I've been working on determining on how to do the platform part of this. My first guess is we could do something like this in the platform files:(from gcc.cmake)IF(CMAKE_COMPILER_IS_GNUCC) SET (CMAKE_C_FLAGS_INIT "") SET (CMAKE_C_FLAGS_DEBUG_INIT "-g") SET (CMAKE_C_FLAGS_MINSIZEREL_INIT "-Os
Hi,
We are currently using MSBuild to select the build configuration. The
command line looks like this:
msbuild All.sln /property:Configuration=Release
/logger:FileLogger,Microsoft.Build.Engine;logfile=All.log
This line will build the ALL_BUILD target in release mode. (We tried to
use VCBuild,
At 02:39 AM 9/8/2006, Michael Casadevall wrote:
I've been working on determining on how to do the platform part of this. My
first guess is we could do something like this in the platform files:
(from gcc.cmake)
IF(CMAKE_COMPILER_IS_GNUCC)
SET (CMAKE_C_FLAGS_INIT )
SET
On 9/6/06, Brandon J. Van Every [EMAIL PROTECTED] wrote:
Actually it's not invalid to compile it. That's what a cross-compiler
does. It's invalid to run it. CMake already has this distinction
between TRY_COMPILE and TRY_RUN, so if TRY_RUN is used inappropriately,
that's user error. If CMake
Hrm, interesting idea, I admit that it didn't occur to me, although I
don't see how it really is that different. In your case, you simply
setting the same variables twice for local and cross-compilation,
which seems a little kludgy. I mean, you still need to handle all the
CMAKE_CROSS
At 12:28 PM 9/8/2006, Michael Casadevall wrote:
Hrm, interesting idea, I admit that it didn't occur to me, although I
don't see how it really is that different. In your case, you simply
setting the same variables twice for local and cross-compilation,
which seems a little kludgy. I mean, you
Is there a way to override the mechanism for finding dependencies? Fortran 90
with conditionally compiled modules is not handled correctly. It would be nice
to be able to take the results of cpp -M on a source file and append
additional dependencies found by looking at the preprocessed source. I
Warren Turkal wrote:
Is there a way to override the mechanism for finding dependencies?
No.
I am trying to work around the crappy support for preprocessed Fortran
dependency generation. See bug #2361 [1].
Unfortunately we do not have time to solve this problem ourselves right
now. Your time
At 03:02 PM 9/8/2006, Alan W. Irwin wrote:
On 2006-09-08 13:59-0400 William A. Hoffman wrote:
As a CMake developer I think it would be much easier to do it with
two makefiles. As far as I can tell there are two modes of cross compiling.
1. The whole project is being built for some other
On Friday 08 September 2006 13:11, you wrote:
#ifdef FOO
use abc
#else
use xyz
#endif
That's exactly the type of thing causing the problems.
wt
--
Warren Turkal, Research Associate III/Systems Administrator
Colorado State University, Dept. of Atmospheric Science
Hi,
Von: William A. Hoffman [EMAIL PROTECTED]
...
The trouble is that in the case of 2, when you build
some sort of code generation executable as part of the build and run it
during
the build, it has to be built for the host or local machine.
Exactly.
Just imagine somebody might want to
-Original Message-
From: [EMAIL PROTECTED]
To: [EMAIL PROTECTED]
Sent: Fri, 8 Sep 2006 4:06 PM
Subject: Re: [CMake] Adding cross-compiler support to CMake ...
lol, I actually intended to try cross-compiling KDE as a demostration of CMake's cross-compiling abilities once it
At 04:06 PM 9/8/2006, [EMAIL PROTECTED] wrote:
-Original Message-
From: [EMAIL PROTECTED]
To: [EMAIL PROTECTED]
Sent: Fri, 8 Sep 2006 4:06 PM
Subject: Re: [CMake] Adding cross-compiler support to CMake ...
lol, I actually intended to try cross-compiling KDE as a demostration of
At 04:33 PM 9/8/2006, Alexander Neundorf wrote:
How should all that work when cross-compiling ?
The FindFoo.cmake files as far as I can see cannot and are not written to find
libraries in strange directories for cross-compiling.
All the values could be set manually by hand, but this is not
On 2006-09-08 22:33+0200 Alexander Neundorf wrote:
[...]Once it is possible to do this, the cross-compiling abilities are really
perfect.
In KDE we do about one million configure checks, checking the availability of
various functions, types and headers. And we use many extra libraries, which
Benjamin Reed wrote:
On 9/6/06, Brandon J. Van Every [EMAIL PROTECTED] wrote:
Actually it's not invalid to compile it. That's what a cross-compiler
does. It's invalid to run it. CMake already has this distinction
between TRY_COMPILE and TRY_RUN, so if TRY_RUN is used inappropriately,
that's
It would just need to be in the cross-compiler's directory. The test
needs to do TRY_COMPILE on the cross-compiler, and as long as it
compiles, the library is in the right place. TRY_RUN should be able
to run tests remotely, or have a manual answer put in place just
incase the target
17 matches
Mail list logo