Re: [bug-inetutils] Generated files in CVS?

[Sergey Poznyakoff]

Alfred M\. Szmidt <address@hidden> wrote:

> I fail to see how it makes `cvs diff' useless, it is trivial to ignore
> files in a big diff,

You cannot do it with .cvsignore, since it applies only to files present
in the local copy and missing from the repository. Of course you can do
piping CVS output through a rather trivial sed script, but in my opinion
it is very inconvenient.

Am I missing something?

> We already keep partially generated files in
> the CVS (GNUlib), so I don't see why we can't keep the rest.

Well, yes, that is true. On a side note, these can be omitted from the
repository as well, the way we do in tar, cpio and cflow
repositories. However, here I tend to agree with you that it is better
to have gnulib sources in our repository, simply because we have got no
control over the way they change in gnulib, and sometimes their changes
in gnulib break their usage in the application. Besides, these files do not
change as often as autogenerated files do.

> Also, I find that it isn't easy enough to generate them, sometimes you
> simply do not have the tools on the machine to do it.  Which for my
> part atleast, happens quite often.

Yes, I understand your point. 

Regards,
Sergey

PS: I'd like to ask our colleagues to share their opinions, too.