of symbol.

It turns out that `index` is one deprecated one function in posix.

Using `gcc -std=c99` fixes the problem, but `clang -std=c99` doesn't do the
trick. Any explanation?

Re: Determining whether Container()==Container()can compile?

Mailing List Email — 2013-05-24T13:21:28

It can be done yourself, or am I misunderstanding your question?
Example:

#include <iostream>
#include <vector>
#include <type_traits>

template<typename T, typename U>
struct IsEqComparable
{
private:
    typedef struct { char dmy; } yes;
    typedef struct { char dmy[2]; } no;

    template<typename T2, typename U2>
    static auto test(void*) -> typename std::enable_if<std::is_same<
            decltype(std::declval<T2>() == std::declval<U2>()),
            decltype(std::declval<T2>() == std::declval<U2>())
        >::value, yes>::type;

    template<typename T2, typename U2> static no test(...);
public:
    static const bool value = sizeof(test<T, U>(nullptr)) == sizeof(yes);
};

struct bar {};

int main()
{
    if (IsEqComparable<std::vector<int>, std::vector<int>>::value)
        std::cout << "Yes!\n";
    else
        std::cout << "No!\n";
    if (IsEqComparable<std::vector<int>, bar>::value)
        std::cout << "Yes!\n";
    else
        std::cout << "No!\n";
    return 0;
}


On 24 May 2013 10:41,

Re: [Static Analyzer]How to track Dead memory Region ofC++ objects

Karthik Bhat — 2013-05-24T09:14:34

Thanks Gabor.
I was able to achieve the required functionality in checkDeadSymbols by
using SymbolReaper's isLiveRegion function.



On Fri, May 24, 2013 at 12:20 PM, Gábor Kozár <kozargabor-Re5JQEeQqe8AvxtiuMwx3w< at >public.gmane.org> wrote:

_______________________________________________
cfe-dev mailing list
cfe-dev-Tmj1lob9twqVc3sceRu5cw< at >public.gmane.org
http://lists.cs.uiuc.edu/mailman/listinfo/cfe-dev

Determining whether Container()==Container() cancompile?

Stephen Kelly — 2013-05-24T08:41:09

Hi there,

Code such as 

  std::vector<A> veca, vecb;
  veca == vecb;

can only compile if A()==A() can compile, if you'll forgive me taking some 
descriptive shortcuts.

I tried writing a template to determine if a type can be equality-compared 
for the purpose of type-erasure:

 http://thread.gmane.org/gmane.comp.lib.qt.devel/11120

Adding auto and trailing return type seems to be helpful:

 http://thread.gmane.org/gmane.comp.lib.qt.devel/11120/focus=11157

Would the suggestion from Olivier Goffart be something that could be applied 
to libcxx to make it possible to write such a template for its stl container 
implementations?

Does anyone else have any other ideas for something that would work with 
todays stl implementations?

Thanks,

Steve.

Re: [Static Analyzer]How to track Dead memory Region ofC++ objects

Gábor Kozár — 2013-05-24T06:50:59

Hi,

What you need is checkDeadSymbols(). In it, iterate through all your
tracked MemRegion-s, and ask the SymbolReaper whether they are dead (I
can't remember the name of the method, but it should be trivial).

Symbols are created when the Static Analyzer has no knowledge of that
particular value. You were correct in trying to work with MemRegion-s.

Gabor


2013/5/24 Karthik Bhat <blitz.opensource-Re5JQEeQqe8AvxtiuMwx3w< at >public.gmane.org>

_______________________________________________
cfe-dev mailing list
cfe-dev-Tmj1lob9twqVc3sceRu5cw< at >public.gmane.org
http://lists.cs.uiuc.edu/mailman/listinfo/cfe-dev

DeclRefExpr - function call arguments

yjninja — 2013-05-24T06:09:25

Hi All,I have a got declrefexpr of the form 'change_value(x)', where I find
the references to 'x'I am writing a tool to change 'x' to 'y'.But, I am not
able to find a way to identify 'x' field in the DeclRefExpr, which only
shows 'change_value' when I print the characters within the sourcerange of
the DeclRefexpr.Can anyone help me with this asap ? I only need the 'x'
portion of 'change_value(x)'.Much Appreciated,..Yogi.



