Bug 29545

Summary:	[bisected glsl2]piglit glslparsertest_preprocess1.frag fails
Product:	Mesa	Reporter:	fangxun <xunx.fang>
Component:	Mesa core	Assignee:	Carl Worth <cworth>
Status:	VERIFIED FIXED	QA Contact:
Severity:	normal
Priority:	high	CC:	cworth
Version:	unspecified	Keywords:	regression
Hardware:	All
OS:	All
Whiteboard:
i915 platform:		i915 features:

Description fangxun 2010-08-12 23:08:51 UTC

System Environment:
--------------------------
Arch    x86_64
Platform    piketon
Libdrm:         (master)2.4.21-15-gb04515c5d6c95f573457a94267b855cceb639105
Mesa:           (glsl2)5a805079a8d209e843661941730ecfebb65d2913
Xserver:      (master)xorg-server-1.8.99.905-22-g6dae7f3792611aace1df0cca63bf50c50d93de43
Xf86_video_intel: (master)2.12.0-66-g19c48d3b3f33582baa87a9b3a9189e320e4cea45
Kernel: 2.6.35 (for-linus)9fe6206f400646a2322096b56c59891d530e8d51


Bug detailed description:
-------------------------
Piglit case glslparsertest_preprocess1.frag fails on mesa glsl2 branch. Bisect find 764e096647ec8c0f20ea3b5191499af806ad23f8 is first bad commit:
commit 764e096647ec8c0f20ea3b5191499af806ad23f8
Author: Carl Worth <cworth@cworth.org>
Date:   Wed Aug 11 12:45:55 2010 -0700

    glcpp: Regenerate glcpp-parse.c

    After a recent change to glcpp-parse.y


Reproduce steps:
----------------
1.xinit&
2.glslparsertest preprocess1.frag fail

Comment 1 Carl Worth 2010-08-17 23:30:01 UTC

Thanks for the report.

This is an interesting case. The test case tries to verify correct operation
of the preprocessor, but only tests in a very vague way, (provides a broken
file and tests that the compiler fails to compile it).

So before the bisected commit, the compiler was failing on this test (due
to a compiler bug), but was not failing in the way the test actually intended
(which really wanted a "macro redefined" error).

The bisected commit fixed the compiler bug so that the code actually compiled,
but this caused the test case to fail.

The new bug, then, has nothing to do with the bisected commit, but was simply
that the preprocessor was failing to generate errors in the case of redefined
macros.

I've now implemented this, (along with two new test cases to the preprocessor's
test suite---which is much more thorough in what it looks for), and I've pushed
the changes to mesa master.

So this bug is fixed, now.

-Carl

Comment 2 Gordon Jin 2010-08-18 19:51:26 UTC

verified with mesa master d442a01ac14382d83cdaac87d2832315ceb3e963.

Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.