Fix #17 - Error suppression operator is not fully honoured

[kAlvaro]

Warnings suppressed with @ were still altering the test status.

This is still not a full solution, because status is assigned separately for each error severity and I've only changed warnings. I wanted to get some feedback before spending more time on this.

[Bilge]

All changes look good to me 👍🏻

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 100.00%. Comparing base (e71b10a) to head (f44ae07).
Report is 1 commits behind head on master.

Additional details and impacted files

@@             Coverage Diff             @@
##              master       #18   +/-   ##
===========================================
  Coverage     100.00%   100.00%           
- Complexity        61        64    +3     
===========================================
  Files              6         6           
  Lines            143       146    +3     
===========================================
+ Hits             143       146    +3     

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

[kAlvaro]

I've added the same fix for notices and deprecations. I don't think you can use @ for any other severity types (I tried triggering E_USER_ERROR and it wasn't being caught).

While working on this, I found another issue, but it's a different thing and I don't know how it's meant to work so I'll probably open an issue to discuss it.

[Bilge]

E_USER_ERROR is a separate event. It should be possible to handle it in Printer::trace, although I haven't tried. In any case, I think we can add support for USER level events later.

[kAlvaro]

E_USER_ERROR is a separate event. It should be possible to handle it in Printer::trace, although I haven't tried. In any case, I think we can add support for USER level events later.

Honestly, I only tried that because I couldn't think of a way to trigger a regular error on purpose.

[Bilge]

E_USER_ERROR is a separate event. It should be possible to handle it in Printer::trace, although I haven't tried. In any case, I think we can add support for USER level events later.

Honestly, I only tried that because I couldn't think of a way to trigger a regular error on purpose.

Easiest way is to search php-src for E_ERROR and pick one that's easy to reproduce.

[Bilge]

Thanks for this @kAlvaro!