Status Meetings:2007-01-17:Log

From Camino Wiki
Jump to navigation Jump to search
[12:00pm] froodian: i thought pink was going to be here
[12:01pm] cl: he's coming 
[12:02pm] pinkerton joined the chat room.
[12:02pm] pinkerton was promoted to operator by chanserv.
[12:02pm] pinkerton: moof
[12:02pm] pinkerton: sorry
[12:02pm] pinkerton: i only have to 9:30
[12:02pm] mento joined the chat room.
[12:02pm] mento was promoted to operator by chanserv.
[12:02pm] cl:    
[12:03pm] cl was promoted to operator by froodian.
[12:03pm] murph was promoted to operator by froodian.
[12:04pm] cl: OK, let's get started.
[12:04pm] cl: Apparently I'm running things this week. 
[12:05pm] cl: And I hadn't looked at the agenda until 30 seconds ago when I posted that link, so...
[12:05pm] pinkerton: good times!
[12:05pm] cl: everyone pull that up 
[12:05pm] cl: 1.1 is coming along
[12:06pm] cl: either we're really good at killing crashers or a2 isn't getting enough testing.
[12:06pm] cl: we're not sure which
[12:06pm] cl: do we have any numbers on a2 downloads, ss?
[12:06pm] cl: (or anyone else?)
[12:07pm] pinkerton: i need to update to it 
[12:07pm] cl: haha
[12:07pm] pinkerton: oh, btw, people at google say a2 is really good
[12:07pm] cl: excellent
[12:07pm] cl: b1 will be even gooder
[12:08pm] cl: i guess it goes without saying that a2 is, in theory, our last release before b1.
[12:08pm] cl: and b1 is string-freeze time
[12:08pm] cl: mento: is early Feb a realistic date for b1?
[12:09pm] mento: as far as spinning it, sure
[12:09pm] cl: OK, good.
[12:09pm] cl: I think we can have all the code where we need it to be.
[12:09pm] mento: but does that leave enough time for development above a2?
[12:09pm] pinkerton: i think our b1 list is short, we need to just get this out the door
[12:09pm] cl: Yeah, it is fairly short.
[12:10pm] mento: no problems on my end, then.
[12:10pm] froodian: speaking of which, we should get to the decisions that need to be made
[12:10pm] froodian: while pink's here
[12:10pm] cl: which brings me to specific bugs, unless anyone has anything to say about other general stuff
[12:10pm] cl: first:
[12:10pm] cl: Make "Keyboard Shortcuts" menu item pointing to online documentation
[12:11pm] thebot joined the chat room.
[12:11pm] pinkerton: on the security stuff, we asked for ss and stuart to be given security bug access
[12:11pm] froodian:                           
[12:11pm] pinkerton: we'll keep folks posted on how that goes
[12:11pm] cl: k
[12:11pm] cl: thx
[12:11pm] cl: Do we want to bother with a keyboard shortcuts menu item?
[12:11pm] cl: I personally think it's a good idea.
[12:12pm] cl: But I know menu bloat is something we've been concerned with in th epast
[12:12pm] froodian: I do too.  The question is whether adding a menu item is bloat.  Personally, we got rid of one, so I think adding one is fair
[12:12pm] froodian: wash out
[12:12pm] pinkerton: heh
[12:12pm] pinkerton: i'm ok i guess
[12:12pm] cl: froodian: make it so.
[12:13pm] froodian: an enthusiastic approval if i ever heard one 
[12:13pm] cl: haha
[12:13pm] cl: pinkerton|emo 
[12:13pm] cl:        bug 343299    
[12:13pm] murph: i'm cool with the keys menu item.  I'm amazed how many people don't use the focus shortcuts for the url/search fields.
[12:13pm] thebot: cl: Bug nor, P3, Camino1.1,, ASSI, Spell-check context menu missing Ignore and Learn Spelling entries
[12:14pm] cl: That's gonna be complicated.
[12:14pm] murph: pink, I've been meaning to ask you about mozOSXSpell.  That's fundamentally the engine underneath the mozilla API's we're using for spell checking, correct?
