On the run up to ER10 and O2, I am regularly reviewing code which is locally modifed and not committed to the SVN repository. I have cleared out most of the outstanding mods for: front end source code, filter module files and guardian code. Items which are scheduled for install tomorrow have been kept locally modified for now.
I noticed that not all matlab mdl source files have been generated with matlab version 2012b (v 8.0). Here are the ones which are different:
2010a /opt/rtcds/userapps/release/isi/common/models/Offset_DOF.mdl
2015b /opt/rtcds/userapps/release/isc/h1/models/h1oaf.mdl
2015b /opt/rtcds/userapps/release/isi/h1/models/h1isietmx.mdl
2015b /opt/rtcds/userapps/release/isi/h1/models/h1isietmy.mdl
The version can be found by running my script which_matlab_version.py
I did a quick test to see if matlab 2015b produces the same binary code as 2012b. I used h1pemmx as the test model. On my ubuntu14 workstation running matlab2015b, I opened h1pemmx.mdl and then saved it making no file changes. The 2015b file was compiled on h1build (but not installed). The h1pemmx.ko files in the target area (2012b) and the build area (2015b) were compared. They have the same size but different checksums. Converting binary to strings using the string command shows the only differences are the build datetime which is encoded into the binary.
This is not an exhausive test, h1pemmx is a very simple model. Users are reporting much less graphical issues running matlab 2015b on iMacs, perhaps we should consider upgrading?