[darcs-users] boringfile and binary problem

[Josh Hoyt]

On 9/17/05, zooko at zooko.com <zooko at zooko.com> wrote:
> 
> > We have a script that generates .pyc files from .py files. We ran it on
> > the root of the repository, which also compiled the .py files in
> > _darcs/current!
> 
> Question: if _darcs were named ".darcs", would your script have skipped it by
> default?

It would not have skipped it, regardless of the name. The script was
just a little to dumb for its own good.

Our main source branch is in its own directory, so compiling only the
files in that directory would have yielded most of the benefit that we
expected, and would have avoided the issue. Now that we realize that
the tool is attempting to work on _darcs, it is just as easy to add it
as a special case to skip.

If we added a whole level of indirection to the top of the repository,
it would simplify other scripts, like find | grep type searches, but
it would make the directory structure more complex. An idle thought is
to invert the relationship between _darcs and the working copy, so
that the working copy is inside of _darcs instead of the other way
around. Adding another level to the hierarchy is basically the same
thing, I suppose.

In my environment, I would set it up like:

~/repos/project1/_darcs/...
                /project1/...
~/work/project1 -> ~/repos/project1/project1

Darcs would be able to find the _darcs directory, and my code would
not have to have special cases for excluding _darcs. This works as
long as I do not add any additional files or directories to the root.
I'll have to try this for my next large-scale project.

Josh