1359491 - stylo: do we really not need to handle copy-on-write stuff for servo stylesheets?

Status: RESOLVED DUPLICATE of bug 1339629

(Core :: CSS Parsing and Computation, enhancement)

Description

[Boris Zbarsky [:bzbarsky]]

Now that bug 1290218 is fixed, is the code/comment in nsPresContext::EnsureSafeToHandOutCSSRules still correct?  If so, it could use an explanation of why there is no need to handle copy-on-write there, because it's entirely non-obvious to me: we're sharing a mutable data structure.

Relatedly, the comments on the SheetInfo getters on StyleSheet.h seem to be out of date; the info lives on the inner in both cases...

Comment 1

[Boris Zbarsky [:bzbarsky]]

Note that if we _do_ need this handled, we also need to handle various cases when Gecko calls EnsureUniqueInner in general.

Comment 2

[Cameron McCormack (:heycam)]

(In reply to Boris Zbarsky [:bz] (still a bit busy) (if a patch has no decent message, automatic r-) from comment #0)
> Now that bug 1290218 is fixed, is the code/comment in
> nsPresContext::EnsureSafeToHandOutCSSRules still correct?  If so, it could
> use an explanation of why there is no need to handle copy-on-write there,
> because it's entirely non-obvious to me: we're sharing a mutable data
> structure.

No, I don't think that comment is correct.  I'm pretty sure we do actually need to clone sheet contents in response to CSSOM calls...

Comment 3

[Cameron McCormack (:heycam)]

(In reply to Cameron McCormack (:heycam) from comment #2)
> No, I don't think that comment is correct.  I'm pretty sure we do actually
> need to clone sheet contents in response to CSSOM calls...

What it seems like we don't need to do is support cloning the C++ wrapper objects like ServoStyleRule, since they should never have been instantiated before CSSOM calls happen.

I think this bug can just be duped to bug 1339629.