This is the mail archive of the
mailing list for the libstdc++ project.
Re: Patch: for libstdc++/2071 test case
- To: pedwards at disaster dot jaj dot com
- Subject: Re: Patch: for libstdc++/2071 test case
- From: Loren James Rittle <rittle at latour dot rsch dot comm dot mot dot com>
- Date: Wed, 6 Jun 2001 15:41:34 -0500 (CDT)
- CC: libstdc++ at gcc dot gnu dot org
- References: <200105100918.f4A9InP94151@latour.rsch.comm.mot.com> <200105161630.f4GGURY31700@latour.rsch.comm.mot.com> <20010605163320.A23471@disaster.jaj.com>
- Reply-to: rittle at labs dot mot dot com
Sorry, I have been away from e-mail for 24 hours.
> On Wed, May 16, 2001 at 11:30:27AM -0500, Loren James Rittle wrote:
>>> Here is a tentative patch for libstdc++/2071 which conforms to the
> I just made some notes in 2071 and marked it high-priority.
> With this patch, the bug still occurs for me on sparc-sun-solaris2.8 (and
> yes, I did add the #define to os_defines.h :-). I'm trying to figure
I can test solaris2.7 locally but I haven't built it in ages.
> out how to avoid the seeks altogether, but I'm not sure what's best to
> test against. Anything done inside needs to be fast...
Any approach that rids us of seek calls yet retained syncing with
stdio would be best. My patch was mainly to uncover the non-portable
use of seek and one possible work-around. I surely don't like having
to call ungetc...
> The RSI is acting up, so I'm going to give my wrist a break and ponder
> this for a while.
I can look into this after threading patches are submitted.