Re: [libvirt] [PATCH] Fix Win32 portability problems

> +#ifdef WIN32 > +int > +virHookCall(int driver ATTRIBUTE_UNUSED, > + const char *id ATTRIBUTE_UNUSED, > + int op ATTRIBUTE_UNUSED, > + int sub_op ATTRIBUTE_UNUSED, > + const char *extra ATTRIBUTE_UNUSED, > + const char *input ATTRIBUTE_UNUSED) { > + virReportSystemError(ENOSYS, "%s", > + _("spawning hooks not supported on this platform")); > + return -1; Good enough for now. But gnulib supports posix_spawn ported to mingw (currently LGPLv3, so we'd have to get it relaxed to LPGLv2 first); perhaps if we rewrite hooks to use posix_spawn() instead of fork()/exec(), then we can support hooks on mingw.

> @@ -8425,7 +8430,7 @@ cmdSnapshotList(vshControl *ctl, const vshCmd *cmd) > &creation) < 0) > continue; > localtime_r(&creation, &time_info); > - strftime(timestr, sizeof(timestr), "%F %T %z", &time_info); > + strftime(timestr, sizeof(timestr), "%Y-%m-%d %H:%M:%S %z", &time_info); Is this a case where we want localized output? Or is switching to fixed format a good move independently of mingw lacking localization? Gnulib provides strftime (but it is currently LGPLv3, and would need relaxing), if we want to go with localized output.