Stephen Leake, 2007-06-29:

 > The code is certainly nicely structured and easy to follow.

Thanks :)


 > I'd like to add options to control what is displayed (no knowns, no
 > ignored). Is there a naming convention for user options like this?

Actually, the code attempts to hide known and ignored files (as the only 
option for now), but I accidentally inverted the conditional, so it 
shows just the files it's supposed to hide.  I fixed this in my branch.

So, at the moment, the default probably suits your needs.  But if you 
want to have the ability to turn things on and off, I think it should be 
implemented as an interactive, buffer-local toggle.  It would make sense 
to have the global default value customizable.

xmtn doesn't currently have any defcustom forms because I haven't gotten 
around to introducing any.  In my description of xmtn's architecture, 
options for the status display would fall under "user-visible 
functionality", so they should go in xmtn-dvc.el.  The alternative would 
be creating a file xmtn-custom.el if there is some technical reason why 
it's better to collect all defcustom forms in a separate file.  (I don't 
know.)


 > I'm guessing 'xmtn-status-display-known',
 > 'xmtn-status-display-ignored' (both of which I would set to 'nil')
 > would be good.

Yes, but nil should be the default anyway.


 > I'll experiment with the format; I definitely prefer words (like
 > "added") to letters ("a"). 'status' can contain up to three values if
 > complex renaming is involved. That could make the lines very long. I
 > don't think ewoc handles two-line elements?

Yes, feel free to change the format.

Ewoc does handle multi-line elements, nothing special about them IIRC. 
However, I think we only need them for renamings.  My guess is that 
other elements should fit in a single line if we print only the file 
name and show the directory in a separate line (like PCL-CVS does).  The 
layout will be easier to scan visually if the majority of entries uses 
just one line each in the common case.

The entire dvc-diff-printer machinery could definitely use some 
restructuring, and maybe this is a good time to start with that.  The 
code and interfaces are still rather tla-specific.  Oh, and the fact 
that status and diff use the same major mode and data structures really 
bugs me.  But fixing that is probably rather disruptive for other back-ends.

Christian.


_______________________________________________
Dvc-dev mailing list
[email protected]
https://mail.gna.org/listinfo/dvc-dev

Reply via email to