Added initialization and shuthown fo ecore_file library and replaced realpath function to ecore_file_realpath.
Created attachment 92765 [details] Patch
Comment on attachment 92765 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=92765&action=review > Tools/ChangeLog:8 > + Added initialization and shuthown fo ecore_file library and replaced s/fo/for/ > Tools/EWebLauncher/main.c:802 > - rpath = realpath(theme, NULL); > + rpath = ecore_file_realpath(theme); > if (!rpath) > return NULL; ecore_file_realpath always allocate a string, even when realpath would return an error (exception: when there's an out of memory condition). In this case you need to check whether the returned string is empty. > Tools/EWebLauncher/main.c:888 > + return quit(EINA_FALSE, "ERROR: could not created tmp directory.\n"); s/created/create/
Created attachment 92784 [details] Fixed patch
Comment on attachment 92784 [details] Fixed patch View in context: https://bugs.webkit.org/attachment.cgi?id=92784&action=review > Tools/EWebLauncher/main.c:800 > - rpath = realpath(theme, NULL); > - if (!rpath) > - return NULL; > - > - if (stat(rpath, &st)) { > + rpath = ecore_file_realpath("/etc"); /etc ???? Why? Maybe you used as a test and let it here?
Created attachment 92923 [details] Patch 3
It looks OK to me, informal r+ from my side.
rpath is not a parameter of ecore_file_realpath, right? rpath = ecore_file_realpath(theme);
Created attachment 93084 [details] Patch 4
(In reply to comment #7) > rpath is not a parameter of ecore_file_realpath, right? Right. Tomasz, we can't apply every patch and test when reviewing. Please, test your changes to see if it really fixes the issue before uploading a new patch. Now it looks good.
Comment on attachment 93084 [details] Patch 4 Clearing flags on attachment: 93084 Committed r86349: <http://trac.webkit.org/changeset/86349>
All reviewed patches have been landed. Closing bug.
http://trac.webkit.org/changeset/86349 might have broken GTK Linux 32-bit Release