I would like to include a MANIFEST file automatically when doing 'darcs
dist'

This doesn't work, because the _darcs directory is not present at
predist time:

 darcs setpref predist "darcs query manifest>MANIFEST" 

Since I can't think of a good workaround, I request the following:

 darcs dist --manifest=FILE Adds current manifest to FILE, defaulting to the name MANIFEST  

Then we can set this behavior as a default preference if we always want it.

	Mark

One workaround is to type (or put in the makefile)
'DARCS_REPO_DIR=$PWD darcs dist', and use 'darcs query manifest
--repo-dir=$DARCS_REPO_DIR' in the predist script.

Perhaps darcs should export DARCS_REPO_DIR to the predist
command?

On Thu, May 25, 2006 at 07:57:16PM +0000, Tommy Pettersson wrote:
> 
> Tommy Pettersson <ptp@lysator.liu.se> added the comment:
> 
> One workaround is to type (or put in the makefile)
> 'DARCS_REPO_DIR=$PWD darcs dist', and use 'darcs query manifest
> --repo-dir=$DARCS_REPO_DIR' in the predist script.
> 
> Perhaps darcs should export DARCS_REPO_DIR to the predist
> command?

That would help.

Thanks for the response, Tommy.

	Mark

Two questions:

1. Do we still want dist --manifest, if we export DARCS_REPO_DIR in predist?

2. Would we be making it easy for the user to shoot themselves in the foot? 
Easy in the sense that they would do something which looks innocent, but is
actually a really bad idea?

On Monday 06 August 2007 11:45, Eric Kow wrote:
> Eric Kow <eric.kow@gmail.com> added the comment:
>
> Two questions:
>
> 1. Do we still want dist --manifest, if we export DARCS_REPO_DIR in
> predist?
>
> 2. Would we be making it easy for the user to shoot themselves in the foot?
> Easy in the sense that they would do something which looks innocent, but is
> actually a really bad idea?

As the original reporter, my desire for this feature has diminished. 

I find I nearly always need to include some generated files in MANIFEST, which 
aren't managed by darcs, meaning this feature wouldn't be a solution for me. 

  Mark

Ok... I'm marking this as a wont-fix, especially since we have a seemingly
acceptable workaround anyway.  Objections, etc, welcome.

This remains wont-fix, but I've opened Tommy's suggestion as a new ticket, issue1555