info-cvs
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

behavior change for 'cvs update -jrev1 -jrev2' soon to occur


From: Mark D. Baushke
Subject: behavior change for 'cvs update -jrev1 -jrev2' soon to occur
Date: Wed, 22 Oct 2003 11:11:42 -0700

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1


If you attempt to issue a

  cvs update -j xxx -j yyy

where yyy is the base version of a file in your
directory, cvs currently ignores the command, because
it assumes that the change must already be in your
local version, and there's no reason for you to have
gone to the effort of deleting the code only to request
it be readded.

There are some people who effectively do wish to delete
code and then request it to be re-applied (it takes all sorts).

Does anyone object to CVS actioning the "re-apply" request?

As an example, 

    commit a file with the lines
    aaa
    bbb
    ccc
    
    add "ddd" to the end.

    then commit that.
    
    then delete ddd from the file.
    
    then do a cvs update -j 1.1 -j 1.2 to put the ddd back in.

The current versions (1.11.9 and 1.12.1) of CVS will
ignore you.

Another example:

    commit a file with the lines
    aaa
    bbb
    ccc
    ddd

    remove the "ddd" line.

    then commit that.

    then add "ddd" to the end of the file.

    then do a cvs update -j 1.1 -j 1.2 to remove it again.

The current versions (1.11.9 and 1.12.1) of CVS will
ignore you.

A final example:

    commit a file with the lines
    aaa
    bbb
    ccc

    add a "ddd" line.

    then commit that.

    then change "ddd" to "dddd".

    then do a cvs update -j 1.1 -j 1.2 to cause
    conflict markers to appear along with both the
    "ddd" line and the "dddd" line so that you can more
    carefully consider how to accomodate the original
    version and your new change.

The current versions (1.11.9 and 1.12.1) of CVS will
ignore you.

A change that will be committed shortly will tell CVS
to not ignore your request and instead will let diff3
figure out what needs to happen.

It is believed that this new behaviour, acknowledging
the "re-apply" request, is consistent with both the
documentation, logic, and the behaviour of diff3.

If you rely on the old behaviour, please elaborate on the
circumstances where you issue the join command expecting
it to do nothing.

Thanks go to Paul Edwards for reporting the problem and
for most of the description above (expanded and changed
by me, so any errors in clarity are likely my fault).

A full thread of this discussion is available on the
address@hidden mailing list archived at mail.gnu.org
the first message in the thread begins with the URL:
http://mail.gnu.org/archive/html/bug-cvs/2003-10/msg00199.html

        Thank you,
        -- Mark
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.3 (FreeBSD)

iD8DBQE/lshe3x41pRYZE/gRAnznAKDEFAFu+YRQAW6xVH8ZbHoPjEoRFgCgshbX
/EP59ufuva1qrELqd3rc15Y=
=Y2eS
-----END PGP SIGNATURE-----




reply via email to

[Prev in Thread] Current Thread [Next in Thread]