SConstruct : Race condition when building Cortex (with many threads or in debug mode)

[boberfly]

Generally describe what this PR will do, and why it is needed.

• Attempted fix for SConstruct : Race condition when building Cortex in debug mode (linking IECorePython when it isn't ready yet) #1308
• Based on this old one but a bit simpler: Build : Add more SCons 'Depends' onto the python modules #1185

Related Issues

• SConstruct : Race condition when building Cortex in debug mode (linking IECorePython when it isn't ready yet) #1308

Dependencies

• NA

Breaking Changes

• NA

Checklist

• I have read the contribution guidelines.
• I have updated the documentation, if applicable.
• I have tested my change(s) in the test suite, and added new test cases where necessary.
• My code follows the Cortex project's prevailing coding style and conventions.

[johnhaddon]

As I described in the previous PR, I don't think this is the right approach - we don't want to make the build depend on the install. We should be able to build, test and only then install. Does my description of the problem in #1185 (comment) apply in your case Alex? I think the solution is to ensure that the -L./lib for the build dir comes before the -L for the dependencies dir...

[boberfly]

I'm not too familiar with SConstruct so I don't really know here how to solve that one. I tried putting in any prepends I found with an additional "./lib" to ensure it is first but it didn't seem to work for me...

[johnhaddon]

When it didn't work for you, was that because ./lib wasn't coming first on the command line (i.e. the SConstruct still had problems) or because ./lib was coming first but it didn't stop the linker erroring?

[boberfly]

By the looks of it the BUILD_DIR is coming in first before the ./lib for those ones having issues.

[johnhaddon]

OK, then I think that's the issue we need to fix.

[johnhaddon]

I'm going to close this. As discussed in the comments, I don't think this is the right approach, and the urgency has also now been reduced since we added a workaround in GafferHQ/dependencies, where the problem crops up most.