buffer zero alloc

[mcakircali]

fixes zero size allocation issue such when size=0. namely, new char[size]
in such state, buffer.data() != nullptr is not expected but buffer.data() == nullptr

[codecov-commenter]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 63.79%. Comparing base (3598aa9) to head (723f9d2).
Report is 24 commits behind head on develop.

Additional details and impacted files

@@             Coverage Diff             @@
##           develop     #146      +/-   ##
===========================================
+ Coverage    63.72%   63.79%   +0.06%     
===========================================
  Files         1065     1065              
  Lines        55141    55358     +217     
  Branches      4084     4084              
===========================================
+ Hits         35139    35314     +175     
- Misses       20002    20044      +42     

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

[wdeconinck]

Please could you provide a description to the issue and what this is solving?

[mcakircali]

Please could you provide a description to the issue and what this is solving?

done

[wdeconinck]

This looks like a good and sensible change to me.

[Ozaq]

Looking at Buffer this could be simplified by using std::vector<>char as backing container.

This would remove some code such as allocate/deallocate

Further It would be good to add documentation to the public interface since you already touched this class. Construction from std::string and copy from std::string deserve a short hint in the api doc imo as they treat std::string as a null terminated c-string. i.e. the resulting buffer length is string.size() + 1.

[FussyDuck]

All committers have signed the CLA.

[Ozaq]

pushed doc after talking to @mcakircali

Edit: Also the Buffer dtor should be virtual as this class is subclassed and has protected members. I think making this class final would be better but that chan get the API