IRC Log Viewer » #firebreath » 2011-06-01

IRC Nick Time (GMT-7) Message
dougma 01:06 firebreath-89 commit crashing on examples\FBTestPlugin\test.html :/
but i think it has more problems than my comment on github
sorry, can't offer any more insights atm
dougma 01:06 taxilian_away: ^^ :)
taxilian 02:06 FIREBREATH-89
FireBreathBot 02:06 FIREBREATH-89: Summary: WinMessageWindow should allow a custom winproc
FIREBREATH-89: Assigned To: richard
FIREBREATH-89: Priority: Major, Status: Open,
taxilian 02:06 Hmm. I will look at it tomorrow
_pq_ 02:06 doesn't it already kinda do?
_pq_ 04:06 does anyone has got problems catching right mouse click on ff4 on macos?
i mean middle mouse
middle mouse events are reported as left mouse events
_pq_ 04:06 safari has similar issue
right mouse is completely ignored
is that expected/known bug/my fault?
taxilian: i got timer to work as expected
no issue with that yet
FireBreathBot 08:06 JIRA issue issue created by geoffc
FireBreathBot 10:06 Commit fafa665 on master by Richard Bateman: "FIREBREATH-89 fixed race condition w/ window creation"
JIRA issue issue resolved by richard "Thanks for the heads up; sorry about that; it's a race condition, so I wasn't hitting it on the p..."
taxilian 10:06 dougma: that should fix the messagewin issue
FireBreathBot 11:06 JIRA issue issue commented by geoffc "No problem Richard, happy to help out. I know that the dev branch is nearly always stable, so I l..."
JIRA issue issue commented by richard "If you get it from git the only way is to check the commit you are on. If you download it from gi..."
JIRA issue issue commented by geoffc "I get it from github, but the version file no longer has the version number at the top of the fil..."
JIRA issue issue commented by richard "That indicates that you got it from git directly; if that's the case you'll have to look at the g..."
FireBreathBot 11:06 JIRA issue issue created by geoffc
FireBreathBot 14:06 Commit c50b114 on master by Richard Bateman: "FIREBREATH-91: fixed deadlock on WinMessageWindow shutdown"
JIRA issue issue resolved by richard "Ever noticed how windows uses reentrant code all over the place? =]
JIRA issue issue commented by geoffc "Thanks Richard!"
sabotaged|wk 14:06 so just to verify, for a clean plugin installation (i.e. no plugin with the same name or mimetype exists already), no browser restart is required on any platform, right? though you may need to call navigator.plugins.refresh
taxilian 14:06 sabotaged|wk: I fixed the _tmp/ issue
if you isntall using the MSI/registry method, that is correct
sabotaged|wk 14:06 oh really? what was it
taxilian 14:06 weird handling in the way that I was using configure_file
sabotaged|wk 14:06 ok cool
taxilian 14:06 instead I use cmake -E copy_if_different now
and it seems to work fine
at least, for me
see if you can confirm that =]
when conveninet
sabotaged|wk 14:06 k
hmm, why do you think flash player requires a browser restart (firefox in my case) when doing an install, even when flash doesn't exist already
taxilian 15:06 it's easier
also depends on how you install
sabotaged|wk 15:06 you mentioned MSI, what's so special about that and "registry" method?
taxilian 15:06 well, the MSI uses the registry install method
it's really the registry method that is important, though
sabotaged|wk 15:06 as opposed to what, running regsvr?
taxilian 15:06 if you register the plugin in the registry (which regsvr also does) then you don't have to restart for it to be detected
if you install it as an XPI then you do
sabotaged|wk 15:06 oh
taxilian 17:06 I think I need to combine PluginCore and ScriptingCore into a single project
they were originally seperate-ish, but the more time passes the less that is true
dougma 19:06 was it the _tmp issue which made visual studio want to reload a newly prep'd .sln?
cause that's fixed now which is nice. :)
taxilian 19:06 yes
that was caused by the same issue =]
my track record recently has been somewhat below par on fixes :-/ let's hope I do better on this proxy stuff
dougma 19:06 hehe
taxilian 19:06 bbl, afk
dougma 19:06 everyone jumping on the master and testing stuff. :)
FireBreathBot 19:06 JIRA issue issue commented by dougma "So another way to make it crash is run switcher.html and click away in the control while it's aut..."
taxilian 20:06 dougma: so your pull request includes a full fix for FIREBREATH-87 or just a partial?
