Re: toolchain/38804

[christos%zoulas.com@localhost (Christos Zoulas)]

On Jun 25,  7:33pm, jmc%netbsd.org@localhost (James Chacon) wrote:
-- Subject: Re: toolchain/38804

| 
| On Jun 25, 2008, at 6:30 PM, Valeriy E. Ushakov wrote:
| 
| > The following reply was made to PR toolchain/38804; it has been  
| > noted by GNATS.
| >
| > From: "Valeriy E. Ushakov" <uwe%netbsd.org@localhost>
| > To: gnats-bugs%NetBSD.org@localhost
| > Cc:
| > Subject: Re: toolchain/38804
| > Date: Thu, 26 Jun 2008 05:25:26 +0400
| >
| > On Thu, Jun 26, 2008 at 05:07:49 +0400, Valeriy E. Ushakov wrote:
| >
| >> The problem is that, as -fdump-class-hierarchy tells me,
| >> std::basic_filebuf<char, std::char_traits<char> > has shrunk from
| >> 500 to 492 bytes, so all <fstream> stuff gets broken.
| >
| > And that's b/c it has a __c_lock _M_lock; and in current
| > pthread_mutex_t has different size... *sigh*
| >
| > Oh, and libstdc++ doesn't list libpthread as NEEDED.
| >
| > -uwe
| >
| 
| Even if it did this isn't a libstdc++ issue then. Ugg...
| 
| So, when the replacement pthread library went in I'm assuming  
| everything that depended on the old one was simply bumped in version?
| 
| Here it still sounds like that's the solution then.

Yes, the solution is to bump libstdc++ *and* list -lpthread as a dependency
so that this does not happen again.

christos