[svn.haxx.se] · SVN Dev · SVN Users · SVN Org · TSVN Dev · TSVN Users · Subclipse Dev · Subclipse Users · this month's index

Re: svn:keywords ... a plea

From: marc gonzalez-carnicer <carnicer.lists_at_gmail.com>
Date: Tue, 25 Mar 2008 11:20:00 +0100


i don't quite understand the point of your email. you seem
to just make an informative comment to share. i am replying
because i've found it funny to find somebody else in this list
who has also used SCCS, which is what i used before i found
svn. and also, i'd like to point out that i find keywords in the
sense that you use them pretty useless (my opinion :)

2008/3/24, Graeme Vetterlein <graeme_at_vetterlein.com>:
> OK, I've finally worked up enough enthusiasm to
> put this down in text. (and register for mailing-list etc ...)
> In case I get the mood wrong, this is a plea, not a rant. ... honest.
> I guess I've been using Source code control systems on Unix type
> platforms for 20-25 years. I've used SCCS, RCS, CVS, PCVS, Clearcase
> and most recently SVN. I've used SVN at work for a couple of years and
> I'm finally moving my home systems over to SVN.
> One thing I've always impressed on people developing software is
> 1: Make sure you have them
> 2: Make sure people can identify them
> For part2, I'm usually meaning things like -v (--version) and ident(1)
> and what(1) strings. All these rely on version keywords. Simple:
> SCCS: static const char sccid[] = "%Z%%M% %I% %E%"

hehe :) this line triggered many memories. btw, i think some
SCCS / teamware implementations had an error. whenever your
code contained some date string, like "2001/01/01", that could be
mistook by the sccsid string, that date was replaced by something
like "%Z...", and that caused the code not to work. it was pretty tough
to find out what happened because when the file were checked out,
the file was correct, but when checked in, it was not : so we finally decided
not to use sccsid.

but back to my point. i really don't understand why you need to
set these date and author properties / keywords : it is *all* in the

i remember that we had the policy to set sccsid properly in all
our files in case a bug was found in the code, it was possible to know
which version of each file the executables had been produced from, with
the "strings <executable> | grep sccsid" command, and assuming
that the deployed executable had not been properly "freezepointed" (in
the SCCS sense).

the point is that with SVN, as well as with other *modern* VCSs, there
is no need to track date, version and author of each file because
every commit is a "freezepoint", and the yielded revision number is
all you need, and the date, author, etc info is on the log.

so, i am not gonna get into the details on how to set autoprops or
other things, just that i want to suggest that perhaps there is no
need to use them. but it that has been implemented i may be
missing something. my apologies if i misunderstood the point of
your email.

To unsubscribe, e-mail: users-unsubscribe_at_subversion.tigris.org
For additional commands, e-mail: users-help_at_subversion.tigris.org
Received on 2008-03-25 11:20:41 CET

This is an archived mail posted to the Subversion Users mailing list.

This site is subject to the Apache Privacy Policy and the Apache Public Forum Archive Policy.