|
From: Colin S. <col...@ex...> - 2004-07-29 12:10:09
|
Jérôme BERNARD wrote: > Colin Sampaleanu wrote: > >> Jérôme BERNARD wrote: >> >>> Colin Sampaleanu wrote: >>> >>>> Dmitriy Kopylenko wrote: >>>> >>>>> >>>>>> I like id keywords...they're mainly in the code through my >>>>>> classes or >>>>>> classes I've edited... (Damn, if it's standardized, I won't know >>>>>> if I've >>>>>> worked on a particular file any more :-) >>>>>> >>>>>> >>>>> I also insert $Id:$ when I work on existing classes or create new >>>>> ones ;-) >>>>> >>>>> Dmitriy. >>>> >>>> >>>> >>>> >>>> >>>> I personally like $Id tags. But they do suck big time when trying >>>> to merge files back in after a branch, causing a conflict/change to >>>> exist when there shouldn't be any... There seems to be a trend to >>>> going away from using them in files, by some people, for this reason. >>> >>> >>> >>> >>> Simply use the "-kk" option. >>> Have a look at: http://www.loria.fr/~molli/cvs/doc/cvs_toc.html#TOC96 >>> Keywords are not an issue at all for branches & merges. >>> >> I think I'm missing something. If you just use -kk, then you don't >> get any actual values substituted, only the keyword itself gets left >> in there. Are you suggesting that for doing the merge, you check out >> both branches with -kk, and then merge using those copies? Sort of a >> pain, although it would work. >> >> Colin > > > I should have sent this link instead > http://www.rons.net.cn/htp/cvs/cvs_5.html#SEC64 :-) > Basically, with the "-kk" option, it simply ignores the "expanded > values". No need to checkout with "-kk". Ok, I see what you are getting at. The issue is that this overrides the binary mode on files that are binary. Not very realistic in a large codebase. Colin |