Breaking Compatibility Between Beta and Final

[Shawn Wilsher]

Hey all,

Over in bug 496019 [1], a 1.9.2 blocker, I've had to change the way a
method works, and added a method on the same interface (both changes
require a uuid revision). It is not possible to fix this bug without
making those changes, so I am at a loss as to what I should do here.

Is it OK to break compatibility here? Do you need more information from
me before deciding this?

Cheers,

Shawn

[1] https://bugzilla.mozilla.org/show_bug.cgi?id=496019

[dietrich ayala]

when was the api introduced? what has it shipped in?

can you get info from AMO about what extensions use the api?

if the api can't be used without risk of crash, then it seems like we
want to "help" the ecosystem uplift here. is this the case?

[Shawn Wilsher]

Since I've had zero responses on this, I guess I'll provide more
information in hopes of getting some kind of decision.

This bug was made a blocker because it is the cause of a random orange
bug we are seeing on Linux tinderboxes (bug 519438 [1]). It's a crash
in gtk code when spinning a nested event loop on the main thread. We
may be vulnerable to this crash in other places where we spin the event
loop on the main thread (Weave, sync XHR), but bug 496019 would make it
go away on tinderbox at least.

It should be noted that we shipped 1.9.1 with this behavior (spinning
the event loop).

Cheers,

Shawn

[1] https://bugzilla.mozilla.org/show_bug.cgi?id=519438

On 11/5/09 1:20 PM, Shawn Wilsher wrote:
> Hey all,
>
> Over in bug 496019 [1], a 1.9.2 blocker, I've had to change the way a
> method works, and added a method on the same interface (both changes
> require a uuid revision). It is not possible to fix this bug without
> making those changes, so I am at a loss as to what I should do here.
>
> Is it OK to break compatibility here? Do you need more information from
> me before deciding this?
>
> Cheers,
>
> Shawn
>
> [1] https://bugzilla.mozilla.org/show_bug.cgi?id=496019

> _______________________________________________
> dev-platform mailing list
> dev-pl...@lists.mozilla.org
> https://lists.mozilla.org/listinfo/dev-platform

[Shawn Wilsher]

On 11/9/09 12:29 PM, dietrich ayala wrote:
> when was the api introduced? what has it shipped in?

The Close API? It's been around since day one of Storage as far as I
can tell. This behavior of spinning the event loop on Close was
introduced in Gecko 1.9.1 with the new asynchronous storage API.

> can you get info from AMO about what extensions use the api?

Would love to get that information, but that's blocked on bug 358591 [1].

> if the api can't be used without risk of crash, then it seems like we
> want to "help" the ecosystem uplift here. is this the case?

It's a crash only on Linux, and it's a crash in system libraries.

/sdwilsh

[1] https://bugzilla.mozilla.org/show_bug.cgi?id=358591

[Gavin Sharp]

----- "Shawn Wilsher" <sdw...@mozilla.com> wrote:
> > can you get info from AMO about what extensions use the api?
> Would love to get that information, but that's blocked on bug 358591
> [1].

http://mxr-test.konigsberg.mozilla.org/addons/ is out of date, but it's better than nothing in the interim (you may have to ask timeless to whitelist your IP). Either way, MXR is only useful to find JS callers, right?

Gavin

[Shawn Wilsher]

On 11/9/09 12:45 PM, Gavin Sharp wrote:
> http://mxr-test.konigsberg.mozilla.org/addons/ is out of date, but it's better than nothing in the interim (you may have to ask timeless to whitelist your IP). Either way, MXR is only useful to find JS callers, right?

Yeah, and there are consumers of it. Doesn't look like any of them use
the asynchronous API, but then I also don't know how old this snapshot
is. It's possible nobody in that snapshot could use it.

Cheers,

Shawn

[Gijs Kruitbosch]

On 09/11/2009 21:39 PM, Shawn Wilsher wrote:
>> can you get info from AMO about what extensions use the api?
> Would love to get that information, but that's blocked on bug 358591 [1].

In the meantime, there is http://mxr-test.konigsberg.mozilla.org/addons/

http://mxr-test.konigsberg.mozilla.org/addons/search?string=mozIStorageConnection&find=&findi=&filter=^[^\0]*%24&hitlimit=&tree=addons
is small, and it seems only

https://addons.mozilla.org/en-US/firefox/addon/5817

actually calls the method in question.

The bad news is the scrape on mxr-test dates from november last year, and there
might well be new add-ons or new versions of add-ons that now use it and didn't
before...

~ Gijs