X-Authentication-Warning: ieva01.lanet.lv: pavenis owned process doing -bs
Date: Fri, 9 Jul 1999 20:21:27 +0300 (WET)
From: Andris Pavenis <pavenis AT lanet DOT lv>
To: Michel de Ruiter <mdruiter AT cs DOT vu DOT nl>
cc: djgpp-workers AT delorie DOT com
Subject: Re: DJGPP cvs and gcc-2.95
In-Reply-To: <m112b44-0000BJC@sloep108.cs.vu.nl>
Message-ID: <Pine.A41.4.05.9907092015150.95434-100000@ieva01.lanet.lv>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Reply-To: djgpp-workers AT delorie DOT com
X-Mailing-List: djgpp-workers AT delorie DOT com
X-Unsubscribes-To: listserv AT delorie DOT com
Precedence: bulk


I'll try that when I'll return home tomorrow (or in next week)
At least I have worked with egcs-19990623 and later with egcs-19990629
much and didn't found evident problems (with my own applications in C++)
However I haven't tried to build CVS version of DJGPP in latest 2 weeks.

I don't think gcc should crash on such not very complicated sources.
Are You sure that was not due to Wincrash95 DPMI server bugs (descriptor
leaks)? 

Andris

On Fri, 9 Jul 1999, Michel de Ruiter wrote:

> Dear workers,
> 
> I have tried to compile the current DJGPP cvs libsource, with gcc-2.95
> (as compiled by Andris IIRC).
> 
> Trying that, gcc crashed twice, on unlzw.c and redir.c. I am currently
> investigating the reason, which is difficult as the crashes only
> happen *some* of the times, not always. It may be my computer, as I
> have new SIMMs and an other motherboard.
> 
> Also, I have some questions about the current libsource:
> 
> - There are 8 filenames ending in .orig in manifest/djlsr202.mft. I
>   realize the *203.mft have not been created yet, so the *.orig will
>   not be in them. True?
> - I remember having discussed a patch (proposed by me :-) to
>   src/libc/dos/io/setmode.c, so setting a tty to O_BINARY would at
>   least give an interruptible situation. Is the situation fixed, in
>   some other way, or do we want to leave the situation as is (at least
>   for 2.03)?
> - src/libc/dos/io/setmode.txh discusses to some length the possibility
>   to read the console in binary mode. Is there any situation where
>   this could be useful? And, as the input is line buffered,
>   __djgpp_set_ctrl_c will not really help to generate signals, will
>   it? Or do I still miss something, about console in binary mode?
> 
> During compiling, gcc issued the following warnings. I have not had
> the time to "correct" them (yet). Someone else might?
> 
> dbgcom.c:25: warning: `id' defined but not used
> fullscr.c: In function `redraw':
> fullscr.c:1618: warning: long double format, double arg (arg 3)
> stubify.c: In function `coff2exe':
> stubify.c:50: warning: `used_temp' might be used uninitialized in this function
> scheck.c: In function `scheck':
> scheck.c:56: warning: cast discards qualifiers from pointer target type
> 
> Groente, Michel.
>