When similar warnings are grouped they shouldn't be truncated in the output

[ardrabczyk]

For example:

$ sarif emacs results.sarif
Writing results for results.sarif to results.txt
$ cat results.txt
-*- compilation -*-

Sarif Summary: CodeQL
Document generated on: 2025-03-07 16:03:25.768691
Total number of distinct issues of all severities (error, warning, note): 2



Severity : error [1]
dir/abc.c:506: cpp/wrong-type-format-argument This format specifier for type ' ...
dir/def.c:107: cpp/wrong-type-format-argument This format specifier for type ' ...
dir/123.c:160: cpp/wrong-type-format-argument This format specifier for type ' ...
dir/124.c:160: cpp/wrong-type-format-argument This format specifier for type ' ...



Severity : warning [1]
dir/www.c:132: cpp/comparison-with-wider-type Comparison between [n](1) of type uint8_t and [size](2) of wider type uint32_t.



Severity : note [0]

Notice the This format specifier for type ' ... lines. As a user I'd prefer to see a full line here such as This format specifier for type 'int' does not match the argument type 'unsigned long'., current output isn't very useful.

[johnsaigle]

Agreed, I came here to post this. Maybe a column length parameter passed to the summary, html, ..., modes would be a good way to approach this.