svn commit: r282572 - vendor/OpenBSD/dist/usr.bin/rcs

[Pedro Giffuni]

On 05/07/15 09:42, Bryan Drewery wrote:
> On 5/6/2015 6:52 PM, Pedro F. Giffuni wrote:
>> Author: pfg
>> Date: Wed May  6 23:52:15 2015
>> New Revision: 282572
>> URL: https://svnweb.freebsd.org/changeset/base/282572
>>
>> Log:
>>    Bring OpenRCS 20150326 to the vendor area.
>>    
>>    A BSD-licensed alternative to GNU Revision Control System.
>>    
>>    Obtained from:	OpenBSD
>>
>> ...
> I'm seeing a lot of discussion about this in private areas. The majority
> of it negative. We might as well get some conversation going.

Well, last time something was done ... it was basically the *surprise*
removal of RCS during a code freeze. There was a lot less discussion
there (and it went consequently bad).

I have a patch to make OpenRCS build but I have no plans to do a
surprise merge. I did plan to raise the issue up on the lists.

> I have no major opinion on RCS except that it can easily be a package.
> We can provide a package on the release media as well.

GNU RCS is already packaged (devel/rcs57). Still, having a better
license means we can accept patches and do some basic
maintenance.

> However, as long as we have $FreeBSD$ tags in our own source files and
> compiled binaries I find ident(1) to be very useful for debugging. I
> used it just the other day to see what revision someone's /bin/sh was.
> We use it in the ports package builder to get an idea of what revision
> the port is on (it is not 100% accurate since someone may have modified
> distinfo and we are checking Makefile's ident, but it is still very
> useful to give an idea of the range of commits).
>
> Can we please consider at least importing ident(1) and disassociating it
> from WITHOUT_RCS?
>

I don't see myself doing much more than offering the patch for
openrcs. Rest assured that I won't be committing openrcs
to the base system (at least not soon and certainly not without
discussion).

Pedro.