Обсуждение: $PostgreSQL$ for revision info

Would anyone object to replacing $Header$ and $Id$ with $PostgreSQL$ for
a file version marker?

This would some of us to work in our own repo's without losing the
PostgreSQL version information -- not to mention making diffs easier to
make.

The CVSROOT change consists of adding the below lines to the 'options'
file, then modifying the source files to use $PostgreSQL$

---- options file content ----
tag=PostgreSQL=CVSHeader
tagexpand=iPostgreSQL
------------------------------

This, and it would be really nice if CVSUP repositories could be made
for contrib and interfaces -- or have earthdistance and libpqxx rolled
back into the pgsql-server module.

--
Rod Taylor <rbt@rbt.ca>

PGP Key: http://www.rbt.ca/rbtpub.asc

Rod Taylor <rbt@rbt.ca> writes:
> Would anyone object to replacing $Header$ and $Id$ with $PostgreSQL$ for
> a file version marker?

I don't really care to depend on a nonstandard CVS setup...

> This would some of us to work in our own repo's without losing the
> PostgreSQL version information -- not to mention making diffs easier to
> make.

You might consider checking out with one of the -k options to suppress
the ID markers in your own copy.
        regards, tom lane

> > Would anyone object to replacing $Header$ and $Id$ with $PostgreSQL$ for
> > a file version marker?
> 
> I don't really care to depend on a nonstandard CVS setup...

Data point: Most large projects that have their source routinely
imported into other repositories or import other bits of code, make
heavy use of non-standard $Id$'s for the sake of reducing diff's and
collisions.  The BSDs are the biggest examples and users of this.  I
personally would very much like to see this change as I've grumbled
about doing imports of PostgreSQL into repos that I don't manage and
having lost version information because of colliding $Id$ tags.

Using the options file is pretty standard and has a long history of
use.

http://www.freebsd.org/cgi/cvsweb.cgi/CVSROOT/options

-sc

-- 
Sean Chittenden

Tom Lane wrote:
> Rod Taylor <rbt@rbt.ca> writes:
>> Would anyone object to replacing $Header$ and $Id$ with $PostgreSQL$
>> for a file version marker?
>
> I don't really care to depend on a nonstandard CVS setup...

This is pretty 'standard' and well understood in the *BSD and other worlds.
I cannot recall the exact details, but I believe it requires a minor change
on the CVS server (only).

> You might consider checking out with one of the -k options to suppress
> the ID markers in your own copy.

But then it becomes much harder to generate diffs to the 'real' version
versus your own working copy, which may go through your local repository
many times.

Peter

>>>>> "Sean" == Sean Chittenden <sean@chittenden.org> writes:
   >> > Would anyone object to replacing $Header$ and $Id$ with   >> $PostgreSQL$ for > a file version marker?   >>
>>I don't really care to depend on a nonstandard CVS setup...
 
   Sean> Data point: Most large projects that have their source   Sean> routinely imported into other repositories or
importother   Sean> bits of code, make heavy use of non-standard $Id$'s for the   Sean> sake of reducing diff's and
collisions. The BSDs are the
 

Would be useful for us at TelegraphCQ. I just imported 7.3.2 and went
through a little pain .. 

-- 
Pip-pip
Sailesh
http://www.cs.berkeley.edu/~sailesh

Added to TODO:
* Change CVS $Id$ to $PostgreSQL$


---------------------------------------------------------------------------

Tom Lane wrote:
> Bruce Momjian <pgman@candle.pha.pa.us> writes:
> > Is this a dead issue?  I know several people said they wanted it.
> 
> I'm willing to do it if people want it.  Some of my Red Hat coworkers
> say it'd make their lives easier.
> 
> I'd actually suggest that we standardize on $PostgreSQL$ working like
> $Id$ not $Header$, ie omitting the full-path info.  While I'd rather
> have the path in it myself, this would have the advantage that checkouts
> from the master and anoncvs servers would look the same.  Several people
> have griped about differences between distribution tarballs and anoncvs
> checkouts due to the repository path not being the same.
> 
> Plan B would be to get the repository path to be the same on both
> servers, but that would cause a flag day for either committers or
> noncommitters ...
> 
>             regards, tom lane
> 

--  Bruce Momjian                        |  http://candle.pha.pa.us pgman@candle.pha.pa.us               |  (610)
359-1001+  If your life is a hard drive,     |  13 Roberts Road +  Christ can be your backup.        |  Newtown Square,
Pennsylvania19073