Hi,
Ted Unangst wrote on Fri, Mar 08, 2019 at 03:24:56PM -0500:
> Matthieu Herrb wrote:
>> I would prefer a diff that just add a &&!defined(__OpenBSD__) to the
>> condition before the definition of systemWcwidthOk(). This will cause
>> less risk of conflicts in future updates and clearly show the
Matthieu Herrb wrote:
> I would prefer a diff that just add a &&!defined(__OpenBSD__) to the
> condition before the definition of systemWcwidthOk(). This will cause
> less risk of conflicts in future updates and clearly show the
> intention.
If you prefer that, I would suggest the following to avo
On Fri, Mar 08, 2019 at 02:39:32PM +0100, Ingo Schwarze wrote:
> Hi,
>
> Lauri Tirkkonen wrote on Fri, Mar 08, 2019 at 12:43:11PM +0200:
>
> > I feel like xterm should just use the system wcwidth() to avoid these
> > mismatches, so rudimentary diff to do that below.
>
> Absolutely, i strongly ag
Hi,
Ted Unangst wrote on Fri, Mar 08, 2019 at 12:57:17PM -0500:
> Lauri Tirkkonen wrote:
>> in other words, xterm is (and was) using its own idea of how many
>> columns characters take up. The way this manifested itself to me was
>> that I received some email that contained emoji characters in th
Lauri Tirkkonen wrote:
> in other words, xterm is (and was) using its own idea of how many
> columns characters take up. The way this manifested itself to me was
> that I received some email that contained emoji characters in the
> subject, and they look fine in mutt when I use another terminal (st
Hi,
Lauri Tirkkonen wrote on Fri, Mar 08, 2019 at 12:43:11PM +0200:
> I feel like xterm should just use the system wcwidth() to avoid these
> mismatches, so rudimentary diff to do that below.
Absolutely, i strongly agree with that sentiment.
Having a local character width table in an applicatio
Hi,
it appears xterm tests the system's wcwidth() function on startup,
comparing the results between it and the wcwidth implementation it ships
itself (mk_wcwidth()), for each wchar_t value between 0 and 0x. This
is done in systemWcwidthOk(), called from decode_wcwidth().
I turned some TRACE(
