[PATCH v4] test: emacs: add test for `notmuch-search-operate-all'

Dmitry Kurochkin dmitry.kurochkin at gmail.com
Wed Feb 1 15:25:40 PST 2012


Hi Pieter.

On Wed, 01 Feb 2012 21:32:32 +0100, Pieter Praet <pieter at praet.org> wrote:
> On Wed, 01 Feb 2012 18:06:05 +0400, Dmitry Kurochkin <dmitry.kurochkin at gmail.com> wrote:
> > On Wed, 01 Feb 2012 14:44:48 +0100, Pieter Praet <pieter at praet.org> wrote:
> > > On Mon, 30 Jan 2012 12:13:48 +0400, Dmitry Kurochkin <dmitry.kurochkin at gmail.com> wrote:
> > > > Hi Pieter.
> > > > 
> > > > On Mon, 30 Jan 2012 08:45:50 +0100, Pieter Praet <pieter at praet.org> wrote:
> > > > > `notmuch-search-operate-all' (bound to "*") adds and removes tags
> > > > > to/from all messages which match the query used to populate the
> > > > > current search buffer.
> > > > > 
> > > > > ---
> > > > > 
> > > > > Rebased to current master.
> > > > > 
> > > > > Previous versions (chronologically):
> > > > > - id:"1309450108-2793-1-git-send-email-pieter at praet.org"
> > > > > - id:"1309762318-4530-5-git-send-email-pieter at praet.org"
> > > > > - id:"1310313335-4159-5-git-send-email-pieter at praet.org"
> > > > > 
> > > > 
> > > > This looks like a useful patch series.  We definitely need more tests
> > > > for tagging operations in the Emacs UI.  Do you plan to revive it?
> > > > 
> > > 
> > > Absolutely.  In fact, it's still alive and kicking on a local branch :)
> > > 
> > > 
> > > > Note that not so long ago I posted a bunch of tagging-related patches
> > > > [1] that would conflict with this patch at least because of
> > > > `notmuch-search-operate-all' being renamed to `notmuch-search-tag-all'.
> > > > 
> > > 
> > > Nice series!  Will probably be pushed in the next few days,
> > > so I'll hold off on resubmitting the tests until then.
> > > 
> > > 
> > > > > 
> > > > >  test/emacs |   19 +++++++++++++++++++
> > > > >  1 files changed, 19 insertions(+), 0 deletions(-)
> > > > > 
> > > > > diff --git a/test/emacs b/test/emacs
> > > > > index 8ca4c8a..e94ad94 100755
> > > > > --- a/test/emacs
> > > > > +++ b/test/emacs
> > > > > @@ -124,6 +124,25 @@ test_emacs "(notmuch-show \"$os_x_darwin_thread\")
> > > > >  output=$(notmuch search $os_x_darwin_thread | notmuch_search_sanitize)
> > > > >  test_expect_equal "$output" "thread:XXX   2009-11-18 [4/4] Jjgod Jiang, Alexander Botero-Lowry; [notmuch] Mac OS X/Darwin compatibility issues (inbox unread)"
> > > > >  
> > > > > +test_begin_subtest "Add/remove tags to/from all matching messages."
> > > > > +test_emacs '(notmuch-search "tag:inbox AND tags")
> > > > > +	    (notmuch-test-wait)
> > > > > +	    (notmuch-search-operate-all "+matching" "-inbox")
> > > > > +	    (notmuch-search "tag:matching AND NOT tag:inbox")
> > > > > +	    (notmuch-test-wait)
> > > > > +	    (test-output)'
> > > > > +cat <<EOF >EXPECTED
> > > > > +  2009-11-18 [3/3]   Adrian Perez de Castro, Keith Packard, Carl Worth  [notmuch] Introducing myself (matching signed unread)
> > > > > +  2009-11-18 [1/3]   Carl Worth, Israel Herraiz, Keith Packard       [notmuch] New to the list (inbox matching unread)
> > > > > +  2009-11-18 [2/2]   Keith Packard, Carl Worth    [notmuch] [PATCH] Make notmuch-show 'X' (and 'x') commands remove inbox (and unread) tags (matching unread)
> > > > > +  2009-11-18 [1/2]   Keith Packard, Alexander Botero-Lowry    [notmuch] [PATCH] Create a default notmuch-show-hook that highlights URLs and uses word-wrap (inbox matching unread)
> > > > > +  2009-11-18 [1/1]   Jan Janak            [notmuch] [PATCH] notmuch new: Support for conversion of spool subdirectories into tags (matching unread)
> > > > > +  2009-11-18 [1/1]   Stewart Smith        [notmuch] [PATCH] Fix linking with gcc to use g++ to link in C++ libs. (matching unread)
> > > > > +  2009-11-17 [1/2]   Ingmar Vanhassel, Carl Worth  [notmuch] [PATCH] Typsos (inbox matching unread)
> > > > > +End of search results.
> > > > > +EOF
> > > > > +test_expect_equal_file OUTPUT EXPECTED
> > > > 
> > > > I am worried that this test would break because of changes in other
> > > > tests.  E.g. if a new test adds a new message which matches "tag:inbox
> > > > AND tags", this test would have to be updated.  I think we should avoid
> > > > this.  I see the following options here:
> > > > 
> > > > * Search for messages which are less likely to change, e.g. "from:carl".
> > > > 
> > > > * Rework the test to avoid using any fixed expected results, e.g.:
> > > > 
> > > >   - count all messages with tag:inbox
> > > > 
> > > >   - remove inbox tag, add some other distinct tag for all messages with
> > > >     tag:inbox
> > > > 
> > > >   - count all messages with tag:inbox again, check that it is 0
> > > > 
> > > >   - add the inbox tag back, remove the previously added tag, check the
> > > >     message count
> > > > 
> > > > I like the latter approach because it does not compare Emacs UI output
> > > > and hence would not break when that output changes.  What do you think?
> > > > 
> > > > Also, we should leave notmuch db in the same state as it was before the
> > > > test if possible.
> > > > 
> > > 
> > > Good point(s)!
> > > 
> > > How about this:
> > > 
> > >   #+begin_src sh
> > >     test_begin_subtest "Add/remove tags to/from all matching messages."
> > >     expected=$(notmuch count from:cworth AND tag:inbox)
> > >     test "${expected}" == "0" && expected="Need more matches!" # prevent false positives
> > >     test_emacs "(notmuch-search \"from:cworth AND tag:inbox\")
> > >             (notmuch-test-wait)
> > >             (notmuch-search-operate-all \"+matching\" \"-inbox\")"
> > >     output=$(notmuch count from:cworth AND tag:matching AND NOT tag:inbox)
> > >     notmuch tag -matching +inbox -- from:cworth AND tag:matching AND NOT tag:inbox # restore db state!
> > >     test_expect_equal "$output" "$expected"
> > >   #+end_src
> > > 
> > 
> > I think "from:cworth" is not needed here.  In my previous email I meant
> > this as an alternative option to testing using counts.
> > 
> 
> I know, but I decided to include it anyway, on the basis that it
> increases the probability of the search buffer containing threads with
> matched *as well as* unmatched messages, thus providing more insurance
> wrt `notmuch-search-operate-all' doing The Right Thing (TM).
> 
> 
> > Probably we should add more counts in the test?  E.g.:
> > 
> > * before tagging
> >   - count tag:matching, it must be zero
> > 
> > * after -inbox +matching
> >   - count tag:matching, it must be equal to $expected
> >   - count tag:inbox, it must be zero
> > 
> > * after +inbox -matching
> >   - count tag:matching, it must be zero
> >   - count tag:inbox, it must be equal to $expected
> > 
> > We can compare it in a single test_expect_equal call.  Actual and
> > expected results may be smth like:
> > 
> >   initial tag:matching count: n1
> >   tag:inbox count after tagging: n2
> >   tag:matching count after tagging: n3
> >   tag:inbox count after restoring: n4
> >   tag:matching count after restoring: n5
> > 
> > In case of failure, we should get a nice diff.
> > 
> 
> OK, how about this?:
> 

