On 07/08/2010 09:32 AM, Igor Sereda wrote:
>> I like the idea of this change, but I wonder if it can be made without
>> introducing a new command-line option. Your "expectations" as listed above
>> certainly make sense. That is, until you actually read the built-in
>> documentation found in the program's usage message. :-)
>
> I see what you mean. Some thoughts:
>
> 1. I think "--drop-empty-revs" could have been something like
> "--filter-revisions" instead, which would mean, keep revision if and
> only if at least one node passes include/exclude filter. In that case,
> empty revisions could be filtered by "include --filter-revisions /".
Heheh. "Could have been" is a luxury we can always entertain but rarely do
anything about.
A new option for the behavior you propose could be as simple as
--drop-all-empty-revs.
> 3. While indeed help message for --drop-empty-revs matches the
> behavior, the *comment* in the current code does not:
>
> /* Revision is written out in the following cases:
> 1. No --drop-empty-revs has been supplied.
> 2. --drop-empty-revs has been supplied,
> but revision has not all nodes dropped
> 3. Revision had no nodes to begin with.
> */
>
> Obviously, the guy who wrote the comment expected the same behavior as
> I did, but the resulting code was a bit different. :)
Yeah, that may be true. But as you well know, our users will be banking on
the user-visible documentation and current behavior more so than our code
comments. :-)
--
C. Michael Pilato <cmpilato_at_collab.net>
CollabNet <> www.collab.net <> Distributed Development On Demand
Received on 2010-07-08 16:25:47 CEST