[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
FW: Why does 'cvs admin -l<branchname>' complain that my branch is absen
FW: Why does 'cvs admin -l<branchname>' complain that my branch is absent?
Thu, 10 Jan 2002 12:32:38 -0800
From: Teala Spitzbarth
Sent: Thursday, January 10, 2002 11:32 AM
To: 'Whitlock, Ginger'; address@hidden
Subject: RE: Why does 'cvs admin -l<branchname>' complain that my branch
I know from other email that Larry has sent me that the way CVS handles
branches can be problematic in some commands - i.e. CVS doesn't actually
create a branched version of each file when the branch is created (that
is why you
see the magic zero in that fake branch revision number for files that
haven't been modified yet). This behavior can sometimes cause conflict
with how RCS
works under the covers. For example, the log command will have
for files on branches when they are first modified, due to the fact that
the "previous" (magic zero) revision on the branch doesn't really exist
at the RCS level, causing the log code to fall into an error condition.
have fixed this is the development version by now...)
I believe the admin -l command is fundamentally an RCS level command -
so I bet
it can't lock unmodified versions on a branch, because at the RCS level,
versions of the file don't really "exist".....
There are versions of commit-info scripts kicking around that allow you
lock branches - you may want to consider one of those. See for example:
(Note: I'm pretty sure this script got munged somehow and the exit 0
should be outside
of the for loop, otherwise recursive commits might slip through onto a
From: Whitlock, Ginger [mailto:address@hidden
Sent: Wednesday, January 09, 2002 7:33 PM
Subject: Why does 'cvs admin -l<branchname>' complain that my branch is
We're running Solaris 8, with CVS 1.11
Might someone help me with an error I'm receiving with attempting to
branch in CVS?
I receive an error complaining that my branch is absent:
.java,v: branch customer1-0_5C1_00E absent
running cvs log to verify the branch exists :
and it does.
I thought I'd found a work around , (cvs admin -bcustomer1-0_5C1_00E
<repository>, to set the default to this branch then
cvs admin -l -L <repository> , to lock what is now the branch at 'tip' ,
followed by cvs admin -b <repository> to set it back to the trunk)
....and this works well as long as every-single file on the branch has
actually been modified, if there are any that have not, the cvs command
lock bombs with an assertion failed error from hash.c
I appreciate any assistance with this matter.
Info-cvs mailing list