2012/1/31 Jacek Cała <[email protected]>

> > fossil commit -m "Some changes in dir" --filerange [2-3]
>
> or
>
> > fossil commit -m "Some changes in somefiles" --filerange 1,3
>

That's an interesting idea. The first syntax wouldn't be ideal because the
[ and ] operators actually mean something to unix shells, but the second
form, including 1-3 could be done. In fact i have some C code for parsing
ranges like that in another project of mine, and it would be easy enough to
port that in. The only problem i see at the moment is that such numbers are
transient - they potentially change with any given file-level change. e.g.

fossil changes
1 abc
2 def
3 xyz

then i edit file "lmn" and run:

fossil commit -m '...' 1-3

it "could be" (and likely will be) that file lmn is now #3 (assuming
lexical ordering).

Fossil "could", as part of "changes" or "status" record the last known
list, and check that list before allowing a commit to be performed, and
erroring out if (A) the list has changed since the last "changes"/"status"
and (B) the user supplied a numeric list on the CLI.

A secondary (but probably minor) problem is that this syntax prohibits
numerically named files (probably a rarity, but someone out there probably
has 1 or 2 of them (no pun intended).

useful. Also, could Richard, Stephan and others who spent more time on
>

i think that's the first time someone has accused me of being familiar with
the fossil innards ;).

i think the idea is interesting and quite possibly doable with relatively
little effort, but i also know that i tend to miss some of the "bigger
picture" pitfalls which others on this list historically point out for, so
i'd like to wait and see what some of the others say...

-- 
----- stephan beal
http://wanderinghorse.net/home/stephan/
http://gplus.to/sgbeal
_______________________________________________
fossil-users mailing list
[email protected]
http://lists.fossil-scm.org:8080/cgi-bin/mailman/listinfo/fossil-users

Reply via email to