[12:14pm] cl: yeah, you guys discuss this, since i have no idea what's going on
[12:14pm] cl: ping me when we're ready to move to the next bug 
[12:14pm] pinkerton: mozosxspell is a replacement for the myspell stuff using the cocoa libs
[12:14pm] pinkerton: NSSPellCheck, etc
[12:15pm] pinkerton: just reimplements the same api with the osx stuff instead of shipping a library
[12:15pm] pinkerton: er, shipping a dictionary
[12:15pm] murph: ok, so what ends up happening when a word is learned?
[12:15pm] pinkerton: right now, probably nothing
[12:15pm] murph: because NSSpellChecker doesn't provide a public API for doing that
[12:15pm] murph: i see..
[12:16pm] pinkerton: we might be able to use a 'learned word' dict on the side like the myspell stuff
[12:16pm] pinkerton: and then we'd have to look in both on every word
[12:16pm] froodian: but we'd want learned words to persist across other apps
[12:16pm] froodian: wouldn't we?
[12:17pm] pinkerton: i guess that's the only difference between learn and ignore ,right?
[12:17pm] murph: I was testing the learning behavior, and they do persist even after quitting Camino, so they must be filed in some dict
[12:17pm] pinkerton: learn ignores in other apps 
[12:17pm] pinkerton: ok, maybe we are doing that then
[12:17pm] murph: ignoring does not persist between editing sessions, but learning does (right now)
[12:17pm] pinkerton: oh
[12:17pm] pinkerton: so the only problem we have is that "learn" doesn't learn for other applications?
[12:17pm] murph: the main learning trouble is that it's not propagated to other apps.
[12:18pm] murph: yeah
[12:18pm] froodian: learn doesn't learn for us either
[12:18pm] pinkerton: murph just said it did
[12:18pm] froodian: i mean, it's never presented in the spelling options
[12:18pm] murph: hmm, I thought it did
[12:18pm] pinkerton: oh right
[12:18pm] froodian: if it really learned it, it could be a correct spelling if you spelled something similar to it
[12:19pm] murph: ohh, right, I overlooked that part of it
[12:19pm] pinkerton: so i guess we need to work that into the suggestions as well
[12:19pm] pinkerton: if we can get "learn" to learn and suggest for just camino, i'm ok with that for 1.1
[12:19pm] pinkerton: since FF can't do other apps either
[12:19pm] pinkerton: and we can work on it for the future
[12:19pm] cl: sounds good to me
[12:20pm] pinkerton: if we can't get learn, we can fall back to only having ignore, which is annoying, but not the end of the world at all
[12:20pm] cl: i was just about to ask that
[12:20pm] cl: move on to next bug?
[12:20pm] pinkerton: it's still hugely better than not having spellcheck, so we shouldn't hold the world for it
[12:20pm] murph: yeah, I agree.
[12:20pm] pinkerton: ok
[12:20pm] pinkerton: next?
[12:20pm] cl:        bug 361463    
[12:20pm] thebot: cl: Bug nor, --, Camino1.1,, NEW, Address security concerns with keychain password auto-fill in forms
[12:21pm] pinkerton: i got 10mins
[12:21pm] pinkerton: stuart and i talked a little about this yesterday, but i don't remember what we were talking about 
[12:21pm] ardissone|away: there's no smorgan to desc this
[12:21pm] You are now known as ardissone|food.
