Aha, interesting, I'd forgotten about that.
I will see if I can cleanly cherry-pick those commits. If I can do so without
breaking anything (and double checking that the changes don't alter any public
APIs), I'll merge it.
The other alternative is that you could pick a recent tag in the master branch,
where this has already been fixed. Any place where I have tagged it (e.g.,
"Release-1.5.6dev") will be pretty solid, and we're very close to branching a
stable 1.5 anyway.
-- lg
On Nov 28, 2014, at 1:16 PM, Justin Israel <[email protected]> wrote:
>
> I spent some time looking at this, and I realized that you had actually done
> some work on the error reporting, but they aren't picked into the 1.4 branch:
> https://github.com/OpenImageIO/oiio/commit/73af64b64867785e8c1ccca1ad3a8dff87777f0e
> https://github.com/OpenImageIO/oiio/commit/65eeee465edf030867ed7468632a5e6dcadfbb5b
>
> I see that you now have an error() function that exits non-zero, and it gets
> called in the output logic when there is a write error. Can this stuff get
> merged into 1.4?
>
>
>
> On Sat Nov 29 2014 at 6:58:12 AM Larry Gritz <[email protected]> wrote:
> If you look in oiiotool.cpp, in action_diff, you'll see something like:
>
> if (ret != DiffErrOK && ret != DiffErrWarn)
> ot.return_value = EXIT_FAILURE;
>
> This is the value that's eventually returned when oiiotool exits.
>
> There may be other places, and other error conditions, where we should be
> doing this.
>
>
> On Nov 28, 2014, at 2:44 AM, Justin Israel <[email protected]> wrote:
>
>> I can definitely take a look at it tomorrow. Just wanted to check the
>> intention. As long as oiiotool exits non-zero when there are write errors,
>> that would be a win.
>>
>>
>> On Fri, 28 Nov 2014 8:36 PM Larry Gritz <[email protected]> wrote:
>> I think returning 0 from this function is fine, but I agree that if some
>> output file is unable to write, a flag should be set so that oiiotool as a
>> whole gives a shell return code indicating an error (i.e., not zero).
>>
>> This is not intentional, and returning an error code so it can be scripted
>> with error checks is a good thing. So it should definitely be fixed.
>>
>> I can do it if you want, or feel free to give it a stab. Let me know which
>> you prefer.
>>
>> -- lg
>>
>>
>>
>> On Nov 27, 2014, at 4:09 PM, Justin Israel <[email protected]> wrote:
>>
>>> I'm curious about some behaviour I am seeing, and whether or not anyone has
>>> encountered this and found it to be problematic?
>>>
>>> oiiotool seems to exit with an exitcode 0 even if the file it was writing
>>> failed:
>>> https://github.com/OpenImageIO/oiio/blob/master/src/oiiotool/oiiotool.cpp#L540
>>>
>>> So what I end up seeing in my render pipeline process is that oiiotool kind
>>> of blows through a sequence "successfully" even if the output directory did
>>> not exist. Is this expected behaviour or a bug? Do most people tack on a
>>> post check that all of the frames are there? I had assumed I could rely on
>>> the exitcode to know if the call to oiiotool failed to do the job.
>>>
>>> Is this something that would break people if it were corrected to return a
>>> non-zero exit code if the write fails?
>>>
>>
>> --
>> Larry Gritz
>> [email protected]
>>
>>
>>
>> _______________________________________________
>> Oiio-dev mailing list
>> [email protected]
>> http://lists.openimageio.org/listinfo.cgi/oiio-dev-openimageio.org
>> _______________________________________________
>> Oiio-dev mailing list
>> [email protected]
>> http://lists.openimageio.org/listinfo.cgi/oiio-dev-openimageio.org
>
> --
> Larry Gritz
> [email protected]
>
>
>
> _______________________________________________
> Oiio-dev mailing list
> [email protected]
> http://lists.openimageio.org/listinfo.cgi/oiio-dev-openimageio.org
> _______________________________________________
> Oiio-dev mailing list
> [email protected]
> http://lists.openimageio.org/listinfo.cgi/oiio-dev-openimageio.org
--
Larry Gritz
[email protected]
_______________________________________________
Oiio-dev mailing list
[email protected]
http://lists.openimageio.org/listinfo.cgi/oiio-dev-openimageio.org
