851908 - Assertion failure: !domClass || clasp->mPrototypeID != prototypes::id::_ID_Count (Why do we have a hasInstance hook if we don't have a prototype ID?) without clobber

Status: RESOLVED FIXED

(Core :: DOM: Core & HTML, defect)

Description

[:Ms2ger (he/him; ⌚ UTC+1/+2)]

See <https://tbpl.mozilla.org/php/getParsedLog.php?id=20742583&tree=Firefox> for example.

Comment 1

[Andrew McCreight [:mccr8]]

My patches in bug 827486 caused a similar problem in non-clobber builds.  It is especially weird because I'm not sure what it actually changed.  I'll look into that at some point.

Windows debug.

M4: assertion failure isEmpty(), same signature (mozilla::dom::InterfaceHasInstance)

https://tbpl.mozilla.org/php/getParsedLog.php?id=20730944&tree=Mozilla-Inbound

M1, bc: same thing as comment 0.
Assertion failure: !domClass || clasp->mPrototypeID != prototypes::id::_ID_Count
in InterfaceHasInstance

https://tbpl.mozilla.org/php/getParsedLog.php?id=20731519&tree=Mozilla-Inbound
https://tbpl.mozilla.org/php/getParsedLog.php?id=20730924&tree=Mozilla-Inbound

Comment 2

[Kyle Huey (Exited; not receiving bugmail, old account, do not use)]

Yeah the build system is still broken.

Comment 3

[Boris Zbarsky [:bzbarsky]]

So wait.

Which part is not getting rebuilt right?  The HasInstance hook is just part of the binding file, so if anything changed about the binding I'd think it would get rebuilt...  Kyle, do you have any idea what's going on here?

Do we have any sort of reliable steps to reproduce locally with dep builds?

Comment 4

[Olli Pettay [:smaug][[e-mail]]]

At least I haven't managed to reproduce locally, and didn't see the problem when pushing to try
(but try does clobber builds). But when I pushed some event patches to m-i, it reported about this
issue. Clobber helped.

Comment 5

[Kyle Huey (Exited; not receiving bugmail, old account, do not use)]

(In reply to Boris Zbarsky (:bz) from comment #3)
> Kyle, do you have any idea what's going on here?

Yes, the problem is that we are blowing out the MSYS command line limit when we process the dependency files, which causes the spawned shell to crash, which means we don't actually have any dependency data.  This is limited to Windows because MSYS.  Bug 850153 was supposed to fix this but didn't cover all of the cases.  I think bug 462463 fixed this for real by killing off the need to spawn the shell at all.

Please reopen if you see it again.

Comment 6

[Andrew McCreight [:mccr8]]

blocking bug backed out