Yes, so this looks like it needs addressing urgently. I was really
hoping it was just a regression but no such luck. I need to read back
through all these posts but if anyone can summarise the problems
they've encountered so far here that would be a great help.
On Tue, Jun 30, 2009 at 5:19 PM, Zeno<zeno.crive...@gmail.com> wrote:
> Yes, Firefox v3.5 (final) has just been announced and many (Low Pro)
> stuff seems to be broken
> Grrrr
> Maybe we should start a new thread about FF related issues (talking
> about this in the Safari 4 thread is not very appropriate)
> On 29 June, 00:17, pushrod_admin <pushrodme...@googlemail.com> wrote:
>> This prob seemed to go away, but has now come back (for me) with the
>> latestFirefoxRC for Mac. ANyone else seeing this?
>> On 31 May, 11:51, Dan Webb <d...@danwebb.net> wrote:
>> > Yep, that's good news. I expected it to be the case as it looked more
>> > like a regression inFirefoxthan a bug in Low Pro.
>> > Thanks for letting us know, Nathan.
>> > On Sun, May 31, 2009 at 11:48 AM, MRoderick <roderick.mor...@gmail.com> wrote:
>> > > On 31 May, 09:46, Nathan <XGamer...@gmail.com> wrote:
>> > >> Well good news! Downloaded a newfirefox3.5 nightly (Shiretoko/
>> > >> 3.5pre) and this version works perfectly with my sample case and with
>> > >> my full website. They must have fixed the bug. This is excellent
>> > >> excellent news for me and lowpro users everywhere. Crisis averted. I
>> > >> can't wait for 3.5b5 which will hopefully incorporate this change.
>> > > That is great news! Thanks for reporting it :-)
Woah, just from running the test pages it seems that things are pretty
messy. It looks like its something to do with the inheritance code
which is nearly identical to prototypes except for the way the
constructor is created...and that bit is pretty hairy to be fair. So,
I'm going to dig in to this and try to root out the problem. This
should probably fix most if not all of the problems we are
experiencing though.
If anyone else wants to dig in and assist that would be great.
On Tue, Jun 30, 2009 at 5:22 PM, Dan Webb<d...@danwebb.net> wrote:
> Yes, so this looks like it needs addressing urgently. I was really
> hoping it was just a regression but no such luck. I need to read back
> through all these posts but if anyone can summarise the problems
> they've encountered so far here that would be a great help.
> Thanks,
> Dan
> On Tue, Jun 30, 2009 at 5:19 PM, Zeno<zeno.crive...@gmail.com> wrote:
>> Yes, Firefox v3.5 (final) has just been announced and many (Low Pro)
>> stuff seems to be broken
>> Grrrr
>> Maybe we should start a new thread about FF related issues (talking
>> about this in the Safari 4 thread is not very appropriate)
>> On 29 June, 00:17, pushrod_admin <pushrodme...@googlemail.com> wrote:
>>> This prob seemed to go away, but has now come back (for me) with the
>>> latestFirefoxRC for Mac. ANyone else seeing this?
>>> On 31 May, 11:51, Dan Webb <d...@danwebb.net> wrote:
>>> > Yep, that's good news. I expected it to be the case as it looked more
>>> > like a regression inFirefoxthan a bug in Low Pro.
>>> > Thanks for letting us know, Nathan.
>>> > On Sun, May 31, 2009 at 11:48 AM, MRoderick <roderick.mor...@gmail.com> wrote:
>>> > > On 31 May, 09:46, Nathan <XGamer...@gmail.com> wrote:
>>> > >> Well good news! Downloaded a newfirefox3.5 nightly (Shiretoko/
>>> > >> 3.5pre) and this version works perfectly with my sample case and with
>>> > >> my full website. They must have fixed the bug. This is excellent
>>> > >> excellent news for me and lowpro users everywhere. Crisis averted. I
>>> > >> can't wait for 3.5b5 which will hopefully incorporate this change.
