On 03/14/2011 01:49 PM, 'Daniel Shahaf' wrote:
> Stefan Sperling wrote on Mon, Mar 14, 2011 at 18:36:21 +0100:
>> I haven't yet seen any good suggestion for an alternative approach.
>>
>> Dry run implies carrying out every step made during an update expect
>> making modifications to the working copy. The suggestion to put dry-run
>> into a separate editor is even worse because it duplicates at lot more
>> conditionals (for update logic) in a new place than this patch is adding
>> to existing places.
>>
>> If there is a cleaner way to do it, fine. Maybe Arwin would like to learn
>> about this better approach and try to implement it, for update and for merge.
>> But is there?
>
> I'd have expected this discussion to take place prior to anyone writing a patch.
That is the ideal, sure, but sometimes(?) Reality misses the mark of
perfection. It's easy for us all to get busy doing our own thing, and
easier still to assume somebody else will handle something that we don't
wish to handle -- or at least don't wish to take the time to handle right now.
Back in the day (before more of us understood our server logic), we used to
joke that the best way to get Greg Stein to improve mod_dav_svn was just to
commit a broken implementation of whatever feature we were hoping he'd
write. :-) Greg was always good for code review, but not always so readily
available for pre-coding consultation. I don't say this as an indictment of
Mr. Stein (whose contributions are voluntary just like most everyone else's
are here), but as an example of just the realities of open source software
dynamics.
And sometimes, for whatever reason, it's just easier to communicate about
certain topics in C than in English. 'Round here, we'll all expected to be
bilingual that way. :-)
--
C. Michael Pilato <cmpilato_at_collab.net>
CollabNet <> www.collab.net <> Distributed Development On Demand
Received on 2011-03-14 19:01:00 CET