On 17-Jun-07, at 3:41 AM, [EMAIL PROTECTED] wrote:
I just thought of 2 disgusting "solutions" :)
Way 1:
- Create a new repository.
- Add those files, record a patch. Save the "_darcs/inventory" file
somewhere.
- You can now work locally with this repository.
- If you publish it, make sure nobody can read the first patch or
anything in "_darcs/pristine" or anything outside of "_darcs".
- To make other people able to pull from your repository,
distribute the
saved "_darcs/inventory" file and instruct people how to build a
working repository (darcs init, replace inventory, put the set of
files into the right place in the top dir AND under _darcs/pristine)
I tested this locally and it should work (it breaks "darcs repair" for
everybody since they are missing a patch)
Tobi, I appreciate that you tested this, but it does not satisfy my
criteria because it requires that I distribute the inventory, which
contains the files that I do not want to distribute.
However, could I not do as you suggest - not distribute the
inventory, but rather give instructions on how to reproduce the
inventory? Would that work? (I'm a Darcs newbie, so I can't see if
it will work or not.)
Way 2:
Make the first patch (the one with your big files) 100% reproducible.
In the current darcs repository structure this is easily possible
using
"darcs record -y -l --pipe" and always passing the same input to
it. So
you could distribute a script which checks for the presence of the
files, creates the init patch and then pulls from your patch
repository.
This has the risk that users could use darcs to pass this init patch
around...
I think this might work. But, what is the "-y" flag? Also, does --
pipe suggest that darcs is going to get more than one file piped to
it? What exactly does --pipe cause darcs to expect?
I'm okay with users passing that initial patch around - that's their
business, not mine.
_______________________________________________
darcs-users mailing list
[email protected]
http://lists.osuosl.org/mailman/listinfo/darcs-users