[12:21pm] pinkerton: and he's not here now
[12:22pm] cl: well, that pretty much covers all the important stuff
[12:22pm] froodian: well, those are all the 1.1 bugs (yay, we got through before 12:30)
[12:22pm] pinkerton: heh
[12:23pm] cl: pinkerton: if you get some time, I'd like for you to at least think about bug 350733 and give a yes/no/maybe-in-certain-circumstances
[12:23pm] thebot: cl: Bug enh, --, ---,, UNCO, Make cmd/shift modifiers work on items in the back/forward button menus
[12:23pm] cl: doesn't have to be today, just sometime 
[12:23pm] cl: because I get all twitchy when there are UNCOs in my queries
[12:23pm] pinkerton: heh
[12:23pm] cl: especially if they're older than a month 
[12:24pm] ardissone|food: that's a compelling reason to fix a bug 
[12:24pm] cl: it's a compelling reason to make some sort of decision 
[12:24pm] pinkerton: i guess it can't really hurt to do it from a user perspective, just makes that code more confusing
[12:24pm] pinkerton: more features to screw up and regress 
[12:24pm] cl: well, it screws with UI standards, too
[12:24pm] pinkerton: hrm, i guess it does
[12:24pm] cl: since you can't hold down the Command key and click, or the button will move
[12:25pm] cl: and by introducing this inconsistency, there's potential for introducing user confusion
[12:25pm] pinkerton: my gut is to say no, let's keep it simple
[12:25pm] cl: mento: i think you're the point man on bug 337261
[12:25pm] thebot: cl: Bug nor, --, Camino2.0,, NEW, Review dom_storage build changes on the Camino/1.8 branch
[12:25pm] mento: oh, ok
[12:26pm] mento: thanks for the ping, i'll take a look.
[12:26pm] cl: bug 361157 needs some thought on the part of somebody
[12:26pm] froodian: the more i think about it, the more i think it'd be weird
[12:26pm] thebot: cl: Bug nor, --, ---,, ASSI, Opening files should open them in a new window
[12:26pm] froodian: though i filed it
[12:26pm] cl: dunno who 
[12:26pm] froodian: (it = the cmd-click bug)
[12:26pm] pinkerton: ok, i gotta jet if nobody needs me
[12:27pm] cl: thanks
[12:27pm] froodian: later pinkerton 
[12:27pm] pinkerton: sorry i've been so MIA lately, keep up the good work, people do like a2 a lot!
[12:27pm] pinkerton: l8r
[12:27pm] pinkerton left the chat room.
[12:27pm] murph: have a good one pink.
[12:27pm] cl: anyone have anything else to discuss?
[12:27pm] murph: not really, but, FWIW, I think I can get behavior in pretty soon to switch spell checking languages.
[12:27pm] cl: excellent
[12:28pm] ardissone|food: mento: pawn vanished again on Sunday 
[12:28pm] froodian: kicks pawn, just for fun
[12:28pm] mento: box is not up
[12:28pm] ardissone|food: lovely
[12:29pm] ardissone|food: #flog 
[12:29pm] mento: pings #build
[12:30pm] ardissone|food: so, everyone saw the notes about b1 in the agenda?
[12:31pm] ardissone|food: it's string/nib freeze for l10n, so if anyone has bugs that might change those
[12:31pm] ardissone|food: and aren't b1+ or b1?, please b1? them
[12:31pm] murph: ok.
[12:31pm] cl: ardissone|food: you, me, froodian, and ss should go through the target-1.1 list and make sure everything is kosher re: b1
[12:31pm] ardissone|food: i hope we have them all
[12:32pm] ardissone|food: also
[12:32pm] ardissone|food: no new features post-b1
[12:32pm] ardissone|food: we have enough bugs/holes in features we added this cycle that we won't fix them all already
[12:32pm] ardissone|food: and
[12:33pm] cl: heh
[12:33pm] ardissone|food: be very stingy about giving bugs a 1.1 target from now on
[12:33pm] ardissone|food: we re-triaged earlier in the week and kicked 20 bugs
[12:34pm] ardissone|food: but we're still 35 bugs from done (5 are website)
[12:34pm] cl: 34
[12:34pm] ardissone|food: unless it's critical or a new regression, don't 1.1 it
[12:34pm] smorgan joined the chat room.
[12:35pm] smorgan was promoted to operator by you.
[12:35pm] froodian: I've got to head.  See you later everybody
[12:35pm] murph: later froodian 
[12:35pm] ardissone|food: later froodian 
[12:36pm] ardissone|food: smorgan: can you describe the password stuff you and pink were talking about, so we can be clued in?
