Re: Advice/help wanted on bzr fast-export-from-cvs

Michael Haggerty wrote:
> Ian Clatworthy wrote:

>> 1. Can we remove the restriction that the --options file is
>>    incompatible with most other options? I'd prefer to see other options
>>    override settings in there. See below.
> 
> For some options, it would be easy to remove this restriction.  For
> others, it would be rather awkward, because the options file is generic
> Python code; it doesn't just write information to some "options"
> dictionary, but actually does nontrivial configuration of things that
> are sometimes order-dependent.
> 
> So I suggest that you give me a list of the options that you would like
> to use in conjunction with an options file, and I'll check them out on a
> case-by-case basis.  Would that be doable?

Yes. See below.

>> 2. This one is more a nice to have: if --blob-file and --dump-file
>>    have the same value, then the dump information should be appended
>>    to the blob information.
> 
> This would be easy to implement, with the caveat that OutputPass can
> never be restarted.  Well, I guess one could remember the previous
> length of the file and truncate it if necessary at the beginning of
> OutputPass.

On further thought, let's forget #2. I'd prefer to have a cvs2bzr script
that avoids this hack.

> So why don't we add a top-level cvs2bzr script and a
> cvs2bzr-example.options file within the cvs2svn project?  (Most of the
> code can be shared with cvs2git and maybe the upcoming cvs2hg.)  With
> luck we can select a few more command-line options to be made compatible
> with --options and cover your needs that way.

Right. My cvs2svn branch (see my previous email) has a cvs2bzr script
already. It still needs cvs2bzr-example.options and probably
www/cvs2bzr.html added though.

The options I want to use in combination with --options are:

* --sort
* --trunk-only
* --encoding

Ian C.

------------------------------------------------------
http://cvs2svn.tigris.org/ds/viewMessage.do?dsForumId=1667&dsMessageId=2384203

To unsubscribe from this discussion, e-mail: [[email protected]].