Sender:	rich AT phekda DOT freeserve DOT co DOT uk
Message-ID:	<3D01DCB7.BD837FBA@phekda.freeserve.co.uk>
Date:	Sat, 08 Jun 2002 11:30:15 +0100
From:	Richard Dawe <rich AT phekda DOT freeserve DOT co DOT uk>
X-Mailer:	Mozilla 4.77 [en] (X11; U; Linux 2.2.19 i586)
X-Accept-Language:	de,fr
MIME-Version:	1.0
To:	djgpp-workers AT delorie DOT com
Subject:	Re: GCC 3.1 & CVS LIBC Uncommited patches
References:	<003001c20ebd$d2149fd0$0102a8c0 AT acceleron>
Reply-To:	djgpp-workers AT delorie DOT com

Hello.

Andrew Cottrell wrote:
[snip]
> There have been a number of patches circulating over the past two week
> mainly for variable argument parameter passing issues found when building
> the CVS LIBC code with GCC 3.1. I would like to get back to being able to
> build the LIBC code from the CVS source instead of what I currently have.
> 
> I just fetched the latest LIBC code a few hours ago and found that the
> marjority of the patches have not been comiited. Is there any reason for not
> comitting the majority of the patches, except the possibly the
> src/libc/compat/ioctl/ioctl.c and src/libc/posix/fcntl/fcntl.c ?

I've just committed some of the patches. I think we decided a while ago that
there should be a gap of a week between a patch being approved and the patch
being committed.

> The following is my list of C & H files that are diffrenet between the CVS
> LIBC code and my old SRC directory which I sucessfully built and then used
> to re-build majority of the other packages I normally build. I have yet to
> figure out an issue when building G77.
> 
>       File Src
>       fprintf.c libc\ansi\stdio
>       printf.c libc\ansi\stdio
>       remove.c libc\ansi\stdio
>       sprintf.c libc\ansi\stdio
>       ioctl.c libc\compat\ioctl
>       _irdlink.c libc\compat\unistd
>       fchown.c libc\compat\unistd
>       llseek.c libc\compat\unistd
>       symlink.c libc\compat\unistd
>       _close.c libc\dos\io
>       _creat.c libc\dos\io
>       _creat_n.c libc\dos\io
>       _open.c libc\dos\io
>       _read.c libc\dos\io
>       _write.c libc\dos\io
>       fstat.c libc\posix\sys\stat
>       lstat.c libc\posix\sys\stat
>       confstr.c libc\posix\unistd
>       lseek.c libc\posix\unistd
>       doscan.c libc\ansi\stdio
>       select.c libc\compat\time
>       conio.c libc\pc_hw\co80
>       link.c libc\posix\unistd
>       write.c libc\posix\unistd
>       emu387.cc libemu\src
>       screen.c debug\fsdb
> 
> There were also some patches that were to do with some of the makefiles.

Here's what I've just committed:

* fixes for libemu/src;
* fixes for debug/fsdb;
* fixes for ansi/stdio/*printf.c;
* disable warnings when compiling fcntl.c & ioctl.c.

That leaves patches in the following category:

* _doscan*() fixes;
* FSEXT handling;
* confstr.c fix.

_doscan*() needs more testing. The FSEXT patch needs reworking, to use a
library-private header for the common functions, as suggested by Eli. I need
to rework the confstr.c patch, having realised why we use snprintf().

I'm on vacation for the next week or so, so hopefully I can test all this
stuff out sometime in the middle of relaxing and having
non-computer-programming-related fun. 8)

Bye, Rich =]

-- 
Richard Dawe [ http://www.phekda.freeserve.co.uk/richdawe/ ]

- Raw text -