This is the mail archive of the
mailing list for the GCC project.
Re: [v3] patch for Missing explicit instantiations in libstdc++
- From: Loren James Rittle <rittle at latour dot rsch dot comm dot mot dot com>
- To: gcc-patches at gcc dot gnu dot org
- Cc: dave at hiauly1 dot hia dot nrc dot ca
- Date: Fri, 7 Dec 2001 15:42:55 -0600 (CST)
- Subject: Re: [v3] patch for Missing explicit instantiations in libstdc++
- Organization: Networks and Infrastructure Lab (IL02/2240), Motorola Labs
- References: <Pine.SOL.3.91.1011204114830.8366Bfirstname.lastname@example.org>
> /usr/ccs/bin/ld: Unsatisfied symbols:
> std::_Swap_lock_struct<0>::_S_swap_lock (data)
> I don't have much time to look at this in the next few days but possibly
> the test should check whether __GTHREAD_MUTEX_INIT is defined. This build
> was done with `Thread model: single'.
Um, I don't expect anyone to remember any particular e-mail, but this
failure on your platform was mentioned when I checked in ext/rope.cc:
On Nov 2., I wrote to the libstdc++ and gcc-patches lists (the message
was directed at you since you reported breakage on your platform but
it is unclear that I CC'd or BCC'd you):
> FYI, ext/rope.cc will not pass on your platform until this is fixed
> properly so there is an outstanding record of the situation.
We do not want to disable the test case or force it to pass on ports
that are missing functionality since that is the record that some work
is still required to bring your platform up to other ports. It could
be XFAIL'd on your port but do nothing to mask the failure please.
Yet another reason not to mask the problem: No port that reports
`Thread model: single' should fail that test case since:
#define __GTHREAD_MUTEX_INIT 0
is in gthr-single.h and that file should be used by all ports that report
`Thread model: single'.