>>> > > That is great news! Thanks for reporting it :-)
I've fixed the problems with Lowpro 0.5 in FF3.5 by changing the
second definition of "var behavior" to "var behaving" -- it seems
FF3.5 did not like the duplicate reference within the function block
for the initial "var behavior".
I've tested this so far in FF3.5 on XP and Leopard, FF3.0.11 on
Leopard, Safari 4 on Leopard, Safari 3.2 on XP, Chrome, and our good
friends Internet Explorer 6 and 7. Worksforme, testing with an
interface chock-full of fairly complex Behaviors.
Best,
Alex Wallace
On 30 June, 13:16, Dan Webb <d...@danwebb.net> wrote:
> Woah, just from running the test pages it seems that things are pretty
> messy. It looks like its something to do with the inheritance code
> which is nearly identical to prototypes except for the way the
> constructor is created...and that bit is pretty hairy to be fair. So,
> I'm going to dig in to this and try to root out the problem. This
> should probably fix most if not all of the problems we are
> experiencing though.
> If anyone else wants to dig in and assist that would be great.
> On Tue, Jun 30, 2009 at 5:22 PM, Dan Webb<d...@danwebb.net> wrote:
> > Yes, so this looks like it needs addressing urgently. I was really
> > hoping it was just a regression but no such luck. I need to read back
> > through all these posts but if anyone can summarise the problems
> > they've encountered so far here that would be a great help.
> > Thanks,
> > Dan
> > On Tue, Jun 30, 2009 at 5:19 PM, Zeno<zeno.crive...@gmail.com> wrote:
> >> Yes, Firefox v3.5 (final) has just been announced and many (Low Pro)
> >> stuff seems to be broken
> >> Grrrr
> >> Maybe we should start a new thread about FF related issues (talking
> >> about this in the Safari 4 thread is not very appropriate)
> >> On 29 June, 00:17, pushrod_admin <pushrodme...@googlemail.com> wrote:
> >>> This prob seemed to go away, but has now come back (for me) with the
> >>> latestFirefoxRC for Mac. ANyone else seeing this?
> >>> On 31 May, 11:51, Dan Webb <d...@danwebb.net> wrote:
> >>> > Yep, that's good news. I expected it to be the case as it looked more
> >>> > like a regression inFirefoxthan a bug in Low Pro.
> >>> > Thanks for letting us know, Nathan.
> >>> > On Sun, May 31, 2009 at 11:48 AM, MRoderick <roderick.mor...@gmail.com> wrote:
> >>> > > On 31 May, 09:46, Nathan <XGamer...@gmail.com> wrote:
> >>> > >> Well good news! Downloaded a newfirefox3.5 nightly (Shiretoko/
> >>> > >> 3.5pre) and this version works perfectly with my sample case and with
> >>> > >> my full website. They must have fixed the bug. This is excellent
> >>> > >> excellent news for me and lowpro users everywhere. Crisis averted. I
> >>> > >> can't wait for 3.5b5 which will hopefully incorporate this change.
> >>> > > That is great news! Thanks for reporting it :-)
That's really interesting. I'd just worked out the fix was to remove:
var behavior = arguments.callee;
You fix actually works but only bby mistake. The reason it works is
because of this:
return function() {
var initArgs = [this].concat(args);
behavior.attach.apply(behaving, initArgs);
};
The fact that you forgot to change the first reference of behavior to
behaving got around the bug. Man, there's something strange going on
there. I did learn something though. The function behavior actually
already has a reference to itself through the closure so that
arguments.callee stuff is completely redundant.
Thanks very much for that Alex. That really pointed me in the right
direction....I'd have been there for hours otherwise. If we ever meet
I owe you a beer :)
On Tue, Jun 30, 2009 at 8:57 PM, alex wallace<alexmlwall...@gmail.com> wrote:
> I've fixed the problems with Lowpro 0.5 in FF3.5 by changing the
> second definition of "var behavior" to "var behaving" -- it seems
> FF3.5 did not like the duplicate reference within the function block
> for the initial "var behavior".
