Make kill_python a little more forgiving if it can't obtain a snapshot of module information for a given python[_d].exe process. Failing here was too pessimistic; the python[_d].exe process may be owned by another user, which is the case in some buildbot environments.

This commit is contained in:
Trent Nelson 2008-04-03 20:00:08 +00:00
parent 8d69c1f95f
commit 2aae1d92eb
1 changed files with 9 additions and 5 deletions

View File

@ -118,11 +118,15 @@ main(int argc, char **argv)
/* It's a python process, so figure out which directory it's in... */ /* It's a python process, so figure out which directory it's in... */
hsm = CreateToolhelp32Snapshot(TH32CS_SNAPMODULE, pe.th32ProcessID); hsm = CreateToolhelp32Snapshot(TH32CS_SNAPMODULE, pe.th32ProcessID);
if (hsm == INVALID_HANDLE_VALUE) { if (hsm == INVALID_HANDLE_VALUE)
printf("CreateToolhelp32Snapshot[3] failed: %d\n", GetLastError()); /*
CloseHandle(hsp); * If our module snapshot fails (which will happen if we don't own
return 1; * the process), just ignore it and continue. (It seems different
} * versions of Windows return different values for GetLastError()
* in this situation; it's easier to just ignore it and move on vs.
* stopping the build for what could be a false positive.)
*/
continue;
if (!Module32FirstW(hsm, &me)) { if (!Module32FirstW(hsm, &me)) {
printf("Module32FirstW[2] failed: %d\n", GetLastError()); printf("Module32FirstW[2] failed: %d\n", GetLastError());