This is the mail archive of the gcc@gcc.gnu.org mailing list for the GCC 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: libstdc++ tests assumes all locale data are available


On Mon, Dec 02, 2002 at 03:36:30PM -0500, Phil Edwards wrote:
> On Mon, Dec 02, 2002 at 12:27:33PM -0800, H. J. Lu wrote:
> > > the libstdc++ testsuite will not pass the named locale tests. If this 
> > > isn't an issue, don't worry about it. If named locales are needed, the 
> > > underlying locale information must be installed. Note that rebuilding 
> > > libstdc++ after the "C" locales are installed is not necessary.
> > 
> > Those tests are meaningless.
> 
> Beg pardon?
> 
> If you're using locales, then you'll have locales installed.  And we
> test them.  When the tests break, we know something is wrong.
> 

You can't tell the cause of the failure. Since I am ignoring all
libstdc++ test failures, I may not notice there is a regression. Also
even if a locale, say de_DE, is added later, I still don't know if
de_DE works or not. 

As I said, if a test can't tell what causes the failure, it is next to
useless. Another way to fix it is to check available locales during
configure time. You may use "locale -a" for that.

H.J.


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