> I've tested this so far in FF3.5 on XP and Leopard, FF3.0.11 on
> Leopard, Safari 4 on Leopard, Safari 3.2 on XP, Chrome, and our good
> friends Internet Explorer 6 and 7. Worksforme, testing with an
> interface chock-full of fairly complex Behaviors.
> Best,
> Alex Wallace
> On 30 June, 13:16, Dan Webb <d...@danwebb.net> wrote:
>> Woah, just from running the test pages it seems that things are pretty
>> messy. It looks like its something to do with the inheritance code
>> which is nearly identical to prototypes except for the way the
>> constructor is created...and that bit is pretty hairy to be fair. So,
>> I'm going to dig in to this and try to root out the problem. This
>> should probably fix most if not all of the problems we are
>> experiencing though.
>> If anyone else wants to dig in and assist that would be great.
>> On Tue, Jun 30, 2009 at 5:22 PM, Dan Webb<d...@danwebb.net> wrote:
>> > Yes, so this looks like it needs addressing urgently. I was really
>> > hoping it was just a regression but no such luck. I need to read back
>> > through all these posts but if anyone can summarise the problems
>> > they've encountered so far here that would be a great help.
>> > Thanks,
>> > Dan
>> > On Tue, Jun 30, 2009 at 5:19 PM, Zeno<zeno.crive...@gmail.com> wrote:
>> >> Yes, Firefox v3.5 (final) has just been announced and many (Low Pro)
>> >> stuff seems to be broken
>> >> Grrrr
>> >> Maybe we should start a new thread about FF related issues (talking
>> >> about this in the Safari 4 thread is not very appropriate)
>> >> On 29 June, 00:17, pushrod_admin <pushrodme...@googlemail.com> wrote:
>> >>> This prob seemed to go away, but has now come back (for me) with the
>> >>> latestFirefoxRC for Mac. ANyone else seeing this?
>> >>> On 31 May, 11:51, Dan Webb <d...@danwebb.net> wrote:
>> >>> > Yep, that's good news. I expected it to be the case as it looked more
>> >>> > like a regression inFirefoxthan a bug in Low Pro.
>> >>> > Thanks for letting us know, Nathan.
>> >>> > On Sun, May 31, 2009 at 11:48 AM, MRoderick <roderick.mor...@gmail.com> wrote:
>> >>> > > On 31 May, 09:46, Nathan <XGamer...@gmail.com> wrote:
>> >>> > >> Well good news! Downloaded a newfirefox3.5 nightly (Shiretoko/
>> >>> > >> 3.5pre) and this version works perfectly with my sample case and with
>> >>> > >> my full website. They must have fixed the bug. This is excellent
>> >>> > >> excellent news for me and lowpro users everywhere. Crisis averted. I
>> >>> > >> can't wait for 3.5b5 which will hopefully incorporate this change.
>> >>> > > That is great news! Thanks for reporting it :-)
Nice one. I thought the trick was to separate the variable names so
that attach could have the newly distinct var applied to it - but it
seems the naming conflict was easily avoided. Glad to have it fixed
and back in trunk.
Best,
A
On 30 June, 16:16, Dan Webb <d...@danwebb.net> wrote:
> The fact that you forgot to change the first reference of behavior to
> behaving got around the bug. Man, there's something strange going on
> there. I did learn something though. The function behavior actually
> already has a reference to itself through the closure so that
> arguments.callee stuff is completely redundant.
> Thanks very much for that Alex. That really pointed me in the right
> direction....I'd have been there for hours otherwise. If we ever meet
> I owe you a beer :)
> Fixed version is on github.
> On Tue, Jun 30, 2009 at 8:57 PM, alex wallace<alexmlwall...@gmail.com> wrote:
> > I've fixed the problems with Lowpro 0.5 in FF3.5 by changing the
> > second definition of "var behavior" to "var behaving" -- it seems
> > FF3.5 did not like the duplicate reference within the function block
> > for the initial "var behavior".
