The windows code runs the following code, and there is a FAIL check �to see if there is a duplicate file name based on case sensitivity.�
sqInt sqFileDeleteNameSize(char* fileNameIndex, sqInt fileNameSize) {
� WCHAR win32Path[MAX_PATH+1];
� int sz;
� /* convert the file name into a null-terminated C string */
� sz = MultiByteToWideChar(CP_UTF8, 0, fileNameIndex, fileNameSize, NULL, 0);
� if(sz > MAX_PATH) FAIL();
� MultiByteToWideChar(CP_UTF8, 0, fileNameIndex, fileNameSize, win32Path, sz);
� win32Path[sz] = 0;
� if(hasCaseSensitiveDuplicate(win32Path)) FAIL();
� if(!DeleteFileW(win32Path)) FAIL();
� return 1;
}
On 2010-01-25, at 2:07 PM, Mariano Martinez Peck wrote:
Hi folks. I don't want other than green tests in dev images, thus, I started to fix them as much as possible. There are a lot of failing tests of Nile (only fail in windows). The test is NSCompressedSourceStreamTest� and I noticed something weird.� If you look to tearDown method, it tries to remove two files that it actually were created in the setUp method.
But it is weird because it tries to remove a file that actually exist and it is not being removed.
If I debug in DosFileDirectory >> deleteFileNamed: localFileName ifAbsent: failBlock
everything looks ok, but the file is not removed and the failBlock is executed :(
localFileName ->� 'testCompressedSourceStream.stc'
(self fullNameFor: localFileName) asVmPathName� -> 'C:\mariano\squeak\imagenes\Pharo1.0-10508-rc2dev10.01.2\testCompressedSourceStream.stc'
And that file exist, as you can see in screenshot that I attach.
Can someone help me with this ?
Cheers
Mariano
Pharo-project mailing list
Pharo-project@lists.gforge.inria.fr
http://lists.gforge.inria.fr/cgi-bin/mailman/listinfo/pharo-project
--
===========================================================================
===========================================================================