Re: [PATCH 1/3 v2] Limit width of formatted text by characters rather than bytes

On Sun 2017-09-17 21:31:08 +0200, Jeremie Courreges-Anglas wrote:
> Actually the behavior was rather incorrect.  After my patch,
> concat_width was copying up to 'width' bytes, not up to 'width' UTF-8
> characters.  This gotcha was caught by your test case.  I find the
> latter behavior more useful.
> 
> Note that the function only cares about how many characters we concat,
> not about their actual width on screen.  I believe this behavior is
> reasonable, though.

Sounds good to me!

> > Anyway, I think it is a big improvement as is.
> >
> > It also still might be good to have a few unit tests. Would you be okay
> > with tests in the form of my third patch?
> 
> Yep.  I have applied your 3/3 patch, which introduces a testcase, and
> made further tweaks.  Namely:
> - rename concat_width to something more appropriate
> - move it so sbuf.c
> - shuffled some code to make it easier to link against it
> - move the testcase to its own .c file
> - simplify src/Makefile.am accordingly
> 
> Feedback welcome.

It looks great. Thank you for getting that in!