Author Topic: Bug in mentalray type with mrstand3.10.1-adsk2013  (Read 11848 times)

pwortas

  • Jr. Member
  • **
  • Posts: 3
Bug in mentalray type with mrstand3.10.1-adsk2013
« on: March 22, 2013, 09:54:08 PM »
When attempting to get the job type working on mental ray 3.10, I came across this bug in the mentalray type. This is version 6.4.4.

In MentalrayJob.pm on line 700, the min version for mental ray is pulled from the mi file:
Code: [Select]
if(/^\s*min version\s+"(.*)"/) {
$min_version = $1;
}

In order to handle the new 'filename' keyword in versions of mental ray newer than 3.7, there is a condition:
Code: [Select]
if($min_version >= 3.7) {
# NEW in mental ray 3.7: output file in .mi is now
# specified as in: 'filename "ooga.0001.iff"'
if(/^\s*filename\s+"(.*)"/) {
push @outfiles, $1;
}
}

The problem is that in 3.10, the min version is 3.10, which is less than 3.7 when compared numerically. This causes the code to fall to the sub 3.7 block, which checks the 'output' keyword.

I noticed this failing because the output keyword had colons in it, which broke when attempting to test directory creation on a Windows worker.

Since we know 100% that our workers will be running 3.10, I just modified the type to drop us in the filename block no matter what. Just figured I'd post it here in case anyone else came across a similar problem.

pwortas

  • Jr. Member
  • **
  • Posts: 3
Re: Bug in mentalray type with mrstand3.10.1-adsk2013
« Reply #1 on: March 27, 2013, 07:24:52 PM »
Another issue, mainly stemming from mrstand 3.10, is that ray.exe seems to return 1 on a successful render. This breaks at runRenderCmd in JobTypeLib as the test is expecting 0 and 1 is normally a failed process call. On a success on my end I'm getting an exitcode reported of 256. Since $? is the status word in Perl, you can mod it by 255 ($? & 255) to get the signal. Running the ray.exe manually and checking the code in Windows returns 1.
« Last Edit: March 27, 2013, 07:42:05 PM by pwortas »

jburk

  • Administrator
  • *****
  • Posts: 493
Re: Bug in mentalray type with mrstand3.10.1-adsk2013
« Reply #2 on: April 26, 2013, 05:11:46 PM »
Just saw this at another site, it seems to be limited to a single version of mentalray 3.10, namely 3.10.1.9

I can't remember the rpm name that installed this particular version, but if you get the latest mentalrayStandalone, it should install 3.10.1.11.

The customer also had an older rpm that installed 3.10.1.4, and both that and 3.10.1.11 exit with exit code 0 on success.