[12:36pm] smorgan: The which now?
[12:37pm] smorgan: Sorry I was late, something came up this morning
[12:37pm] cl:        bug 361463    
[12:37pm] thebot: cl: Bug nor, --, Camino1.1,, NEW, Address security concerns with keychain password auto-fill in forms
[12:37pm] ardissone|food: pink couldn't remember what you two discussed 
[12:38pm] smorgan: So the UE hit of not prompting is pretty obvious; you have to remember your login info
[12:38pm] froodian left the chat room. (Ping timeout)
[12:38pm] smorgan: (or get it directly from keychain, but my guess is most users don't know how to do that)
[12:38pm] ardissone|food: "not prompting"?
[12:39pm] smorgan: If the action on a form is different
[12:39pm] ardissone|food: oh, ok
[12:39pm] smorgan: So the case there is that changes to do logins through, so now my password doesn't fill
[12:40pm] smorgan: If we do prompt, I think yes has to be sticky, but not no
[12:40pm] smorgan: Reasoning:
[12:40pm] smorgan: Yes, because there's no reason not to. If it's a dangerous form, they are already hosed, and if not it's what they would want
[12:41pm] smorgan: No not sticky for two reasons
[12:41pm] smorgan: 1) no is the default, and you don't want the default reaction to be desctructive
[12:42pm] smorgan: 2) if you don't know if it's a real login form, you need to be able to say no with the confidence that you can reload and press yes if it turns out to be one
[12:42pm] smorgan: So the issue is this:
[12:42pm] ardissone|food: by sticky, you mean change the keychain entry as well?
[12:42pm] smorgan: Right, meaning they aren't prompted again
[12:43pm] smorgan: So the issue is, what if, for whatever reason, there is a form on that you do want to fill, but a login form that you don't
[12:43pm] smorgan: Every time you go to a page with the latter, it will ask you if you want to fill it
[12:44pm] ardissone|food: how common is that?
[12:44pm] smorgan: That's the question
[12:44pm] ardissone|food: that almost sounds like bug ....
[12:44pm] smorgan: A bug where?
[12:44pm] ardissone|food:        bug 178607    
[12:44pm] thebot: ardissone|food: Bug nor, --, Camino1.2,, NEW, Keychain cannot save multiple passwords for the same host, different urls
[12:44pm] ardissone|food: oh, no, not quite
[12:45pm] smorgan: No
[12:45pm] ardissone|food: i mean, the different urls could have different form targets
[12:45pm] ardissone|food: since they're "different forms"
[12:45pm] smorgan: There could be a "never" option, but then we need UI for controlling that, which would get really ugly (and users wouldn't understand it)
[12:47pm] smorgan: My feeling is that we should try the prompt and see what happens, unless a bettter idea surfaces
[12:47pm] ardissone|food: yeah
[12:47pm] smorgan: But it makes me nervous
[12:50pm] ardissone|food: (you can never have two different forms on the same host both get filled, or will we store two entries?)
[12:52pm] smorgan: I guess that's a question too
[12:52pm] ardissone|food: there's a half-sketchy option 3
[12:53pm] smorgan: I was thinking of just adding another action host to the keychain entry
[12:53pm] ardissone|food: tell them they don't match
[12:53pm] smorgan: We could store a different account per action host
[12:53pm] ardissone|food: ah
[12:53pm] smorgan: But that makes doing the right thing on a site re-org harder
[12:54pm] ardissone|food: wishes this was just written out as a website bug.....
[12:55pm] smorgan: Yeah, I mostly feel that way too
[12:56pm] smorgan: But then the JS stuff is sketchy too
[12:56pm] smorgan: I'm really starting to think that the description of Opera's (? I think?) method is actually the right approach
[12:57pm] ardissone|food: what's that?
[12:58pm] smorgan: I guess there's some sort of magic-wand toolbar button that lights up when you are on a page with stored login info, and you press it to log in.
