724137 - Support debug trace printouts in EGL

Status: RESOLVED FIXED

(Core :: Graphics, defect)

Description

[Joe Drew (not getting mail)]

Attachment: debug output egl

There is currently no MOZ_GL_DEBUG_VERBOSE support from EGL functions, which can be a little annoying. This patch implements that (and probably has some other cruft in it that needs to be fixed before committing).

(Specifically, all the #ifdef DEBUG -> #if 1 changes will need to be reverted)

Comment 1

[Chris Jones [:cjones] inactive; ni?/f?/r? if you need me]

Aside: have you given apitrace a spin?

Comment 2

[Joe Drew (not getting mail)]

Benoit got the apitrace patches to apply and compile again, but says it crashed when he tried to use it. Neither of us have tried more than that.

Comment 3

[Benoit Jacob [:bjacob] (mostly away)]

Comment on attachment 594331 [details] [diff] [review]
debug output egl

Review of attachment 594331 [details] [diff] [review]:
-----------------------------------------------------------------

::: gfx/gl/GLContext.cpp
@@ +58,5 @@
>  
>  namespace mozilla {
>  namespace gl {
>  
> +#if 1

Did you really intend to get review for this patch? The above makes it look more like a useful local patch, not for landing as-is.

Comment 4

[Joe Drew (not getting mail)]

Attachment: add EGL functions output ifdef DEBUG

That's why I said "(Specifically, all the #ifdef DEBUG -> #if 1 changes will need to be reverted)" :)

Anyways, here is a cleaned-up version.

Comment 5

[Benoit Jacob [:bjacob] (mostly away)]

Woops, sorry about that.

Comment 6

[Benoit Jacob [:bjacob] (mostly away)]

Comment on attachment 594812 [details] [diff] [review]
add EGL functions output ifdef DEBUG

Review of attachment 594812 [details] [diff] [review]:
-----------------------------------------------------------------

::: gfx/gl/GLContextProviderEGL.cpp
@@ +672,5 @@
>              return false;
>          }
>  
>  #define SYMBOL(name) \
> +    { (PRFuncPtr*) &mSymbols.f##name, { "egl" #name, NULL } }

Hah, that's clever. I wonder if it can be a little bit dangerous to use such a common name as SYMBOL for that, but that's a cpp, not a header, after all.

Are you going to generalize this trick to all our *GL symbols? I would, for once, support this use of macros!

Comment 7

[Joe Drew (not getting mail)]

Note that I didn't create SYMBOL, I just changed its implementation.

I don't currently have plans to generalize this trick, as I'm focused on EGL, but I have no particular issues with doing it some time in the mysterious future.

Comment 8

[Mozilla RelEng Bot]

Try run for cd3085bdfb4c is complete.
Detailed breakdown of the results available here:
    https://tbpl.mozilla.org/?tree=Try&rev=cd3085bdfb4c
Results (out of 30 total builds):
    success: 29
    warnings: 1
Builds (or logs if builds failed) available at:
http://ftp.mozilla.org/pub/mozilla.org/firefox/try-builds/[e-mail]-cd3085bdfb4c

Comment 9

[Mozilla RelEng Bot]

Try run for 74b43278cbcd is complete.
Detailed breakdown of the results available here:
    https://tbpl.mozilla.org/?tree=Try&rev=74b43278cbcd
Results (out of 28 total builds):
    success: 25
    warnings: 1
    failure: 2
Builds (or logs if builds failed) available at:
http://ftp.mozilla.org/pub/mozilla.org/firefox/try-builds/[e-mail]-74b43278cbcd

Comment 10

[Joe Drew (not getting mail)]

https://hg.mozilla.org/integration/mozilla-inbound/rev/c4799638be1d

Comment 11

[Ed Morley [:emorley]]

https://hg.mozilla.org/mozilla-central/rev/c4799638be1d