On Wed, Jun 10, 2009 at 10:28:00AM -0500, Doug Goldstein wrote: ACK, this looks good Daniel -- |: Red Hat, Engineering, London -o- http://people.redhat.com/berrange/ :| |: http://libvirt.org -o- http://virt-manager.org -o- http://ovirt.org :| |: http://autobuild.org -o- http://search.cpan.org/~danberr/ :| |: GnuPG: 7D3B9505 -o- F3C9 553F A1DA 4AC2 5648 23C1 B3DF F742 7D3B 9505 :|

2009/6/10 Doug Goldstein <cardoe(a)gentoo.org&gt;: +/* + * Finds a requested file in the PATH env. e.g.: + * "kvm-img" will return "/usr/bin/kvm-img" + * + * You must free the result + */ +char *virFindFileInPath(const char *file) +{ + char pathenv[PATH_MAX]; + char *penv = &pathenv; /* this is for glibc 2.10 strsep chnages */ + char *pathseg; + char fullpath[PATH_MAX]; + + /* copy PATH env so we can tweak it */ + strncpy(pathenv, getenv("PATH"), PATH_MAX); + pathenv[PATH_MAX - 1] = '\0'; + + /* for each path segment, append the file to search for and test for + * it. return it if found. + */ + while ((pathseg = strsep(&penv, ":")) != NULL) { + snprintf(fullpath, PATH_MAX, "%s/%s", pathseg, file); + if (virFileExists(fullpath)) + return strdup(fullpath); + } + + return NULL; +} GCC (version 4.3.3 here) warns about "initialization from incompatible pointer type" at this line: char *penv = &pathenv; /* this is for glibc 2.10 strsep chnages */ Changing it to char *penv = pathenv; /* this is for glibc 2.10 strsep chnages */ fixes the warning. At first I was surprised that the first version with & works and gives correct results, then I checked that actual values that are assigned to char *penv and it's the same in both cases. But the point is: the additional & triggers a warning (breaks compilation with --enable-compile-warnings=error) and removing it doesn't change the behavior. PS: What's this comment about changes in glibc 2.10 to strsep() referring to? I looked at the changelog but couldn't find any recent changes to strsep(). Regards, Matthias

On Fri, Jun 12, 2009 at 10:53:01PM +0200, Maximilian Wilhelm wrote: [...] Hum, not for the main list of course. Somehow I don't find mail to be the best way to carry the informations, something like a set of web pages for the runs in the last couple of days would certainly be more convenient. One way could be to send the output by email and aggregate it somewhere, or export the last result on a web page and I could do the aggregation. But I know some people compile with -Werror on a regular basis so in any case this kind of problems should be caught quickly, maybe it's not worth developping a big setup... Daniel -- Daniel Veillard | libxml Gnome XML XSLT toolkit http://xmlsoft.org/ daniel(a)veillard.com | Rpmfind RPM search engine http://rpmfind.net/ http://veillard.com/ | virtualization library http://libvirt.org/