You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@subversion.apache.org by Max Bowsher <ma...@ukf.net> on 2005/12/01 22:52:35 UTC

Reverting Issuezilla policy from 'UNCONFIRMED/NEW' to '---/unscheduled'

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

A while ago, we switched from indicating 'untriaged issue' with
target_milestone = '---', to status = UNCONFIRMED.

This seemed like a good idea at the time, but unfortunately, CEE's
current permissions setup really isn't friendly to this scheme.


The problems:
* Having the Observer role allows people to skip right past the
UNCONFIRMED status during initial issue submission.

* Having the Observer role allows people to change issues from
UNCONFIRMED -> NEW, via a radio button marked "Confirm issue" - to those
not already familiar with Bugzilla, it sounds like they _ought_ to do that.

* Once an issue has been confirmed, there's no way to unconfirm it, if
the confirmation was done erroneously.


Therefore, it's time to go back to the old scheme:

Issues are initially created with target_milestone = '---'. It is _not_
possible to override this during initial issue submission.
This value indicates 'untriaged'. It is set to something else
('unscheduled', if nothing is more appropriate, once triage is done).

Observers are capable of changing the target_milestone, but the UI isn't
such that it leads them to do it. For the very rare case when it is
changed inappropriately, it will be noticed on the issues@ list, and can
easily be undone.


Max.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.1 (Cygwin)

iD8DBQFDj36zfFNSmcDyxYARAg3pAJ0U7Z2v0nlV+/zYd5BZkWArU27S0QCgmn13
HnW1Df8fjBhEjc6u0wtY260=
=/4gM
-----END PGP SIGNATURE-----

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org

Re: Reverting Issuezilla policy from 'UNCONFIRMED/NEW' to '---/unscheduled'

Posted by kf...@collab.net.
Max Bowsher <ma...@ukf.net> writes:
> > Therefore, it's time to go back to the old scheme:
> > 
> > Issues are initially created with target_milestone = '---'. It is _not_
> > possible to override this during initial issue submission.
> > This value indicates 'untriaged'. It is set to something else
> > ('unscheduled', if nothing is more appropriate, once triage is done).
> > 
> > Observers are capable of changing the target_milestone, but the UI isn't
> > such that it leads them to do it. For the very rare case when it is
> > changed inappropriately, it will be noticed on the issues@ list, and can
> > easily be undone.
> 
> Done.

I have a list of (I think) all the issues that were automatically
swept from '---' to UNCONFIRMED during the earlier changeover.  Would
you like that list?  Well, here it is, cheap enough to just include it
in this mail anyway.  This doesn't include issues that were closed or
otherwise disposed of during the interregnum, of course:

  #2113    svn add read-only-dir fails
  #895     renames delete by name, not by object identity
  #553     test suite tree construction/comparison awkward
  #1610    confusing output from no-op merge
  #630     svn_ra_dav__get_file needs to handle base text case
  #1682    'svn blame' should adjust to max username width
  #1663    keyword contraction as a local mod may be invisible to svn
  #2312    HTTP Accept-Encoding header included twice in requests from ra_dav
  #1894    [PATCH] Initial German book translation
  #981     tree copies can be O(N) when copyfrom args not used
  #877     update_tests.py:update_missing() works around test suite bug
  #1125    Python SWIG bindings don't install correctly
  #2019    httpd child processes waiting for locks can hit max_child limit
  #2118    property merging unnecessarily strict
  #950     svn-config installedness test fails when symlinked
  #1700    Hook script (symbolic link) ignored when broken
  #1162    property timestamps not being updated
  #574     WIN32-specific code in libsvn_subr/io.c, config.c
  #1860    svnserve not sending commit-info when post-commit fails
  #958     keywords that expand just once, on commit
  #540     need real doc string for svnlook/main.c:print_ids_tree()
  #1959    svn could report version of server(s) as well as client
  #960     separate canonical repository URL from repos path in wc
  #735     import should 'mkdir -p'
  #948     Future-proofing 'svn status' output with a --describe option
  #885     MBK's patch to add --ignore & related config option
  #1611    'actual' status trees misconstructed in regression test framework
  #1584    prevent 'svn:' revision props from being set on files/dirs
  #1696    svnadmin load can leave repository needing recovery
  #458     does using change_dir_prop to set wcprops affect trace editor?
  #781     some commands could use a depth indicator option
  #427     "make clean" and "make doc" need to be smarter about docs
  #2122    'svn status -u' can crash with partial -N working copy
  #409     writer locking currently suboptimal
  #910     rewrite svnlook's option and argument parsing to use svn_opt
  #436     libsvn_fs/tree.c:txn_body_copy() - Allow copies from txn roots
  #869     import should use .svn area to get props
  #807     gracefully degrade from failed charset conversion
  #1294    rethink .svn/ area read-only files and dirs strategy
  #1957    Merge misreports bogus conflict.
  #1683    svn_auth.h improvements
  #952     Pass stdin to pre-rev-prop-change hook script.
  #2333    'svn diff URL1 URL2' not reverse of  'svn diff URL2 URL1'
  #1971    certain property names cause non-wf XML responses
  #599     Need a feature like CVS's vendor branches

-Karl

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org

Re: Reverting Issuezilla policy from 'UNCONFIRMED/NEW' to '---/unscheduled'

Posted by Max Bowsher <ma...@ukf.net>.
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Max Bowsher wrote:
> A while ago, we switched from indicating 'untriaged issue' with
> target_milestone = '---', to status = UNCONFIRMED.
> 
> This seemed like a good idea at the time, but unfortunately, CEE's
> current permissions setup really isn't friendly to this scheme.
> 
> 
> The problems:
> * Having the Observer role allows people to skip right past the
> UNCONFIRMED status during initial issue submission.
> 
> * Having the Observer role allows people to change issues from
> UNCONFIRMED -> NEW, via a radio button marked "Confirm issue" - to those
> not already familiar with Bugzilla, it sounds like they _ought_ to do that.
> 
> * Once an issue has been confirmed, there's no way to unconfirm it, if
> the confirmation was done erroneously.
> 
> 
> Therefore, it's time to go back to the old scheme:
> 
> Issues are initially created with target_milestone = '---'. It is _not_
> possible to override this during initial issue submission.
> This value indicates 'untriaged'. It is set to something else
> ('unscheduled', if nothing is more appropriate, once triage is done).
> 
> Observers are capable of changing the target_milestone, but the UI isn't
> such that it leads them to do it. For the very rare case when it is
> changed inappropriately, it will be noticed on the issues@ list, and can
> easily be undone.


Done.


Max.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.1 (Cygwin)

iD8DBQFDkC3vfFNSmcDyxYARAixJAJ9VyfsyLZXpDNETHIQ0HNdmbbb/9wCdF4y1
N+w6WBYYRtsuILqVfDlGDh8=
=kjpJ
-----END PGP SIGNATURE-----

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org