ogl-sample
[Top] [All Lists]

Re: [ogl-sample] gl.spec file missing offset values

To: ogl-sample@xxxxxxxxxxx
Subject: Re: [ogl-sample] gl.spec file missing offset values
From: Jon Leech <ljp@xxxxxxxxxxxxxxxxxxxx>
Date: Fri, 3 Nov 2000 08:41:27 -0800
In-reply-to: <39FE0E05.CE10A154@xxxxxxxxxxx>; from Brian Paul on Mon, Oct 30, 2000 at 05:10:45PM -0700
References: <14845.58250.506669.313124@xxxxxxxxxxxxxxxxxxx> <39FDE614.B340AEE7@xxxxxxxxxxx> <14846.3002.95992.559934@xxxxxxxxxxxxxxxxxxx> <39FE0E05.CE10A154@xxxxxxxxxxx>
Reply-to: ogl-sample@xxxxxxxxxxx
Sender: owner-ogl-sample@xxxxxxxxxxx
On Mon, Oct 30, 2000 at 05:10:45PM -0700, Brian Paul wrote:
> I explained the "alias DeleteTextures" in another email last week.
> Didn't that go through either???
>
> I can live without that fix but if it's not an alias function then it
> should have its own dispatch offset.

    DeleteTextures is defined to use different protocol from the EXT
version, so they are not aliases according to the meaning used in the
spec files (which is that not just the GL semantics, but *all* behavior
beyond the subroutine entry point must be identical). Most extension
entry points which get promoted really are aliases in this sense,
because they share the same GLX rendering opcode, but for queries the
protocol is changed from vendor private to GLX "single". So yes, it
should have its own dispatch offset.

    The spec file needs a way to distinguish between aliasing on the
client and server side - probably a new alias keyword - to capture this
distinction, so that more of the GLX client/server code could be
autogenerated.

    Jon

<Prev in Thread] Current Thread [Next in Thread>
  • Re: [ogl-sample] gl.spec file missing offset values, Jon Leech <=