> > I've tested this so far in FF3.5 on XP and Leopard, FF3.0.11 on
> > Leopard, Safari 4 on Leopard, Safari 3.2 on XP, Chrome, and our good
> > friends Internet Explorer 6 and 7. Worksforme, testing with an
> > interface chock-full of fairly complex Behaviors.
> > Best,
> > Alex Wallace
> > On 30 June, 13:16, Dan Webb <d...@danwebb.net> wrote:
> >> Woah, just from running the test pages it seems that things are pretty
> >> messy. It looks like its something to do with the inheritance code
> >> which is nearly identical to prototypes except for the way the
> >> constructor is created...and that bit is pretty hairy to be fair. So,
> >> I'm going to dig in to this and try to root out the problem. This
> >> should probably fix most if not all of the problems we are
> >> experiencing though.
> >> If anyone else wants to dig in and assist that would be great.
> >> On Tue, Jun 30, 2009 at 5:22 PM, Dan Webb<d...@danwebb.net> wrote:
> >> > Yes, so this looks like it needs addressing urgently. I was really
> >> > hoping it was just a regression but no such luck. I need to read back
> >> > through all these posts but if anyone can summarise the problems
> >> > they've encountered so far here that would be a great help.
> >> > Thanks,
> >> > Dan
> >> > On Tue, Jun 30, 2009 at 5:19 PM, Zeno<zeno.crive...@gmail.com> wrote:
> >> >> Yes, Firefox v3.5 (final) has just been announced and many (Low Pro)
> >> >> stuff seems to be broken
> >> >> Grrrr
> >> >> Maybe we should start a new thread about FF related issues (talking
> >> >> about this in the Safari 4 thread is not very appropriate)
> >> >> On 29 June, 00:17, pushrod_admin <pushrodme...@googlemail.com> wrote:
> >> >>> This prob seemed to go away, but has now come back (for me) with the
> >> >>> latestFirefoxRC for Mac. ANyone else seeing this?
> >> >>> On 31 May, 11:51, Dan Webb <d...@danwebb.net> wrote:
> >> >>> > Yep, that's good news. I expected it to be the case as it looked more
> >> >>> > like a regression inFirefoxthan a bug in Low Pro.
> >> >>> > Thanks for letting us know, Nathan.
> >> >>> > On Sun, May 31, 2009 at 11:48 AM, MRoderick <roderick.mor...@gmail.com> wrote:
> >> >>> > > On 31 May, 09:46, Nathan <XGamer...@gmail.com> wrote:
> >> >>> > >> Well good news! Downloaded a newfirefox3.5 nightly (Shiretoko/
> >> >>> > >> 3.5pre) and this version works perfectly with my sample case and with
> >> >>> > >> my full website. They must have fixed the bug. This is excellent
> >> >>> > >> excellent news for me and lowpro users everywhere. Crisis averted. I
> >> >>> > >> can't wait for 3.5b5 which will hopefully incorporate this change.
> >> >>> > > That is great news! Thanks for reporting it :-)
On Tue, Jun 30, 2009 at 9:23 PM, alex wallace<alexmlwall...@gmail.com> wrote:
> Nice one. I thought the trick was to separate the variable names so
> that attach could have the newly distinct var applied to it - but it
> seems the naming conflict was easily avoided. Glad to have it fixed
> and back in trunk.
> Best,
> A
> On 30 June, 16:16, Dan Webb <d...@danwebb.net> wrote:
>> That's really interesting. I'd just worked out the fix was to remove:
>> var behavior = arguments.callee;
>> You fix actually works but only bby mistake. The reason it works is
>> because of this:
>> The fact that you forgot to change the first reference of behavior to
>> behaving got around the bug. Man, there's something strange going on
>> there. I did learn something though. The function behavior actually
>> already has a reference to itself through the closure so that
>> arguments.callee stuff is completely redundant.
