fix SEGFAULT in checking prototype arguments

[lefessan]

No description provided.

[codecov-commenter]

⚠️ Please install the to ensure uploads and comments are reliably processed by Codecov.

Codecov Report

Attention: Patch coverage is 80.00000% with 1 line in your changes missing coverage. Please review.

Project coverage is 65.85%. Comparing base (a2a51fd) to head (9f7df04).
Report is 99 commits behind head on gcos4gnucobol-3.x.

Files with missing lines	Patch %	Lines
cobc/parser.y	80.00%	0 Missing and 1 partial ⚠️

❗ Your organization needs to install the Codecov GitHub app to enable full functionality.

Additional details and impacted files

@@                  Coverage Diff                  @@
##           gcos4gnucobol-3.x     #133      +/-   ##
=====================================================
- Coverage              65.86%   65.85%   -0.01%     
=====================================================
  Files                     32       32              
  Lines                  59481    59483       +2     
  Branches               15708    15709       +1     
=====================================================
- Hits                   39177    39175       -2     
- Misses                 14283    14290       +7     
+ Partials                6021     6018       -3     

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

[GitMensch]

thanks for the test and the fix, only a minor change to the testsuite, an added ChangeLog entry and that's good to go

[GitMensch]

I see, we expect WORD and have a LITERAL.

For level 78 (likely also for level 01 constants) we have several places that result in a strange error message (and like in that place a bit misleading position of the error message). A way to solve that would be passing an extra token type, but then we'd have to add something like the following - and use it in nearly all places where we have LITERAL and some, like that, where we currently use WORD.

literal: LITERAL { $$ = $1; } | CONSTANT { $$ = CB_TREE (CB_CONST_USER ($1)->literal;) }

That's definitely too much for this PR.
As the parser error leads to a "totally different" code path: please leave lvl78 in the test, but commented out with a note.

[lefessan]

@engboris Maybe that could be a first job for you ? If I understand well, the idea is to remove 78-level constants from the lexer, and replace all the places where they can replace a LITERAL by a rule that also accepts a WORD and checks if it is a 78-level ?

[GitMensch]

@lefessan that would be quite a big change and @engboris already took one of the other issues (towards stateless runtime), which is good to be finally handled.

[lefessan]

I commented out the level 78 from the prototype call.

[GitMensch]

Good as is now. For the additional part of the 78 handling - please create an issue in the wish-list tracker with as much technical information as reasonable for someon to work on this later.

[GitMensch]

@lefessan @ddeclerck Who does the commit to SVN now?

[ddeclerck]

Fabrice might be a bit busy at the moment, I'll take care of it (tomorrow I guess).

[ddeclerck]

Merged in SVN at revision 5227.
(@GitMensch Can you reattribute it to Fabrice ?)

[GitMensch]

metadata updated, no bug issue to be closed (nut that's quite an exotic point to reach, and described enough in the ChangeLog - not necessary to create a bug report)

On to the next PRs :-)