This is the mail archive of the binutils@sources.redhat.com mailing list for the binutils project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: PATCH: Export LC_ALL=C (Re: Sorting in nm and testsuite)


On Thu, Jun 27, 2002 at 01:44:49PM -0700, Geoff Keating wrote:
> > From: Andreas Schwab <schwab@suse.de>
> > Date: Thu, 27 Jun 2002 10:11:46 +0200
> 
> > Geoff Keating <geoffk@geoffk.org> writes:
> > 
> > |> [I note, BTW, that whoever made the original change that caused the
> > |> problem seems to have forgotten to add a testcase for it; the obvious
> > |> testcase is to set LC_ALL=en_US and check that the sort order changes.]
> > 
> > You can't really test that portably, since locales are implementation
> > defined.  Even locale names differ among platforms.
> 
> How about doing it only for glibc-based OSs, then?  In practise, if
> you can write a test that works under linux, it will be run often
> enough that any regressions will be found.

You just need to explicitly set LC_ALL in your testcase. You can't
assume anything what LC_ALL is.


H.J.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]