--
View this message in context: http://clang-developers.42468.n3.nabble.com/DeclRefExpr-function-call-arguments-tp4032294.html
Sent from the Clang Developers mailing list archive at Nabble.com.

[Static Analyzer]How to track Dead memory Region of C++objects

Karthik Bhat — 2013-05-24T05:02:42

Hi,
I'm developing a checker for internal purpose were i need to track when Mem
region corresponding to a class object gets invalidated. For example in the
code below

class A {
public:
void myConstruct() {};
};

void fun() {
  A a;
  a. Construct();
  return;
}

int main() {
  fun();
  return 0;
}

the memoryRegion for class A's instance a gets invalidated as soon as we
return from fun() but i'm currently not getting checkRegionChanges
corresponding to this event. ( I'm using getCXXThisVal().getAsRegion() to
get the MemRegion corresponding to instance a.

I also tried to using symbol corresponding to it (i.e
getCXXThisVal().getAsSymbol()) but it always returns NULL.

Could anyone guide me how can i track that MemRegion corresponding to a has
been invalidated after the fun returns?

Thanks
Karthik
_______________________________________________
cfe-dev mailing list
cfe-dev-Tmj1lob9twqVc3sceRu5cw< at >public.gmane.org
http://lists.cs.uiuc.edu/mailman/listinfo/cfe-dev

Re: Clang trunk dev binaries

Kim Gräsman — 2013-05-23T20:42:41

Ben,

On Thu, May 23, 2013 at 10:23 PM, Langmuir, Ben <ben.langmuir-ral2JQCrhuEAvxtiuMwx3w< at >public.gmane.org> wrote:

That's the one! I thought it only had libclang-the-C-wrapper parts,
but it's the full bundle.

Thank you!

- Kim

Re: [PATCH] Fix printing CXXTemporaryObjectExpr withdefault args

Will Wilson — 2013-05-23T20:27:23

Sorry, wrong ML - I've reposed on cfe-commits.

- Will.
_______________________________________________
cfe-dev mailing list
cfe-dev-Tmj1lob9twqVc3sceRu5cw< at >public.gmane.org
http://lists.cs.uiuc.edu/mailman/listinfo/cfe-dev

[PATCH] Fix printing CXXTemporaryObjectExpr with defaultargs

Will Wilson — 2013-05-23T20:23:55

Hi All,

Attached is a patch for a bug in the printing of CXXTemporaryObjectExpr
whereby an extraneous comma will appear before an (implicit) defaulted
argument.

Test case included and all tests passed locally.

Cheers,
Will.
_______________________________________________
cfe-dev mailing list
cfe-dev-Tmj1lob9twqVc3sceRu5cw< at >public.gmane.org
http://lists.cs.uiuc.edu/mailman/listinfo/cfe-dev

Re: Clang trunk dev binaries

Langmuir, Ben — 2013-05-23T20:23:25

Sorry, I misunderstood.  You probably want the "libclang-dev" package then.

Ben

Re: Clang trunk dev binaries

Kim Gräsman — 2013-05-23T19:59:15

Hi Ben,

On Thu, May 23, 2013 at 9:39 PM, Langmuir, Ben <ben.langmuir-ral2JQCrhuEAvxtiuMwx3w< at >public.gmane.org> wrote:

Thanks, those were the ones I found too, but they only contain
executables (clang-3.4) and generated docs (clang-3.4-doc).

I'm looking for headers and static libs to link with a recursive AST visitor.

