You are viewing a plain text version of this content. The canonical link for it is here.
Posted to users@subversion.apache.org by an...@amd.com on 2004/05/27 17:17:31 UTC
RE: (--incremental ) was backup strategy - and unrecoverable
data bases.
> Quite a few of the projects I work on use code generators
> (lex/treecc/jay etc.) -which regenerate 3000+ line files on each
> revision. - it quickly adds up.. (although only 10-50 lines actually
> change on each commit..)
I expect that people wll jump on you, saying
"Don't check in generated files."
Not me: I sometimes check in generated files,
e.g. to protect against the possibility
of some unanticipated dependency like a change
in compiler version resulting in the inability
to regenerate exactly the historical files.
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org
Re: (--incremental ) was backup strategy - and unrecoverable data bases.
Posted by Gary Feldman <g1...@marsdome.com>.
>From: <an...@amd.com>
>Sent: Thursday, May 27, 2004 1:17 PM
> Not me: I sometimes check in generated files,
I do, too, but not in the development tree. When such things are needed,
there's a separate release tree that gets the important results of a build.
They're checked in as part of the build or release automation, and
developers needn't deal with the problem on a day to day basis.
There are also circumstances in which it makes sense to check in the
compiler, linker, other tools, and even a list of OS patches and
configuration items (but not the whole OS).
Gary
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org
Re: (--incremental ) was backup strategy - and unrecoverable data bases.
Posted by Glenn Maynard <g_...@zewt.org>.
On Thu, May 27, 2004 at 10:17:31AM -0700, andy.glew@amd.com wrote:
> > Quite a few of the projects I work on use code generators
> > (lex/treecc/jay etc.) -which regenerate 3000+ line files on each
> > revision. - it quickly adds up.. (although only 10-50 lines actually
> > change on each commit..)
>
> I expect that people wll jump on you, saying
> "Don't check in generated files."
If only 10-50 lines are actually changing, then it doesn't matter that
it's generated; it should be a tiny diff, anyway. It's no different
than checking in a 3000+ line plain source file.
I often check in prebuilt Windows binaries and libraries, since that
environment is painful to automate.
--
Glenn Maynard
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org