(Closes #426) Add convert argument to open().

[arporter]

No description provided.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Comparison is base (0410062) 91.99% compared to head (ce2a081) 91.96%.

❗ Current head ce2a081 differs from pull request most recent head 53f32fa. Consider uploading reports for the commit 53f32fa to get more accurate results

Additional details and impacted files

@@            Coverage Diff             @@
##           master     #435      +/-   ##
==========================================
- Coverage   91.99%   91.96%   -0.04%     
==========================================
  Files          85       85              
  Lines       13681    13641      -40     
==========================================
- Hits        12586    12545      -41     
- Misses       1095     1096       +1     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[arporter]

In the interests of speed (of implementation) I haven't added full testing of the OPEN statement here.
Ready for review from either @sergisiso or @rupertford.

[rupertford]

Some additional tests, refactoring and making the extension optional are required - please see inline comments.

[arporter]

Ready for another look now. Ran across #432 while trying to check the documentation. Happily that's a separate issue though :-)

[arporter]

Taking this back as I've discovered it doesn't work if one requests a F2008 parser instead of F2003!

[arporter]

This has got a bit bigger as, on looking at the Fortran2008 implementation of Open (which I missed first time around), I realised that there was a lot of code duplication that could be factored out. I also slightly refactored the way that we handle the list of supported extensions (we now have a method that returns the list rather than just a module-scope variable containing a list). This makes testing a bit cleaner as monkeypatch can be used more easily.

Ready for another look now.

[sergisiso]

@arporter Again, I gave a quick read to this PR since this was already reviewed once to check that I agree with the general solution and that the previously raised comments from @rupertford are now all resolved. This is ready to merge.