812742 - Sequence codegen is a major footgun

Status: RESOLVED FIXED

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

Description

[Boris Zbarsky [:bzbarsky]]

It passes through "(*arr.AppendElement())" as the declName, so if the inner type conversion template uses ${declName} twice bad things happen.  Luckily nothing does that, as far as I can tell... until bug 767933.

Comment 1

[Boris Zbarsky [:bzbarsky]]

Attachment: Fix sequence codegen to not do AppendElement multiple times by accident.

Comment 2

[Peter Van der Beken [:peterv]]

Comment on attachment 682719 [details] [diff] [review]
Fix sequence codegen to not do AppendElement multiple times by accident.

Review of attachment 682719 [details] [diff] [review]:
-----------------------------------------------------------------

::: dom/bindings/Codegen.py
@@ +2221,5 @@
>                      {
>                          "val" : "temp",
>                          "valPtr": "&temp",
> +                        "declName" : "slot",
> +                        "holderName": "tempHolder",

This tempHolder shouldn't be used, right? Can we keep not passing a holderName? If not, we should at least add a comment about that.

Comment 3

[Boris Zbarsky [:bzbarsky]]

http://hg.mozilla.org/integration/mozilla-inbound/rev/c3d44fd7dce4 with a comment added explaining the holderName bit.

Bug 767933 will make us fail tests without this patch, so in-testsuite+.

Comment 4

[Ed Morley [:emorley]]

https://hg.mozilla.org/mozilla-central/rev/c3d44fd7dce4