localedef during tests suddenly needs a lot of memory

Florian Weimer fweimer@redhat.com
Fri Aug 18 11:28:00 GMT 2017


On 08/18/2017 12:09 PM, Mike FABIAN wrote:
> Attached is muy patch to use ranges instead of single code points in the width
> data of charmaps/UTF-8 whereever possible.

Could you please verify that the charmaps/UTF-8 update does not change
the compiled locale data in any way?  If it does not, I think this part
should go in separately, to unbreak the test suite.

(I do not think we run the script automatically during build, right?)

Thanks,
Florian



More information about the Libc-locales mailing list