Fixup D3D11 types for builds with Win8 SDK and moz SDK version of 601

RESOLVED FIXED in mozilla23

Status

()

RESOLVED FIXED
6 years ago
6 years ago

People

(Reporter: jimm, Assigned: jimm)

Tracking

Trunk
mozilla23
x86_64
Windows 7
Points:
---

Firefox Tracking Flags

(Not tracked)

Details

Attachments

(1 attachment)

(Assignee)

Description

6 years ago
https://bugzilla.mozilla.org/attachment.cgi?id=747398&action=edit

This needs another update for devs who build without enable-metro.
(Assignee)

Comment 1

6 years ago
Created attachment 747910 [details] [diff] [review]
patch

d3dcommon.h is available in both d3d sdk sources we support. If you are using the win8 sdk (and have your paths set right) the 8.0 sdk version should get included. Otherwise the June 2010 d3d sdk should get included. 

Only thing I'm not too sure about is if this needs to be wrapped in some sort of D3D ifdef. Doesn't seem like it since we require D3D headers to build, so I think checks like CAIRO_HAS_D2D_SURFACE are obsolete.
Attachment #747910 - Flags: review?(khuey)
FWIW, this is causing Windows B2G desktop build bustage as well (which we didn't notice because they're hidden by default on inbound/m-c).
When this patch gets r+, please land it on birch rather than inbound.
(Assignee)

Comment 4

6 years ago
Comment on attachment 747910 [details] [diff] [review]
patch

whoever gets to it first.
Attachment #747910 - Flags: review?(bas)
Comment on attachment 747910 [details] [diff] [review]
patch

Review of attachment 747910 [details] [diff] [review]:
-----------------------------------------------------------------

::: gfx/thebes/gfxWindowsPlatform.h
@@ +33,5 @@
>  #ifdef CAIRO_HAS_D2D_SURFACE
>  #include <dxgi.h>
>  #endif
>  
> +// This header is available in the June 2010 SDK and in the Win8 SDK  

nit: whitespace at eol.
Attachment #747910 - Flags: review?(khuey) → review+
(Assignee)

Updated

6 years ago
Attachment #747910 - Flags: review?(bas)
https://hg.mozilla.org/mozilla-central/rev/25c04296c675
Status: NEW → RESOLVED
Last Resolved: 6 years ago
Resolution: --- → FIXED
Target Milestone: --- → mozilla23
Would't it be safer to test for the definition of each of these separately rather than assuming one being undefined means they all are?  There might be another SDK 7 update that breaks this the way it is now.
You need to log in before you can comment on or make changes to this bug.