856215 - Codegen for converting a sequence of sequences is buggy

Status: RESOLVED FIXED

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

Description

[Boris Zbarsky [:bzbarsky]]

We end up with something like this:


      for (uint32_t i = 0; i < length; ++i) {
...
            for (uint32_t i = 0; i < length; ++i) {
...
                if (!xpc::NonVoidStringToJsval(cx, currentValue[i][i], &tmp)) {

which is obviously not a good idea.

Comment 1

[Boris Zbarsky [:bzbarsky]]

Hrm.  So the problem is that we don't actually know the type of currentValue[i] when we're generating this code, so we can't easily create a temporary for it.  :(

I need to think about this a tad....

Comment 2

[Boris Zbarsky [:bzbarsky]]

So this is a bit annoying.

I can hack this together, sort of, but it fails any time the declType of a method return value and dictionary member type are different, which they are for dictionary types for example.

If we had C++11 decltype(), I could just decltype(currentValue)::elem_type.  But we don't have that on all our supported platforms yet.

Maybe I need to special-case sequence-of-sequences conversion to call a function of some sort.

Or maybe I need to keep a global nesting level and name my counter variables with that level included....  That last might be simplest.  :(

Comment 3

[Boris Zbarsky [:bzbarsky]]

Attachment: Fix conversion of sequences of sequences to JS to not screw up its indexing.

This is ugly, but seems to work....

Comment 4

[Boris Zbarsky [:bzbarsky]]

Valentin, does that patch help bug 843865?

Comment 5

[Valentin Gosu [:valentin] (he/him)]

It seems to be working perfectly now. Thanks!

Comment 6

[Boris Zbarsky [:bzbarsky]]

https://hg.mozilla.org/integration/mozilla-inbound/rev/125fb548eaba

Comment 7

[Ed Morley [:emorley]]

https://hg.mozilla.org/mozilla-central/rev/125fb548eaba