[12:58pm] ardissone|food: hmm
[12:58pm] ardissone|food: unless sfri and fx both switched to doing that, we'd get killed
[12:59pm] smorgan: Maybe, maybe not
[12:59pm] smorgan: It's the same amount of work for the user
[12:59pm] smorgan: And it closes the "unclosable" auto-JS-stealing holes too
[12:59pm] ardissone|food: right now it's no work for the user; we just fill
[1:00pm] smorgan: They have to press the login button
[1:00pm] smorgan: My understanding of the wand is that it does that too
[1:00pm] ardissone|food: oh
[1:01pm] smorgan: It would make the multi-account UI much easier to implement too
[1:01pm] ardissone|food: menu-button?
[1:02pm] smorgan: That was my thought
[1:02pm] smorgan: (not sure if they do it)
[1:04pm] smorgan: It may well have problems I haven't thought of, but it does close a lot of security holes by requiring initial user action
[1:05pm] ardissone|food: ok, so homework for everyone is to check out Opera (we think) and for smorgan to talk to pink more
[1:05pm] smorgan: Oh, this probably isn't 1.1 material
[1:06pm] cl: so is there a viable solution for 1.0.4?
[1:06pm] ardissone|food: really?
[1:06pm] smorgan: We'd need more time to consider/test a big UI change like that I think
[1:06pm] cl: because that's the other thing that's being held up.
[1:07pm] smorgan: I think for 1.0.4 and 1.1 we pick an option like prompting on different forms or silently not filling
[1:07pm] ardissone|food: oh, the opera thing, not "fixing this bug somehow"
[1:07pm] smorgan: Right
[1:07pm] ardissone|food: ok
[1:07pm] ardissone|food: whew
[1:08pm] ardissone|food: or we could not fill and tell them why 
[1:08pm] ardissone|food: instead of being silent 
[1:08pm] smorgan: How is that better than prompting and having a fill button?
[1:09pm] ardissone|food: but i still think i prefer the prompt
[1:09pm] ardissone|food: heh, i need caffiene
[1:10pm] ardissone|food: one other question for you before we leave
[1:10pm] ardissone|food: did Eiichi's log mean anything?
[1:10pm] ardissone|food:    
[1:10pm] thebot: ardissone|food: Bug 337750 cri, P2, Camino1.1,, NEW, bookmarks-corrupted-n.plist often written as empty file
[1:10pm] ardissone|food: he's on 10.3.9
[1:11pm] ardissone|food: or was, last i knew
[1:11pm] smorgan: He couldn't have been
[1:12pm] smorgan: And that error makes no sense at all to me.  "I couldn't write a file because it doesn't exist"
[1:12pm] smorgan: I have no idea what that means actually happened
[1:13pm] smorgan: I'm hoping someone else will get something else that makes sense
[1:13pm] ardissone|food:                           
[1:13pm] smorgan: But I have no real hope for that bug making sense anytime soon
[1:14pm] smorgan: I think your backup idea is the only reason we'll be able to ship 1.1
[1:14pm] ardissone|food: heh
[1:14pm] ardissone|food: now if i could only come up with an idea like that for the "page loads but doesn't render" bug
[1:16pm] ardissone|food: anyway, it's a quarter after 1 here, so unless anyone has anything else...
[1:16pm] smorgan: If I could repro it on demand I could see if it's perhaps another case like the scrollbar...
[1:17pm] froodian joined the chat room.
[1:17pm] froodian was promoted to operator by chanserv.
[1:18pm] ardissone|food: most common is a full-page image that's scaled
[1:18pm] ardissone|food: and then random pages thereafter
[1:18pm] ardissone|food: 0r facebook 
[1:18pm] smorgan: That one wasn't fixed already?
[1:19pm] smorgan: I thought the more reproducible version had been fixed somehow
[1:19pm] ardissone|food: the image-dragging from desktop, no
[1:19pm] ardissone|food: that's why the bug is assigned to mento, though
[1:19pm] ardissone|food: we need to split that bug 
[1:29pm] You are now known as ardissone|away.