Re: [bug-gawk] Possible printf %c width multi-byte bug

[Nethox]

Aharon Robbins,  2013-07-02 21:30:
> Thanks for the report and the test files. With the patch below, it passes.
> I assume it covers the original report as well?

Yes, the test covers both bug reports. And also %s, which had no bug but
I think it was untested for multibyte width and precision.


>> I tried changing ENVIRON["LC_ALL"] from the script itself, but the
>> printf function was unaffected, so the script must be manually called by
>> prepending LC_ALL="C.UTF-8", or I guess with a new specific make target.
> 
> That is correct. Changing ENVIRON within the awk program does not affect
> the locale stuff; it's set at startup.

Ok, now I see the initialization at main.c; and the TODO entry talking
about setenv(), which could comprise setlocale() as well.


> Here is the new patch, relative to the gawk-4.1-stable branch in git.
> It includes a modified version of the first patch.
> 
> Once you approve, I'll push this.

It works great. All tests pass, including mbprintf4.awk with
LC_ALL="C.UTF-8". Moreover, the output of mbprintf4.awk with LC_ALL="C"
or --characters-as-bytes (-b) remains identical to the unpatched 4.1.0
one, which was already correct (taking into account that processing
multibyte input as bytes results in human-nonreadadable text).


> Thanks,
> 
> Arnold

Thank you for the programming and maintenance work.