[p4] RCS keywords

Mike mikee at mikee.ath.cx
Thu Nov 9 17:15:27 PST 2006


On Thu, 09 Nov 2006, Ivey, William might have said:

> > -----Original Message-----
> > From: Mike [mailto:mikee at mikee.ath.cx]
>  
> > On Thu, 09 Nov 2006, Ivey, William might have said:
> > 
> > > 
> > > We do use it for some human-readable files that will be distributed
> > > outside our development environment. Script files mostly. For most
> > > source files it makes no sense, and for others our build number
> > > system contains enough information to determine the revisions of the
> > > files and this number is added during the build process.
> > 
> > I have not used perforce enough yet to have an opinion about this
> > with respect to perforce, however with rcs and cvs I found having the
> > expanded (oh, and sccs too) ident marks in source code extremely
> > important in that it allows you to use 'ident BINARY' to verify
> > all headers and support files in a binary are the same version
> > and that there is not a build issue. For sccs use 'what BINARY' (I
> > think, it's been a long time for sccs).
> 
> That kind of thing is taken care of by the Perforce metadata so
> you don't need the information in the files themselves. (The
> information that would be inserted into the files is just a copy
> of the metadata info anyway.) In fact, adding the information
> through keyword expansion just complicates merging between branches
> as it introduces non-functional changes which show up in diffs (not
> a big deal, but why worry about it at all).
> 
> To indicate the source files used to build a binary, you generally
> only need to include a single item in one file created at build
> time. This would probably be a change number.

Only if you have the source files at hand. If you only have the
binary the metadata will do no good.

Mike



More information about the perforce-user mailing list