FireBreathBot 20:06 FIREBREATH-87: Summary: ActiveX crash when creating out of DOM
FIREBREATH-87: Assigned To: richard
FIREBREATH-87: Priority: Major, Status: Open,
dougma 20:06 full fix
taxilian 20:06 git pull requests
FireBreathBot 20:06 TypeError: object of type 'bool' has no len() (file "/home/richard/phenny/", line 186, in call)
taxilian 20:06 lol. wow
that's pretty funny
dougma 20:06 true though :)
taxilian 20:06 lol. yes, also true =]
dougma 20:06 nice to see a bot with character
taxilian 20:06 well, I wrote the github integration for this one
so that gives it some of it's character… like the part we just saw =]
git pull requests
FireBreathBot 20:06 TypeError: object of type 'bool' has no len() (file "/home/richard/phenny/", line 186, in call)
taxilian 20:06 git pull requests
FireBreathBot 20:06 4 open pull requests:
dougma: add 64bit build support for windows
dougma: FIREBREATH-87: do just the necessary and avoid crash
dougma: FIREBREATH-77: ActiveX: focus+capture on mousedown
dougma: FIREBREATH-68 set this for ax & npapi callbacks
taxilian 20:06 that's better =]
dougma 20:06 cool... maybe i should close fb-68 request? still not sure what to do there and it's work-aroundable on the js side
taxilian 20:06 what do you see as being the "ideal"? if we could make it work exactly as you think it should work?
FireBreathBot 20:06 FIREBREATH-86: Summary: crash with FBTestPlugin on switcher.html
FIREBREATH-86: Assigned To: richard
FIREBREATH-86: Priority: Major, Status: Resolved,
dougma 20:06 I think it should set 'this' to the plugin's element (if it's in the dom!) when firing events because other events work that way
taxilian 20:06 even if it isn't fired from the root JSAPI object?
say it is fired from another object?
FireBreathBot 20:06 FIREBREATH-87: Summary: ActiveX crash when creating out of DOM
FIREBREATH-87: Assigned To: richard
FIREBREATH-87: Priority: Major, Status: Open,
dougma 20:06 in that case this would be the other object.... :)
FireBreathBot 20:06 Commit 603e9cd on master by dougma: "FIREBREATH-87: do just the necessary and avoid crash"
taxilian 20:06 that's where it gets tricky
if you want it to be the DOM object, fine… if you want it to be the JSAPI object, fine… but making it so it could be one or the other depending on circumstances is tricky
dougma 20:06 mmm
taxilian 20:06 would it be enough to just make it so events always fire using the JSAPI object that it originated from?
the (Subtle) difference would be that you wouldn't be able to get id, size, position in dom, etc
dougma 20:06 and i think that will tripup things like jquery
taxilian 20:06 jquery doesn't play nice with FireBreath anyway
can't figure out exactly why
dougma 20:06 (although jquery seems to have other problems)
taxilian 20:06 right
dougma 20:06 yeah it has a block list and won't attach events to object (unless it's flash!)
i stepped through it one day
1.6 anyways
silently refuses for extra giggles
taxilian 20:06 of course, if it did, it probably wouldn't handle things correctly on IE9 anyway
with the lovely "addEventListener fails silently" thing
dougma 21:06 well i just moved on with my own event attachment... :)
anyway getting some context on an event would be better than nothing.
taxilian 21:06 right
if there were some way that I could check to see if the JSAPI object were the root, that would work; that would require refactoring some other things, though
of course, I'll probably need to do that whenever I finally get around to doing FIREBREATH-5
FireBreathBot 21:06 FIREBREATH-5: Summary: Reserved JSAPIAuto names (dom element)
FIREBREATH-5: Assigned To: richard
FIREBREATH-5: Priority: Minor, Status: Open,
JIRA issue issue resolved by richard "PCH working in the 1.6 (master) branch!"
dougma 21:06 yikes... fb-5 is interesting!
