Re: CVS update [cvs1-11-x-branch]: /ccvs/src/

[Derek Robert Price]

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

mdb@cvshome.org wrote:

>Tag: cvs1-11-x-branch
>User: mdb     
>Date: 04/11/17 22:52:15
>
>Modified:
> /ccvs/src/
>  ChangeLog, sanity.sh
>
>Log:
> * sanity.sh (basic2-21a): The val-tags file should have
> at least 'rtagged-by-head y' in it.


Actually, this was the behavior previous to the fix I installed on
feature.  Despite some question in the comments as to the validity of
this approach, it was already being tested for and enforced in both
test suites prior to my enhancement to feature.   From sanity.sh:

              # Same "Up-to-date check" issues as in rmadd-14.
              # The "no such tag" thing is due to the fact that we only
              # update val-tags when the tag is used (might be more of a
              # bug than a feature, I dunno).
              dotest_fail rmadd-17 \
    "${testcvs} -q ci -r mynonbranch -m add file4" \
    "${PROG} \[commit aborted\]: no such tag mynonbranch"
              # Try to make CVS write val-tags.
              dotest rmadd-18 "${testcvs} -q update -p -r mynonbranch
    file1" \
    "first file1"
              # Oops, -p suppresses writing val-tags (probably a
    questionable
              # behavior).
              dotest_fail rmadd-19 \
    "${testcvs} -q ci -r mynonbranch -m add file4" \


Comments in the source itself fully document the "add on update"
approach.  I think it was considered desirable once (mabe due to lack
of code duplication?), since tags missing from val-tags sometimes
needed to be searched for in the repository and added on reads anyhow
since an earlier version of CVS (from before val-tags) might have
created the tag.

I suppose if someone else is going to write and apply the patch, I
don't have any major objections.  It shouldn't be very destabilizing,
but I'm not sure it is worth the effort of back porting, myself,
especially as I am hoping that we can bless the current feature as
stable soon (January or February?), probably after a few RC (release
candidate) releases (say, December through January), then retire
1.11.x anyhow.

> (basic2-21b): New test showing a cvs bug when val-tags
> is not properly updated.


Assertion failures are, I will agree, bugs.  :)

Cheers,

Derek

- --
                *8^)

Email: derek@ximbiot.com

Get CVS support at <http://ximbiot.com>!
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.3 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iD8DBQFBnQrCLD1OTBfyMaQRAmJfAJ9T9N0Yoj5gk6CuMnlim/EHDMcdOQCeO2+B
qtVukD1r7JCnzt4thkEu4Ko=
=tfti
-----END PGP SIGNATURE-----