[cfe-commits] r169830 - /cfe/trunk/lib/Serialization/ASTReader.cpp

Argyrios Kyrtzidis akyrtzi at gmail.com
Thu Dec 13 16:48:49 PST 2012


On Dec 13, 2012, at 4:02 PM, Galina Kistanova <gkistanova at gmail.com> wrote:

> Hi Argyrios,
> 
> One of these commits (169830, 169831) added failing test to
> clang-X86_64-freebsd builder:
> Clang :: Misc/dev-fd-fs.c failed
> 
> The builder is here:
> http://lab.llvm.org:8011/builders/clang-X86_64-freebsd
> 
> For r 169829 all tests pass:
> http://lab.llvm.org:8011/builders/clang-X86_64-freebsd/builds/6142
> 
> Please have a look at this?

r169830 is unrelated and I don't see what the issue would be with r169831.

r169831 changed the way we stat, in that if a file descriptor is not needed, then we call 'stat' directly instead of opening the file and then calling fstat on it.
Does this not work with named pipes on freebsd ?

> 
> Thanks
> 
> Galina
> 
> 
> 
> On Mon, Dec 10, 2012 at 11:48 PM, Argyrios Kyrtzidis <akyrtzi at gmail.com> wrote:
>> Author: akirtzidis
>> Date: Tue Dec 11 01:48:14 2012
>> New Revision: 169830
>> 
>> URL: http://llvm.org/viewvc/llvm-project?rev=169830&view=rev
>> Log:
>> [PCH] Remove redundant stat call when validating the input files.
>> 
>> This became redundant after we removed the stat cache.
>> 
>> Modified:
>>    cfe/trunk/lib/Serialization/ASTReader.cpp
>> 
>> Modified: cfe/trunk/lib/Serialization/ASTReader.cpp
>> URL: http://llvm.org/viewvc/llvm-project/cfe/trunk/lib/Serialization/ASTReader.cpp?rev=169830&r1=169829&r2=169830&view=diff
>> ==============================================================================
>> --- cfe/trunk/lib/Serialization/ASTReader.cpp (original)
>> +++ cfe/trunk/lib/Serialization/ASTReader.cpp Tue Dec 11 01:48:14 2012
>> @@ -53,7 +53,6 @@
>> #include <algorithm>
>> #include <cstdio>
>> #include <iterator>
>> -#include <sys/stat.h>
>> 
>> using namespace clang;
>> using namespace clang::serialization;
>> @@ -1560,20 +1559,12 @@
>>     if (Overridden)
>>       return InputFile(File, Overridden);
>> 
>> -    // The stat info from the FileEntry came from the cached stat
>> -    // info of the PCH, so we cannot trust it.
>> -    struct stat StatBuf;
>> -    if (::stat(File->getName(), &StatBuf) != 0) {
>> -      StatBuf.st_size = File->getSize();
>> -      StatBuf.st_mtime = File->getModificationTime();
>> -    }
>> -
>> -    if ((StoredSize != StatBuf.st_size
>> +    if ((StoredSize != File->getSize()
>> #if !defined(LLVM_ON_WIN32)
>>          // In our regression testing, the Windows file system seems to
>>          // have inconsistent modification times that sometimes
>>          // erroneously trigger this error-handling path.
>> -         || StoredTime != StatBuf.st_mtime
>> +         || StoredTime != File->getModificationTime()
>> #endif
>>          )) {
>>       if (Complain)
>> 
>> 
>> _______________________________________________
>> cfe-commits mailing list
>> cfe-commits at cs.uiuc.edu
>> http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits





More information about the cfe-commits mailing list