FireBreathBot 21:06 JIRA issue issue resolved by richard "Fixed; use boost::bind to assign a custom winproc to a messagewindow"
dougma 21:06 this is where IE's design is better. :)
taxilian 21:06 isn't it, though? =]
except that IE, not allowing us to override things, creates issues like the "addEventListener" issue
advantages both ways, really
dougma 21:06 but then you fix that on the js side right?
taxilian 21:06 right; however, anyone who writes their event handling code incorrectly will check for addEventListener, find that it is there, and use it
it will not throw any exceptions, in fact it will act in all ways as though it worked
it just won't work
dougma 21:06 oh...
i guess i lucked out and wrote it right :)
taxilian 21:06 so for anything else in IE9 you can use addEventListener, but not for activex controls
heh. well, if you checked for attachEvent first and use that if it is there then you're good
dougma 21:06 so IE9 is saying addEventListener exists but doesn't involve FB in that conversation when the user calls it?
taxilian 21:06 right
doesn't call any of the connectionpoints apis, nothing
unless there is a new IE9-specific interface that I don't know about that I should be providing
dougma 21:06 maybe there's something to query on IHTMLElement?
taxilian 21:06 hmm. interesting possibility
I hadn't thought of that is the new IE9 IHTMLElement
interesting. there are definitely some events that you can get that way
but it doesn't allow arbitrary event names
dougma 21:06 are you interested in adding support for raising real dom events?
like click
taxilian 21:06 if it can be done in a reasonable way, yes
I'm pretty much interested in anything that makes plugins behave more as a javascript developer would expect
as long as it doesn't require compromising stability or performance in meaningful ways
dougma 21:06 sure.
taxilian 21:06 but I think you figured that part out on your own =]
dougma 21:06 which reminds me...
what is FB's browser support? as in versions.
sorry... afk a sec
taxilian 21:06 FF 3 and later, Chrome 2 and later (I think), Safari 3 and later, Opera… sometimes.
IE 6 and later
that's more or less the list
dougma 21:06 ta. just on fb-87: the other two commits in the pull request fix crashes too.
taxilian 21:06 I already pulled those in
they were combined into a single commit is all
those were the fixes for FIREBREATH-86, no?
FireBreathBot 21:06 FIREBREATH-86: Summary: crash with FBTestPlugin on switcher.html
FIREBREATH-86: Assigned To: richard
FIREBREATH-86: Priority: Major, Status: Resolved,
dougma 21:06 fb-87 was more of the same sort of thing (document being null) but i found them after you'd applied the fb-86 patches
(window being null)?
something was null!
taxilian 21:06 yeah
can you look and verify? I thought the other two commits had already been applied
I need to find a good stopping point on this proxy detection code before I can easy check
dougma 21:06 looking
no, the other two ( and ) haven't been applied
FireBreathBot 22:06 c806716 by dougma: fix more DOM Window null problems
dougma 22:06 oh... weird.
i will test master now. :)
taxilian 22:06 believe it or not, it actually finds that even if it's not in the main repo
anywhere on github
!findfile BrowserHost.cpp
FireBreathBot 22:06 Found 3 matching file(s) in the master branch. First 3 are:
taxilian 22:06 tell you what… if I give you commit access, would you like to put those in yourself? I would appreciate it if you'd edit hte commit messages so that the JIRA ticket is on each one
dougma 22:06 ok, i can try that :)
taxilian 22:06 I assume you already understand the whole "please don't destroy the repository, be careful what you commit, don't commit things to the wrong branch, etc" speech, right? =]
I have given you access to the repository. please be careful =]