Closed Bug 170561 Opened 22 years ago Closed 22 years ago

QA doc/outline for type ahead find

Categories

(Developer Documentation Graveyard :: General, defect, P3)

defect

Tracking

(Not tracked)

VERIFIED FIXED

People

(Reporter: bugzilla, Assigned: owen.marshall+bmo)

References

()

Details

Attachments

(3 files, 10 obsolete files)

summary sez it all: should have a test outline for type ahead find.

ui/dev info at http://www.mozilla.org/projects/ui/accessibility/typeaheadfind.html
oops. over to moi...
Assignee: endico → sairuh
QA Contact: imajes → paw
Status: NEW → ASSIGNED
Priority: -- → P3
*** Bug 169954 has been marked as a duplicate of this bug. ***
Blocks: isearch
hi Owen --sorry to respond here in the bug, but my email reply to you bounced. :-/

Owen Marshall wrote:

> Hey, I have some time, and looking at the Feature Test Plans, I think I can
whip up one for typeahead find. Would you object to this? I will start working
on this, and if you already have a copy, if you would rather write this
yourself, etc, just give the word and I will stop. Otherwise, I will post this
up on Bugzilla soon.

my response:

that would be awesome!

what would think about attaching your page to this bug? we could go over it from
there, and i could check it in to the website.

i already have a few keyboard nav QA docs posted under the following directory,
and the one for typeahead find would prolly reside there, i'd think. fwiw, it'll
give you an idea as to format i use for such docs.

http://www.mozilla.org/quality/browser/front-end/testcases/keyboard-nav/

what do you think?
Hey, it looks good. I am making this as compliant as possible -- when I am done,
you should be able to drop it straight into the test outline. I can get a rough
draft up by tomorrow night, with luck :)

One question, though. I was going to include some info in there, such as QA
contact, development contact, etc. I would assume that would be you and aaronl,
correct?

Also, will a new component be created for this? Or will it fall under the realm
of Keyboard Navigation?

Thanks =)
yep, aaronl and i are the default contacts. and yes, Keyboard Navigation would
be proper bugzilla component for this feature.

thanks a lot, Owen!
Ok, hopefully the last question =) Currently, Type Ahead only applies to the
navigator portion, right? Any plans to make it work in composer or address book?

Thanks!
Definitely not addressbook, but typeaheadfind could potentially be applied to
anyhting that uses a Gecko content window, including composer, mailnews, IM ...

Since these things use letters for editing or for shortcuts (like N for next
unread message), they would need accelerators to initiate type ahead find.

I have a patch in bug 158752 to make it work in composer, but I'm not sure what
our plans are for that yet.
just my own $0.02: for an initial version of the test doc, i'd limit details to
the browser (navigator content). sanity checks in other areas and apps would be
okay --but with all things QA (and development), the outline could grow
endlessly. ;)
Attached file First revision (zipped). (obsolete) —
Comment on attachment 100719 [details]
First revision (zipped).

This is the first revision. I will be going over it tonight. Any
comments/concerns?
Attachment #100719 - Attachment description: First revision (zipped). I will be going over it tonight. Any comments/concerns? → First revision (zipped).
Sorry for that spam -- still trying to figure out attachments on Bugzilla ;)

New location: www.bardstowncable.net/~malachi/typeaheadfind.html
i'll take a look at these today.

Owen, would it be okay if i made edits inline to what you've done, then attach
'em here?
This is going to be really important soon.

I'm working on a fix for bug 168281 (support IME for typeaheadfind), which is
making me break typeaheadfind up into smaller, more sensible methods internally.
I think we're going to have to be careful about regressions for that one.

Naoki, do you have a list of i18n tests for typeaheadfind you can contribute?
Oops, I should have said "Kyle and I" are working on bug 168281, I couldn't do
that without him.

Also, thanks to Owen Marshall for getting this started!

I think we want to make sure to test repeated character find, find next,
backspace, escape to cancel, and status bar messages.

>Owen, would it be okay if i made edits inline to what you've done, then attach
'em here?

Sure, go right ahead!

>I think we want to make sure to test repeated character find, find next,
backspace, escape to cancel, and status bar messages.

