(Closes #457) Fix bug with backslash in strings

[arporter]

Weirdly, fparser was interpreting backslash "\" as an escape character. This is not a standard Fortran feature. i.e. if you want to escape a quotation mark in a Fortran string you simply repeat it: "''" means "'".

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 92.15%. Comparing base (370884f) to head (a7f029e).

Additional details and impacted files

@@            Coverage Diff             @@
##           master     #462      +/-   ##
==========================================
- Coverage   92.15%   92.15%   -0.01%     
==========================================
  Files          86       86              
  Lines       13734    13733       -1     
==========================================
- Hits        12656    12655       -1     
  Misses       1078     1078              

☔ 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.

[arporter]

The PSyclone test suite is happy with this branch of fparser so I declare it ready for review. It is a small but quite fundamental change in behaviour that affects both fparser1 and fparser2. However, since interpreting a \ as an escape character is non-standard, I think this is OK.

[hiker]

Thanks a lot for this. The loop looks a lot simpler that what it was before, that's great. I do have the feeling that there is a bug though, please check the comments (and if I am wrong, it needs some tests :) ).

[arporter]

Somehow I've accidentally started a review while attempting to reply to your comments. Anyhow, I've given up on AI and gone back to NI to write a new implementation. I think it's easier to understand now.

[arporter]

OK, ready for another look now @hiker.

[hiker]

Ready to be merged, thanks!

[hiker]

Dang, I need to take that "ready to be merged" back. One line was not covered by your new tests, and that line is actually wrong:
Test (continued line if no closing quote character is found). This adds two tests, but I believe they are actually the same (but the second tests again the handling of 'the other' quote character :) )

fparser/src/fparser/common/tests$ git diff ./test_splitline.py
diff --git a/src/fparser/common/tests/test_splitline.py b/src/fparser/common/tests/test_splitline.py
index cc155c5..68c2612 100644
--- a/src/fparser/common/tests/test_splitline.py
+++ b/src/fparser/common/tests/test_splitline.py
@@ -261,6 +261,8 @@ def test_splitquote(input_line, expected_parts, expected_unterm):
     "input_line, expected_parts, expected_unterm, stopchar, lower",
     [
         ("this is STILL a quote'", ["this is STILL a quote'"], None, "'", True),
+        ("this is STILL a quote", ["this is STILL a quote"], "'", "'", True),
+        ("this is STILL a quote\"", ["this is STILL a quote\""], "'", "'", True),
         ("'' STILL a quote'", ["'' STILL a quote'"], None, "'", True),
         ("'' STILL a', Quote", ["'' STILL a'", ", quote"], None, "'", True),
         ("'' STILL a', Quote", ["'' STILL a'", ", Quote"], None, "'", False),

I believe this is the fix:

fparser/src/fparser$ git diff common/splitline.py
diff --git a/src/fparser/common/splitline.py b/src/fparser/common/splitline.py
index f5667fa..bdcf55b 100644
--- a/src/fparser/common/splitline.py
+++ b/src/fparser/common/splitline.py
@@ -324,7 +324,7 @@ def splitquote(
             pos = end + 1
         else:
             # Didn't find a closing quotation char.
-            return String(line), stopchar
+            return [String(line)], stopchar
 
     while pos < n:
         start = _next_quote(line, start=pos)

[arporter]

Thanks for finding that bug. I didn't understand your suggested change to the test so have made my own. Hope it's in-line with what you had in mind. Weirdly, I couldn't find any report saying that line wasn't covered - how did you find it?

Ready for another look now.