QA

Observation¶

Context: https://suse.slack.com/archives/C02CANHLANP/p1727447152612169
Latest suggestion by mdoucha

Could you check the database data for any weird unicode bytes in Wei Gao's comments? there might be some weird whitespace throwing the bot off

From tinita:

openqa=> select job_id, t_created, text from comments where job_id = 15491984;
  job_id  |      t_created      |                        text                         
----------+---------------------+-----------------------------------------------------
 15491984 | 2024-09-23 09:13:28 | \r                                                 +
          |                     | \r                                                 +
          |                     | @review:acceptable_for:incident_35521:bsc#1227151\r+
          |                     | 
 15491984 | 2024-09-27 14:19:09 | @review:acceptable_for:incident_35521:bsc#1227151
(2 rows)

So I guess carriage returns are a problem.
We should probably filter them already when saving comments

Acceptance criteria¶

• AC1: qem-bot is more tolerant when parsing review comments, e.g. not be caught up by additional \r characters or other non-printable characters
• AC2: Any text before the @review is ignored
• AC3: Trailing characters that should not be part of the reason are not considered part it.

Suggestions¶

• Consider remove non-printable characters from comments when saving: e.g. $var =~ s/.*[^[:print:]]+//;