154483 - I think notregexp should work the way i want it to, and not the way it did.

Status: RESOLVED WORKSFORME

(Bugzilla :: Query/Bug List, enhancement)

Description

[timeless]

I was trying to search for all bugs where neither the assignee nor the QA was 
timeless.

I think not should work the way i want it to, and not the way it did.

i also suspect i'm filing a dupe, but i can't file as unconfirmed, and i can't 
find it.

Comment 1

[Bradley Baetz (:bbaetz)]

Is it possible for you to be any more obscure? :)

Anyway, the real way to fix this is to move 'not' out to be a separate
operation, for all of these things.

Comment 2

[Stephen Lee]

The email checkboxes as they are make more sense if you use something like:
Any of
[X] bug owner
[ ] reporter
[X] QA contact
[ ] CC list member
[ ] commenter
[doesn't match regexp |v]
[ .*@mozilla.org        ]

(or a more complex regexp to get mozilla.org, netscape.com, bugzilla.org, etc.)

i.e. it's asking "are there any bugs which have 'outsiders' involved" (w.r.t. 
this regexp)


I think that the sort of query you were originally asking for would be to allow 
the "any of" static text at the start to replaced with "none of", e.g.

( ) Any of  (O) None of
[X] bug owner
[ ] reporter
[X] QA contact
[ ] CC list member
[ ] commenter
[ contains     |v]
[ timeless       ]


As for the original query in the url, try this instead:

http://bugzilla.mozilla.org/query.cgi?
query_format=&short_desc_type=allwordssubstr&short_desc=&product=Webtools&compon
ent=Bonsai&component=Build&component=Despot&component=LXR&component=Tinderbox&lo
ng_desc_type=allwordssubstr&long_desc=&bug_file_loc_type=allwordssubstr&bug_file
_loc=&status_whiteboard_type=allwordssubstr&status_whiteboard=&keywords_type=all
words&keywords=&emailtype1=substring&email1=&emailtype2=substring&email2=&bugidt
ype=include&bug_id=&votes=&changedin=&chfieldfrom=&chfieldto=Now&chfieldvalue=&r
emaction=run&namedcmd=My%20bugs&field0-0-0=qa_contact&type0-0-
0=notsubstring&value0-0-0=timeless&field0-1-0=cc&type0-1-0=notsubstring&value0-
1-0=timeless

Comment 3

[Joel Peshkin]

Once it becomes possible to negate boolean charts, it becomes possible to negate
the special charts as well.

Comment 4

[Wayne Mery (:wsmwk)]

Joel in comment #3)
> Once it becomes possible to negate boolean charts, it becomes possible to
> negate the special charts as well.

Joel is there a bug# where this is fixed (or not fixed)? 

Comment 5

[timeless]

Attachment: something like this?l

Comment 6

[Vlad Dascalu]

Comment on attachment 271468 [details] [diff] [review]
something like this?l

The entire testing suite reports this as uncompile-able.

Comment 7

[Vlad Dascalu]

Comment on attachment 271468 [details] [diff] [review]
something like this?l

Hmm, ok, I debugged it too:

+            if ($params->param("negate$id") {

You need above two closing paranthesis instead of one:

s#id")#id"))#

Comment 8

[timeless]

Attachment: with balance

Comment 9

[Vlad Dascalu]

Comment on attachment 276309 [details] [diff] [review]
with balance

This doesn't work, I've tested it on trunk and I get the same results, whether I select "Any" or "None".

As far as I could see,             $params->param("negate$chart-$row", 1); is correctly populated, but it's not read at a later date.

Maybe we need "$chart" instead of "$chart-$row" ? (grepping the source seems to suggest that).

Comment 10

[timeless]

justdave: so, i hit a snag...
the current code assumes that all the specials can cooperate as a single chart (chart 1) which is shorthanded {join:-}{index:1}. That doesn't work as soon as you use negation. But I don't want the UI to reflect a second copy of the query into boolean charts (that'd mean it'd be impossible for a user to change the restrictions).

So, I'm looking for a suggestion as to how to store additional charts. Do I just shift all user charts down, or try to make use of slots after user charts.

Comment 11

[Max Kanat-Alexander]

Nowadays you can just use NOT for this.