499 avoid exponential loop handling clean

[hiker]

Fixes the exponential behaviour with number of non-blocked loops in one Fortran unit. This is done by detecting when searching for a blocked loop in e.g.:

      do 10 i=1, 10
 10     a(i) = i

That the statement with label 10 indicates that there is no matching 10 enddo (or continue) statement coming, and therefore abooting earlier.

It cuts down parse time for a socrates file that was not parsed in 40 minutes(!!) down to 7 seconds.

Problem: I don't know how to test this patch. During development, the existing tests have flagged some issues I had, so the code is definitely used, and would trigger a failed test if it is really broken. But I don't really know how to test if it triggers as expected (since the function itself would return None without this patch as well ... just later)

Suggestions welcome

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 92.21%. Comparing base (a197a85) to head (869c01c).

Additional details and impacted files

@@           Coverage Diff           @@
##           master     #501   +/-   ##
=======================================
  Coverage   92.21%   92.21%           
=======================================
  Files          87       87           
  Lines       13832    13843   +11     
=======================================
+ Hits        12755    12766   +11     
  Misses       1077     1077           

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

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[hiker]

While this indeed fixes the issue, it is only fixed for the 2003 standard. Fab uses (by default) F2008, and F2008 still runs forever 😢

[hiker]

Ah, fixing F2008 was actually trivial :)

Ready for a first review.

[hiker]

I actually got an idea for testing: adding an invalid/not-parsable Fortran statement after a loop - if a do loop is parsed without error, it means that the error wasn't reached, and therefore it stopped early. Marking this as 'in progress' again :)

[hiker]

An incorrect statement didn't really work, but I could just check the error message when blocking doesn't parse. If it points at the labeled statement that finished a non-blocked loop, the shortcut works. If it points at the end of the file, it doesn't - hooray ;)