Looks good.  Minor comments below.

>   #+begin_src sh
>     test_begin_subtest "Add/remove tags to/from all matching messages."

We should add notmuch-search here, because similar functionality would
be available in notmuch-show soon.  Also, I suggest replacing
"add/remove tags to/from" with "change tags".  Consider:

  Change tags for all matching threads in notmuch-search.

>     old="inbox" ; new="xobni" ; filter="AND from:cworth"

I would prefer this to be on separate lines.

>     o1=$(notmuch count tag:"${old}" "${filter}") ; n1=$(notmuch count tag:"${new}" "${filter}")

I would definately prefer this to be on separate lines.

Also, please consider s/o1/old_count_1, s/n1/new_count_1/ and so on.

>     test "${o1}" == "0" && o1="Need more matches!" # prevent false positives
>     test_emacs "(notmuch-search \"tag:${old} ${filter}\")
>             (notmuch-test-wait)
>             (notmuch-search-operate-all \"+${new}\" \"-${old}\")"
>     o2=$(notmuch count tag:"${old}" "${filter}") ; n2=$(notmuch count tag:"${new}" "${filter}")

Same comment about separate lines.

>     notmuch tag -"${new}" +"${old}" -- tag:"${new}" "${filter}" AND NOT tag:"${old}"} # restore db state!

Why "AND NOT tag:$old" is needed here?

Since the line is too long, please move the comment to a separate line
above.

Since we are testing Emacs UI, should we restore db though Emacs as
well?

>     o3=$(notmuch count tag:"${old}" "${filter}") ; n3=$(notmuch count tag:"${new}" "${filter}")

Same comment about separate lines.

>     output="
>     before:   ${old}=$o1 ${new}=$n1
>     after:    ${old}=$o2 ${new}=$n2
>     restored: ${old}=$o3 ${new}=$n3"
>     expected="
>     before:   ${old}=$o1 ${new}=0
>     after:    ${old}=0 ${new}=$o1
>     restored: ${old}=$o1 ${new}=0"

I would change "=" to ":".

Regards,
  Dmitry

>     test_expect_equal "$output" "$expected"
>   #+end_src
> 
> 
> > Regards,
> >   Dmitry
> > 
> > > 
> > > > Regards,
> > > >   Dmitry
> > > > 
> > > > [1] id:"1327901644-15799-1-git-send-email-dmitry.kurochkin at gmail.com"
> > > > 
> > > > > +
> > > > >  test_begin_subtest "Message with .. in Message-Id:"
> > > > >  add_message [id]=123..456 at example '[subject]="Message with .. in Message-Id"'
> > > > >  test_emacs '(notmuch-search "id:\"123..456 at example\"")
> > > > > -- 
> > > > > 1.7.8.1
> > > > > 
> > > > _______________________________________________
> > > > notmuch mailing list
> > > > notmuch at notmuchmail.org
> > > > http://notmuchmail.org/mailman/listinfo/notmuch
> > > 
> > > 
> > > Peace
> > > 
> > > -- 
> > > Pieter
> 
> 
> Peace
> 
> -- 
> Pieter


More information about the notmuch mailing list