>> Thanks very much for that Alex. That really pointed me in the right
>> direction....I'd have been there for hours otherwise. If we ever meet
>> I owe you a beer :)
>> Fixed version is on github.
>> On Tue, Jun 30, 2009 at 8:57 PM, alex wallace<alexmlwall...@gmail.com> wrote:
>> > I've fixed the problems with Lowpro 0.5 in FF3.5 by changing the
>> > second definition of "var behavior" to "var behaving" -- it seems
>> > FF3.5 did not like the duplicate reference within the function block
>> > for the initial "var behavior".
>> > I've tested this so far in FF3.5 on XP and Leopard, FF3.0.11 on
>> > Leopard, Safari 4 on Leopard, Safari 3.2 on XP, Chrome, and our good
>> > friends Internet Explorer 6 and 7. Worksforme, testing with an
>> > interface chock-full of fairly complex Behaviors.
>> > Best,
>> > Alex Wallace
>> > On 30 June, 13:16, Dan Webb <d...@danwebb.net> wrote:
>> >> Woah, just from running the test pages it seems that things are pretty
>> >> messy. It looks like its something to do with the inheritance code
>> >> which is nearly identical to prototypes except for the way the
>> >> constructor is created...and that bit is pretty hairy to be fair. So,
>> >> I'm going to dig in to this and try to root out the problem. This
>> >> should probably fix most if not all of the problems we are
>> >> experiencing though.
>> >> If anyone else wants to dig in and assist that would be great.
>> >> On Tue, Jun 30, 2009 at 5:22 PM, Dan Webb<d...@danwebb.net> wrote:
>> >> > Yes, so this looks like it needs addressing urgently. I was really
>> >> > hoping it was just a regression but no such luck. I need to read back
>> >> > through all these posts but if anyone can summarise the problems
>> >> > they've encountered so far here that would be a great help.
>> >> > Thanks,
>> >> > Dan
>> >> > On Tue, Jun 30, 2009 at 5:19 PM, Zeno<zeno.crive...@gmail.com> wrote:
>> >> >> Yes, Firefox v3.5 (final) has just been announced and many (Low Pro)
>> >> >> stuff seems to be broken
>> >> >> Grrrr
>> >> >> Maybe we should start a new thread about FF related issues (talking
>> >> >> about this in the Safari 4 thread is not very appropriate)
>> >> >> On 29 June, 00:17, pushrod_admin <pushrodme...@googlemail.com> wrote:
>> >> >>> This prob seemed to go away, but has now come back (for me) with the
>> >> >>> latestFirefoxRC for Mac. ANyone else seeing this?
>> >> >>> On 31 May, 11:51, Dan Webb <d...@danwebb.net> wrote:
>> >> >>> > Yep, that's good news. I expected it to be the case as it looked more
>> >> >>> > like a regression inFirefoxthan a bug in Low Pro.
>> >> >>> > Thanks for letting us know, Nathan.
>> >> >>> > On Sun, May 31, 2009 at 11:48 AM, MRoderick <roderick.mor...@gmail.com> wrote:
>> >> >>> > > On 31 May, 09:46, Nathan <XGamer...@gmail.com> wrote:
>> >> >>> > >> Well good news! Downloaded a newfirefox3.5 nightly (Shiretoko/
>> >> >>> > >> 3.5pre) and this version works perfectly with my sample case and with
>> >> >>> > >> my full website. They must have fixed the bug. This is excellent
>> >> >>> > >> excellent news for me and lowpro users everywhere. Crisis averted. I
>> >> >>> > >> can't wait for 3.5b5 which will hopefully incorporate this change.
>> >> >>> > > That is great news! Thanks for reporting it :-)
Thanks guys for addressing this issue so quickly! I was panicking when
I saw that ff3.5 reverted to being problematic with lowpro. Amazing
how the problem resurfaced after RC2 for the final release. I am glad
the fix was so easy as removing that one line with a redundant
variable though.
|