Good point. I will almost definately add this in tonight and get it ready. I 
will tell you later today when the updated versions are ready.
Attached file typeaheadfind.html (obsolete) —
Attached file taf_acceptance.html (obsolete) —
i renamed the acceptance test doc to make it more unique within the keybd nav
doc directory.
Attached file taf_functional.html (obsolete) —
i renamed the functional test doc to make it more unique within the keybd nav
doc directory.

i've also replaced the images in here with text, if that's okay.
Comment on attachment 100719 [details]
First revision (zipped).

see revisions in the three attach html files.
Attachment #100719 - Attachment is obsolete: true
so here's what i did:

* ran the spellchecker ;)
* changed each "step" to a test case with numbered steps --i feel it'll easier
to read that way

suggestions:

* what html editor are you using? fwiw, i used composer (today's trunk build on
linux). it seemed to add some funky styling code. i cleaned it up (just a
little). no biggie, but it might make it easier if you end up editing my
attachments. :)

* in addition to aaron's suggestions, how about adding tests for using ' when
typeahead find is set to search all text? (definitely a functional/advanced
test, not basic acceptance.)

thanks again!
* ran the spellchecker ;)
  hehehe, and I am sure I would be ashamed at what results the spellcheck gave
you =)

* changed each "step" to a test case with numbered steps --i feel it'll easier
to read that way
  Yes, much easier to read. Thanks!

* what html editor are you using? fwiw, i used composer (today's trunk build on
linux). it seemed to add some funky styling code. i cleaned it up (just a
little). no biggie, but it might make it easier if you end up editing my
attachments. :)
   <span style="color: rgb(255, 0, 0);">
   You mean that? Yeah, I am using composer (build 2002090804 on WinME). Same stuff.
 
 * in addition to aaron's suggestions, how about adding tests for using ' when
typeahead find is set to search all text? (definitely a functional/advanced
test, not basic acceptance.)
   On it as we speak =)
><span style="color: rgb(255, 0, 0);">
>   You mean that? Yeah, I am using composer (build 2002090804 on WinME). Same
stuff.

har! yet more unclean code generated by composer *sigh*...
(sorry, not blaming you, o'course. just complaining. okay, i'll stop with the
bugspam now. ;)
Oops, ' doesn't search for all text
' searches for links only
/ searches for all text

Some minor points about repeated character search in typeaheadfind:
* After a repeated character search (such as "aaa" to search links that start
with "a"), a different character will start a new repeated character search with
the new character.
* In repeated character search, Backspace/Shift+F3/Shift+Accel+G go to the
previous match, F3/Accel+G go to the next search.
I should clarify that what I just said about "repeated character find" will
apply once bug 168281 is fixed.

More comments -- hope these are useful!

Here are some points about the order typeaheadfind prefers matches.
* If nothing is focused or selected, or the doc is focused, typeaheadfind will
prefer 1) a visible match, then 2) a match below what is currently visible, and
finally 3) a match above what is currently visible.
* If something is currently focused or selected, then typeaheadfind will move
forward from the current focus or selection and then wrap around to the
beginning of the document
* For normal typeaheadfind (not repeated character find), backspace should hit
the exact same matches that typeaheadfind found in the forward direction

Here are some other fine points:
* Space can be part of a match only if it's not the first character. This way it
doesn't conflict with the spacebar page down command.
* I didn't see any functional tests of findnext/find prev. Those commands should
respect the links-only/all-text setting of typeaheadfind, and will wrap around
even when normal find doesn't.
* Typeaheadfind will change the string in the normal find dialog. Next time you
hit Accel+F that will be there. However, it should not affect any of the other
settings.
* Tabbing, changing the selection, scrolling, waiting or pressing Escape will
cancel the current typeaheadfind
* The special green color should only be used for the selection when
typeaheadfind mode is active. If the selection is green at another time (I have
seen this happen), then there is a bug.

Another point about repeated character find:
* After the same character is typed twice, typeaheadfind will still search for
that exact string. Only when it cannot find an exact match for that, and the
same char has been typed multiple times, will it go into repeated char mode,
where it only matches the first character of links. 
* Repeated character find only looks for links, regardless of the textonly setting.

* Also, all the tests I saw matched to the beginning of a word or link.
Typeaheadfind can match to any part of the word or link.
Here are some more tests that I didn't see in the outline:

Mackeral
Medicine
Mendocino
Mental
Mentor

* When you type "Mento" it should end up on "Mentor", after briefly visiting
each one before.
* When you type Menial, it should beep after the i is pressed, and stay on Mental.
Sorry this is so late =/ I have made many changes to the document, and I will 
upload it later tonight.
Hm, I think I uncovered a bug during testing =) I am now uploading the newest 
functional tests. When I attempted test I on Win2K (build 2002092908), it did 
not properly find the word "peninsula". I designed Test I to test this comment:

* If nothing is focused or selected, or the doc is focused, typeaheadfind will
prefer 1) a visible match, then 2) a match below what is currently visible, and
finally 3) a match above what is currently visible.

Now, the way I see Type Ahead Find is this. Type Ahead Find should always 
prefer a 3 character match over a 2 character match, a 2 char over a 1 char, 
etc. When I read that comment, I assumed Type Ahead would wrap back around in 
search for a more exact string, and designed test I to test this portion. But, 
as you see, it dosen't really work like that =)

Shall I change Test I? Or is this something for you to look at?