- Kim

Re: Clang trunk dev binaries

Kim Gräsman — 2013-05-23T19:52:50

Hi there,

On Thu, May 23, 2013 at 9:28 PM, Mailing List Email
<mailinglistx-Re5JQEeQqe8AvxtiuMwx3w< at >public.gmane.org> wrote:

Yes, but it doesn't include the headers/libraries as far as I can tell [1].

Sorry, I could have been clearer: I'm looking for an SDK-style release
with prebuilt libraries and headers. for trunk. llvm-dev is exactly
what I'm looking for, except it's only LLVM, not Clang.

Thanks!

- Kim

[1]
ubuntu< at >ubuntu-VirtualBox:/usr/lib$ dpkg -L clang
/.
/usr
/usr/share
/usr/share/doc
/usr/share/doc/clang
/usr/share/doc/clang/TODO.Debian
/usr/share/doc/clang/copyright
/usr/share/doc/clang/changelog.Debian.gz
/usr/share/man
/usr/share/man/man1
/usr/share/man/man1/clang.1.gz
/usr/share/man/man1/scan-build.1.gz
/usr/share/man/man1/scan-view.1.gz
/usr/share/clang
/usr/share/clang/scan-build
/usr/share/clang/scan-build/ccc-analyzer
/usr/share/clang/scan-build/scan-build
/usr/share/clang/scan-build/scanview.css
/usr/share/clang/scan-build/sorttable.js
/usr/share/clang/scan-build/set-xcode-analyzer
/usr

Re: Clang trunk dev binaries

Langmuir, Ben — 2013-05-23T19:39:42

The clang package from llvm.org/apt is called "clang-3.4". There are also clang-3.4-doc, and a few versions of libclang.

Ben

Re: Clang trunk dev binaries

Mailing List Email — 2013-05-23T19:28:30

apt-get install clang -y?
Last I looked, clang is available in repositories for Ubuntu.


On 23 May 2013 21:20, Kim Gräsman <kim.grasman-Re5JQEeQqe8AvxtiuMwx3w< at >public.gmane.org> wrote:

_______________________________________________
cfe-dev mailing list
cfe-dev-Tmj1lob9twqVc3sceRu5cw< at >public.gmane.org
http://lists.cs.uiuc.edu/mailman/listinfo/cfe-dev

Clang trunk dev binaries

Kim Gräsman — 2013-05-23T19:20:49

Hi all,

I'm running Ubuntu Precise in Virtual Box to do some cross-platform
testing of a Clang-based project.

I can't realistically build LLVM + Clang in that environment, both due
to time and space concerns.

Are there binary releases for trunk somewhere? I'm looking for
something like the "Clang Binaries for Ubuntu-12.04/x86" release at
http://llvm.org/releases/download.html#3.2.

I found this:
http://llvm.org/apt/

which looked promising, but I can only find llvm-*-dev packages with
the LLVM libs and headers, nothing for Clang.

Am I missing something? I'm pretty new to Linux/package managers in
general so I wouldn't be surprised...

Thanks!

- Kim

Re: [PATCH] libc++: help diagnosing the following std::atomic compile error

Stephan Tolksdorf — 2013-05-23T14:26:02

Thanks a lot!

Re: RFC: What does this matcher mean to you?

Vane, Edwin — 2013-05-23T13:01:31

To help stimulate discussion, let me present the two ways this matcher has been interpreted:

1) An "all or nothing" matcher. equalsBoundNode("X") returns true if there exists any previously bound node called "X" and that node satisfies an identity relationship with the node passed to equalsBoundNode. If no such node exists, the matcher just returns false and the match fails, much like any other matcher.
2) As a filter. Consider the matches found by forEach*() as the set of matches. Then the has() matcher containing equalsBoundNode() is used as a filter over this set. Whenever equalsBoundNode() locates a previously bound node with the correct name and satisfying the identity condition, it's really referring to a node in the match set. If the has() matcher fails, the node gets removed from the match set.

The difference can be seen by looking at what matches get returned from either interpretation. With #1, as long as a functionDecl has at least one call statement to a function returning a type that matches a