add support for f08 do concurrent (closes #403)

[rupertford]

Replaces PR #404 to provide a complete implementation where the matching to do concurrent is limited to f2008 code (f2003 will not match). Also adds appropriate f2008 tests, pulls out and extends the existing f2003 tests and updates the f2003 implementation.

To add in the matching to do concurrent I have changed the tuple that is returned by a match in f2003. This will affect any tools that walk the fparser2 tree structure. The first argument is now a string which specifies the type of match "WHILE" or "COUNTER". In the f2008 match "CONCURRENT" is also allowed. The second argument returns the classes resulting from matching and the third argument returns the optional delimiter. The reason for adding the first argument string is that it helps a tree traversal/transformation tool to determine what to expect in the second argument, as the classes returned are different for the different cases. The previous implementation returned classes in the first argument if matching a WHILE and None in the second and None in the first argument and classes in the second argument if matching a COUNTER, but this does not work well when extending to add CONCURRENT in F08.

[codecov[bot]]

Codecov Report

Patch coverage: 100.00% and project coverage change: +0.01 :tada:

Comparison is base (be09a23) 91.74% compared to head (cf09680) 91.75%.

Additional details and impacted files

```diff @@ Coverage Diff @@ ## master #408 +/- ## ========================================== + Coverage 91.74% 91.75% +0.01% ========================================== Files 37 37 Lines 13361 13384 +23 ========================================== + Hits 12258 12281 +23 Misses 1103 1103 ``` | [Impacted Files](https://app.codecov.io/gh/stfc/fparser/pull/408?src=pr&el=tree&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=stfc) | Coverage Δ | | |---|---|---| | [src/fparser/two/Fortran2003.py](https://app.codecov.io/gh/stfc/fparser/pull/408?src=pr&el=tree&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=stfc#diff-c3JjL2ZwYXJzZXIvdHdvL0ZvcnRyYW4yMDAzLnB5) | `94.40% <100.00%> (-0.01%)` | :arrow_down: | | [src/fparser/two/Fortran2008.py](https://app.codecov.io/gh/stfc/fparser/pull/408?src=pr&el=tree&utm_medium=referral&utm_source=github&utm_content=comment&utm_campaign=pr+comments&utm_term=stfc#diff-c3JjL2ZwYXJzZXIvdHdvL0ZvcnRyYW4yMDA4LnB5) | `99.49% <100.00%> (+0.03%)` | :arrow_up: |

:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Do you have feedback about the report comment? Let us know in this issue.

[rupertford]

I guess an alternative to the first return item having "COUNTER" which is just a made up name and is not part of the Fortran standard is to treat the first argument as a keyword and return None instead of "COUNTER". That would still allow tools that traverse the tree to distinguish. I'll leave it to the reviewer to see is they prefer that option (or some other of course).

[rupertford]

Ready for first review from @arporter or @sergisiso

[rupertford]

Ready for next review from @arporter

[arporter]

@rupertford was this the PR we needed to have a discussion about?

[rupertford]

After discussion with @arporter we have agreed to keep the original order of return values and to add additional return values to the end. This will avoid errors with any tools that use fparser. The best solution, which is to create new classes will be documentated in the docstring and in an issue but will not be implemented until we decide on a suitable time for non-backward compatible changes. A summary of the changes is below:

Backwards compatible
F2003: Loop_Cntl: sle, counter_expr, delim
F2008: Loop_Cntl: sle, counter_expr, delim, (conc_expr)
F2018: Loop_Cntl: sle, counter_expr, delim, (conc_expr, local_x)

F2003: While_Loop_Cntl: sle, delim
F2003: Counter_Loop_Cntl: var, lower, upper, [step], delim
F2008: Concurrent_Loop_Cntl: conc_expr, delim
F2018: Concurrent_Loop_Cntl: conc_expr, local_x, delim

[rupertford]

OK, ready for review again from @arporter