All other document changes are small. I want to really tidy up the HTML code 
that Composer made before I submit the rest.
Attached file Latest functional tests (obsolete) —
> Now, the way I see Type Ahead Find is this. Type Ahead Find should always 
> prefer a 3 character match over a 2 character match, a 2 char over a 1 char, 
> etc. When I read that comment, I assumed Type Ahead would wrap back around in \
> search for a more exact string, and designed test I to test this portion. But, 
> as you see, it dosen't really work like that =)

I'm not sure I would say it "prefers" the larger match, because it will simply
say "link not found" or "text not found" if the sum of what you have typed isn't
found. I would say it *requires* the sum total of what you typed, otherwise it
will beep, stop and put a not found message on the status bar. It should
definitely wrap around if that's what it has to do, to match the whole new string.
Attached file Acceptance Tests (obsolete) —
*Killed the awful red text =)
*Ran through Tidy
Attached file Functional Tests (obsolete) —
*Killed the awful red text =)
*Ran through Tidy
*Added part of a test
Attachment #101251 - Attachment is obsolete: true
Attached file Type Ahead Find Outline (obsolete) —
*Added tests to the outline
*Ran through Tidy
Latest documentation changes. Tidy was able to clean the HTML up fairly well. 
If you see anything missing/have any problems, just shout =)
thanks, Owen!

so to let you know: i might not get to this till friday, as some urgent projects
landed on my plate. (if i get to it sooner, yay! but i thought i warn y'all.)
Attachment #100947 - Attachment is obsolete: true
Attachment #100949 - Attachment is obsolete: true
Attachment #100950 - Attachment is obsolete: true
revised(*) files coming up. just some minor edits, really. eg, for some tests i
chopped some steps into seperate ones for a bit more reading clarity. otherwise,
quite good!

(*) using composer with y'day's trunk build. hope it doesn't mess up the code
that Tidy cleaned up!
Attachment #101347 - Attachment is obsolete: true
Attachment #101344 - Attachment is obsolete: true
Attached file taf_functional.html - functional tests (obsolete) —
Attachment #101345 - Attachment is obsolete: true
i should've added that these get r=sairuh. however, owen / aaron, what do you
think of these latest attachments?
Mmm, no more red text =)

It dosen't look like Tidy's changes were really messed up. Most of the changes
it made were deleting unused <span> tags, and it dosen't look like there are any.

Also, I was looking at the functional tests, and I noticed that some of the text
had a bit of a dark red color to it. This goes back to the functional tests,
attachment 101345 [details]. I must've clicked something wrong... fix inbound.

Other than this, it looks great to me!
Attached file taf_functional - functional testcases (obsolete) —
Red color is now gone =)
oh, yeah: i will add a NOWRAP file so that the acceptance and functional test
pages don't get the usual mozilla.org banner and nav tab (which would interfere
with the tests).
s/nav tab/navigation table

(i've checked in the NOWRAP file.)
Attachment #101856 - Attachment is obsolete: true
Comment on attachment 101856 [details]
taf_functional.html - functional tests

oops --actually, the latest version (attachment 101897 [details]) didn't include the
edits that this one has. i'll update this.
Attachment #101856 - Attachment is obsolete: false
cleaned up version (no more red or dark red text).
Attachment #101856 - Attachment is obsolete: true
Attachment #101897 - Attachment is obsolete: true
I just checked this in to:

http://www.mozilla.org/projects/ui/accessibility/qa/taf_qa.html

I also linked to it off the main typeaheadfind page.

It will take a couple of hours for the changes to show up on mozilla.org.

Thanks you guys -- this is great!
Status: ASSIGNED → RESOLVED
Closed: 22 years ago
Resolution: --- → FIXED
wrong location, reopening.
Status: RESOLVED → REOPENED
Resolution: FIXED → ---
updated to v1.47: projects/ui/accessibility/typeaheadfind.html

checked in v1.1 of the following:

quality/browser/front-end/testcases/keyboard-nav/typeaheadfind.html
quality/browser/front-end/testcases/keyboard-nav/taf_acceptance.html
quality/browser/front-end/testcases/keyboard-nav/taf_functional.html

and reassigning to Owen, since he did the work. :)
Assignee: sairuh → Malachi
Status: REOPENED → NEW
Files uploaded, it seems that everything is in order. Marking Verified.
Status: RESOLVED → VERIFIED
*bugspam*
Component: Mozilla Developer → Documentation Requests
Product: Documentation → Mozilla Developer Center
Component: Documentation Requests → Documentation
Component: Documentation → General
Product: Mozilla Developer Network → Developer Documentation
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Created:
Updated:
Size: