Increase output?

habere-et-dispertire · March 26, 2024, 9:57pm

Debugging grammars with Grammar::Tracer produces a fair amount of tree-like output which is truncated after 500 bytes.

If possible please increase this a tad as 18 lines of output can be insufficient, eg :

truncated-min

Merci !

See :

github.com

exercism/raku-test-runner/blob/main/bin/transform-results.raku#L63


      
              $subtest = @part[1].grep({.[0] eq 'comment'}).map({.[1]}).join;
          }
          
          when 'extra' {
              $output ~= @part[1];
          }
          
          when 'assert' {
              given %results<tests>[$i++] -> %test {
                  %test<name>   = @part[1]<name>;
                  %test<output> = $output.chars <= 500 ?? $output !! ($output.substr(0, 500) ~ '... Output was truncated. Please limit to 500 chars.') if $output;
                  if @part[1]<ok> {
                      %test<status> = 'pass';
                  }
                  else {
                      %test<message> ~= $subtest;
                      %test<status> = 'fail';
                  }
              }
              $output  = '';
              $subtest = '';

m-dango · March 27, 2024, 11:00am

This is a requirement in the test runner specification so unfortunately I cannot make this call myself:

habere-et-dispertire · March 27, 2024, 10:54pm

We can get ~50% saving by removing ANSI.

A Grammar::Tracer issue was raised a decade ago for an option to drop colour so we can’t do it that way round.

With a test regex of :

s:g/\x[1b]\[\d+[\;\d+]*<[mG]>//

It can render :

TOP
|  Codon
Phenylalanine
* MATCH "UUC"
|  * MATCH "UUC"
|  Codon
Phenylalanine
* MATCH "UUC"
|  * MATCH "UUC"
|  Codon
|  |  STOP
|  |  * MATCH "UAA"
|  * MATCH "UAA"
|  Codon
|  |  Tryptophan
|  |  * MATCH "UGG"
|  * MATCH "UGG"
|  Codon
|  * FAIL
* MATCH "UUCUUCUAAUGG"

What do you think @m-dango ?

m-dango · March 28, 2024, 1:06am

I’ve created a pull request which counts the characters without ANSI. If the output exceeds 500 characters, ANSI is removed and the output is truncated, else the color is left in.

github.com/exercism/raku-test-runner

Exclude ANSI colors from character count

exercism:main ← m-dango:colorstrip

opened 01:04AM - 28 Mar 24 UTC

m-dango

+8 -1

I've left the potential coloring of output present if the number of characters s…ans coloring falls below 500 (which could potentially end up being over 500, but I don't know if ANSI is considered part of the counting), else the color is stripped and output truncated. Created in response to https://forum.exercism.org/t/increase-output/10487