Community technical support mailing list was retired 2010 and replaced with a professional technical support team. For assistance please contact: Pre-sales Technical support via email to sales@march-hare.com.
"Tony Hoyle" <tmh at nodomain.org> wrote in message news:atrm905vopr2bjds8u2oh7jhunc29r44uv at 4ax.com... > On Fri, 7 May 2004 08:20:49 +0200, "Kevin" <zzz at zzz.zzz.org> wrote: > > cvs -q checkout -P -r branch -jHEAD module (in directory U:\) > I'm not at all sure that is semantically valid - checkout and merge > are normally two separate operations (you need a sandbox to merge to > normally). More to the point, there is no handling of such a case in > the merge code that I can see. I have tested with an example using update instead of checkout: *same effect*. The file is not removed as it should. About the syntax: I use the syntax and it works most of the time. Shouldn't this have the same effect as a checkout followed by a merge? Can this syntax be supported please? This syntax is very useful as you cannot specify a module in update. If there were new folders added (to trunk, for example) I need to specify -d to make sure the folders are created on the branch too. However update gets *all* folders irrespective of the module that I have checked out. How I can do an update + merge filtering the files by module (and getting any new folders created for the module only)? -- Kevin Agius