X-Authentication-Warning: delorie.com: mail set sender to djgpp-workers-bounces using -f
From: <ams AT ludd DOT ltu DOT se>
Message-Id: <200501152147.j0FLlaRA012354@speedy.ludd.ltu.se>
Subject: Re: *time_r patch
To: DJGPP-WORKERS <djgpp-workers AT delorie DOT com>
Date: Sat, 15 Jan 2005 22:47:36 +0100 (CET)
X-Mailer: ELM [version 2.4ME+ PL78 (25)]
MIME-Version: 1.0
Content-Transfer-Encoding: 7bit
Content-Type: text/plain; charset=US-ASCII
X-ltu-MailScanner-Information: Please contact the ISP for more information
X-ltu-MailScanner: Found to be clean
X-MailScanner-From: ams AT ludd DOT ltu DOT se
Reply-To: djgpp-workers AT delorie DOT com

This message bounced around a little due to DJ's s p a m filters, so
it's been delayed a week.

According to Brian Inglis:
> On Fri, 07 Jan 2005 20:37:36 +0100 (CET), ams AT ludd DOT ltu DOT se wrote:
> 
> >According to Brian Inglis:
> >1. > Index: tests/libc/ansi/time/ctime.c
> >2. > ===================================================================
> >3. > RCS file: /cvs/djgpp/djgpp/tests/libc/ansi/time/ctime.c,v
> >4. > diff -u -p ctime.c
> >5. > --- /dev/null	2005-01-05 23:24:37.000000000 -0700
> >6. > +++ time.c	2005-01-05 23:24:16.000000000 -0700
> >7. > @@ -0,0 +1,49 @@
> >8. > +/* ctime.c - basic ISO and POSIX time function calls */
> >
> >Ohoy there! tests/libc/ansi/time/ctime.c (line 1), ctime.c (line 4 and
> >8) but time.c (line 6)! Note the missing "c".
> >
> >Line 4 hasn't the full name either, which mean (c)time.c won't be
> >created in the proper place. At least it wasn't with my version of
> >patch. (That could be my old patch or bad parameter from me.)
> >
> >Please make a little better patches.
> >
> >While I certainly do appreciate the work you do, you should try to
> >make it easy for those who are trying to apply the patches.
> 
> Unfortunately, cvs diff does not seem to work with the -N option
> when it is a new file with nothing to compare against,

No, it doesn't unless you have cvs add-ed that file (which I use
rather exensivly). But unless you have write access, I don't think you
can cvs add the file.

> so I tried to fake the cvs/rcs header based on a manual diff
> generated before I renamed the file.

That is ok. But please, please, please tell us when there's something
out-of-the-ordinary going on, which will help with any
surprises. (Cf. my "(pasted)" commments for some of the patches of
mine.) What we want is minimal surprise for any users of the submitted
work.

(You get extra confusion points for managing to rename ctime.c to
time.c.)

Next time, just tell us that you had to special treat the file XYZ
because it doesn't exist, or _some_thing (perhaps saying new file
dir/file which contains
---- file start
...
---- file end
. I'm not sure what is the most easiest to work with format.)

Or somebody should tell me the option to patch that makes this work
fine.

An extreme and cool variant would be to let me (or somebody with write
access) cvs add the file, whereafter you should be able to cvs
diff(1). However that would mean a relatively quick interchange (I
want this file -> Ok, added -> Here's diff.), which, this time, in
retrospect probably would have worked given our response times. But
suppose you don't get a reply that the file is ready for diff-ing in
a week or month...?

> Please let me know if the following bare diff works
> with your patch on your system,
> then I know I can just omit the extra cvs/rcs junk.

I won't unless I have a lot of time on my hands (2), because I've
already debugged the problem (strange patch) and corrected it (moved
that time.c to <whatever>/ctime.c


Big thumbs up for the work!


Right,

						MartinS

(1) I don't know if this actually works.

(2) Which is very rare for me (and that probably goes for
everyone). More like unheard of lately, actually...