Commit Graph

10 Commits

Author SHA1 Message Date
David Nadlinger
f85ed37cf1 Revert "Use CMake-detected make program."
CMAKE_MAKE_PROGRAM isn't always (GNU) make, but refers to the
"make" equivalent of the target generator, for example the
"ninja" executable if the Ninja generator is used.

Using find_program or something similar would probably still
be a good idea.

This reverts commit 5a6176316a.
2013-02-13 16:04:01 +01:00
David Nadlinger
aa4543465d Actually run DMD test suite in 32 bit mode on x86_64 multilib builds.
Yet another regression gone unnoticed because CMake functions
silently accept extra arguments.
2012-12-20 03:57:04 +01:00
David Nadlinger
3fc9a0faa3 CMake 2.8.0 compatibility. 2012-12-11 23:08:13 +01:00
David Nadlinger
a74e4abe58 Fix testsuite target dependency setup.
No idea how this worked locally.
2012-12-07 00:04:03 +01:00
David Nadlinger
5a6176316a Use CMake-detected make program.
This probably doesn't make a huge difference in practice, as the
DMD test suite scripts require a Posix-ish environment anyway.
2012-12-06 23:17:32 +01:00
David Nadlinger
6cb0193c53 Clean files from previous runs before executing DMD testsuite. 2012-12-06 23:15:45 +01:00
David Nadlinger
18e0fafe8e Run tests both in 32 and 64 bit mode on multilib builds. 2012-12-06 21:57:18 +01:00
David Nadlinger
0c0e88dda1 Add missing LDMD option handling code.
Reusing the existing parsing code for DMD compatibility was a good
idea, copy-pasting the tool together in the middle of the night not
so much.
2012-10-03 19:51:04 +02:00
David Nadlinger
d5d649b9c9 Build DMD testsuite in a debug and a release configuration. 2012-09-29 23:10:30 +02:00
David Nadlinger
c0980de3af Testsuite updates; run it as part of CI builds.
Only can run a quick build because of Travis time restrictions. Should
try to find a set of most-likely-to-fail command line args, maybe
"-O -gc -fPIC"?
2012-09-29 20:59:06 +02:00