Ian Hickson wrote:
>
> On Fri, 17 Aug 2001, JTK wrote:
> >
> > Ian Hickson wrote:
> >>
> >> On Fri, 17 Aug 2001, JTK wrote:
> >>>
> >>> I see no way to do a simple "cvs update". There's gotta be one, could somebody
> >>> please tell me how to do it? And add it to the web site?
> >>
> >> See:
> >>
> >> http://www.mozilla.org/cvs.html
> >
> > Yeah I saw that. I was able to checkout the tree, though why I couldn't
> > just do a "cvs checkout" the same way I can in all other cvs projects
> > (including GCC) still escapes me.
>
> There is a LOT of stuff in the tree, for example Bugzilla and Mozbot are
> in there. Also, a lot of the stuff is not on the HEAD but is on branches,
> to allow easier modularity and so on, for example libpr0n (the image
> library) and NSPR (the portable runtime library) are not in
> "SeamonkeyAll", if I recall correctly.
>
Right, well there's a lot of stuff in the GCC tree too, yet I don't have
any problem when I simply:
cd <gcc_source_root_path>
cvs -z9 -q update -dP
> > Didn't see any mention of how to do a "cvs update" (like I can in all
> > other CVS projects) though.
>
> The makefile I mentioned:
>
> cvs co mozilla/client.mak
> nmake -f client.mak pull_all
>
> ...will do it for you, including taking into account all the side files on
> different branches and so on.
>
Ok, that's second-best anyway. Shouldn't "pull_all" then be renamed
"update"? Or, in the ever-important interests of "not breaking" that
which is already broken, simply add another pseudotarget thus: "update:
pull_all"?
> > Icidentally I tried to do a "cvs update" from the root and it started
> > pulling in all kinds of crazy crap that I don't think I wanted, so I
> > killed it.
>
> Now you see why we don't recommend that! :-)
>
Indeed. May I humbly suggest there's something wrong with your cvs
setup that prevents it from working as expected? IANAcvsL, so perhaps
there's no other way to set things up, but I have to believe that if GCC
of all crazy build-anywhere-anytime-branches-up-the-wazoo programs can
do it, Mozilla can.
