On Thu, Apr 11, 2013 at 12:34 AM, Pavel Sanda <[email protected]> wrote:
> Nico Williams wrote:
>> I see the smiley.  I hope to convince you to stop the use of RCS :)
>
> Then you will need to show better arguments :) I was talking about single
> lyx documents 50-400 pages, not about linux kernel.
> I'm really happy that I can use RCS, because I suspect the fileformat
> and commandline switches I need to maintain in codebase or tinker with
> will be the same after 20 years when 50% of current VCSs around will be
> unmaintained corpses. So durability from 70s is advantage!

Even for single-file LyX "projects", if I render into text form or
HTML or LaTeX, I like to commit the rendered forms as well: so I can
browse rendered file history.  Suddenly a single-file project ==
multi-file project and I care about commit atomicity.

I also have entire swaths of my home directory in git (being very
careful not to commit files with passwords or keys in them).  This
nicely covers the one-file LyX document case, and it's much easier to
use than RCS (no need to lock/unlock files, checkout/checking, just
git add/commit and move alont).

>> Simple for one file.  Not simple for more than one.  Utterly useless
>> for any project with more than a handful of files, or more than a
>> handful of people editing the same files, or when you can't use
>> NFS/whateverFS.
>
> No doubt, but I was not really talking about zillion file projects
> but about real life scenario with LyX. Another scenarios are possible
> as well, hence is silly to try fight whether git or RCS in general
> terms.

Indeed.  I just want LyX to not ever run a git checkout command.  Not
even give users the option: it can be destructive!  I also want LyX to
limit itself to informational actions only in the case of git-like
VCSes (Mercurial, Fossil, ...).  Please don't let this get lost in the
noise about how to handle simple version control tasks.

Nico
--

Reply via email to