[libvirt] [PATCH] vmware: os x support is broken

[Doug Goldstein]

On Fri, Jan 3, 2014 at 12:14 PM, Eric Blake <eblake at redhat.com> wrote:
> On 01/03/2014 10:57 AM, Denis Kondratenko wrote:
>> Incorrect usage of virAsprintf and vmware-vmx reports to stderr.
>>
>> --- https://bugzilla.redhat.com/show_bug.cgi?id=1036248
>
> Oh my - the triple --- leadin confused 'git am':
>
> Applying: vmware: os x support is broken
> fatal: patch fragment without header at line 42: @@ -271,17 +271,17 @@
> vmwareExtractVersion(struct vmware_driver
> *driver)</div><div> </div><div>     switch (driver->type)
> {</div><div>         case VMWARE_DRIVER_PLAYER:</div><div>-
>  if (virAsprintf(&bin, "%s/%s", vmwarePath,
> "vmplayer"))</div>
> Repository lacks necessary blobs to fall back on 3-way merge.
> Cannot fall back to three-way merge.
> Patch failed at 0001 vmware: os x support is broken
>
> so I had to apply it by hand.
>
>
>>          case VMWARE_DRIVER_PLAYER:
>> -            if (virAsprintf(&bin, "%s/%s", vmwarePath, "vmplayer"))
>> +            if (virAsprintf(&bin, "%s/%s", vmwarePath, "vmplayer") < 0)
>
> Definitely correct!
>
>
>>      cmd = virCommandNewArgList(bin, "-v", NULL);
>>      virCommandSetOutputBuffer(cmd, &outbuf);
>> +
>> +    // OS X 10.9.1 and some earlier ver: vmware-vmx reports ver to stderr
>
> We prefer to avoid C99 comments, and the code is self-explanatory enough
> that it was easier to just drop the comment (and put it in the commit
> message instead).

Sadly the comment actually doesn't make any sense. vmware-vmx is a
program shipped by VMware and doesn't select stdout or stderr based on
OS X version. It selects it based on whether the arguments supplied
are valid or not for the command. If they aren't valid then its stderr
(like most UNIX utilities) and if they're valid it goes via stdout. I
believe we had to use a bad command (and thus stderr) on OS X always
to get the version info as nothing that supplied the version returned
a successful.

Denis,

I don't own a copy of VMware Fusion (or any VMware product) so I can't
personally test. I have a number of branches which have some potential
fixes (I believe 1 series has been posted RFC without a response) but
don't at the moment have anyone to test them. So if you're interested
in testing them I can provide you with an alternate brew install
command and we can maybe land some other fixes.

-- 
Doug Goldstein