[cmake-developers] Initial Attempt at Green Hill MULTI IDE Generator Support

[Geoffrey Viola]

Thanks for the feedback. I made some changes and rebased on the trunk.

> The hunk in Modules/CMakeSystemSpecificInformation.cmake is now just removing a blank line so it should not be needed.
Removed

>> +message("GHS-DetermineCompiler.cmake")
>
>This looks like a leftover debugging message.
Removed

>> there are a few changes to find boost libraries with this generator.
>[snip]
>> -if ( WIN32 AND Boost_USE_STATIC_LIBS AND NOT CYGWIN)
>> +if ("Windows" STREQUAL  ${CMAKE_HOST_SYSTEM_NAME} AND
>> +    Boost_USE_STATIC_LIBS AND NOT CYGWIN)
>
>Why does the host affect the library prefix for binaries meant for the target platform?
I'm still unsure if the WIN32 cmake variables is the target, platform or both. I originally thought this check should be based on the host, but it should be based on the target platform. I added a special case for the GHS Multi platform.

>> Some of the code is C++11. I'm not sure if that is an issue.
>
> It is.  We build on compilers that barely support C++98.
I removed usages of nullptr and std::anyof with a lamda expression.

>> Also, it skips the step where it determines the compiler by using the
>> force compiler macro. I don't see any other cmake module using this
>> functionality.
>[snip]
>> +set(CMAKE_SYSTEM_NAME "GHS-MULTI")
>> +set(CMAKE_SYSTEM_PROCESSOR "ARM")
>...
>> +include(CMakeForceCompiler)
>> +CMAKE_FORCE_CXX_COMPILER("cxarm" "GhsMultiArmCXX")
>
>Those APIs are meant for use in a CMAKE_TOOLCHAIN_FILE.
>
>The cmGlobalGhsMultiGenerator::EnableLanguage method should be able to take care of those.  This is the first time we've had a generator that locks to a specific target >platform (and implicitly cross-compiles).
>
>Is CMAKE_SYSTEM_VERSION always hard-coded on a given system or should the user be able to choose it in some way?
Platform/GHS-MULTI-C.cmake and Platform/GHS-MULTI-CXX.cmake have been moved into the enable language method.

The cmake system version can be changed with the GHS_COMP_ROOT cache variable. The version variable is not currently being used internally, since I've only tested it on what I think is the latest version. There was a bug in the previous use of find_program for CMAKE_MAKE_PROGRAM, where I thought the paths were interpreted as a regular expression.


Geoffrey Viola
SOFTWARE ENGINEER
T

+1.435.755.2980

ext 1077
M

+1.215.896.6521


asirobots.com




-----Original Message-----
From: Brad King [mailto:brad.king at kitware.com]
Sent: Wednesday, November 05, 2014 12:37 PM
To: Geoffrey Viola
Cc: cmake-developers at cmake.org
Subject: Re: [cmake-developers] Initial Attempt at Green Hill MULTI IDE Generator Support

On 11/05/2014 01:11 PM, Geoffrey Viola wrote:
> I rebased and squashed the previous commits and made some new changes.

Thanks.  Here are more comments.

The hunk in Modules/CMakeSystemSpecificInformation.cmake is now just removing a blank line so it should not be needed.

> +message("GHS-DetermineCompiler.cmake")

This looks like a leftover debugging message.

> there are a few changes to find boost libraries with this generator.
[snip]
> -if ( WIN32 AND Boost_USE_STATIC_LIBS AND NOT CYGWIN)
> +if ("Windows" STREQUAL  ${CMAKE_HOST_SYSTEM_NAME} AND
> +    Boost_USE_STATIC_LIBS AND NOT CYGWIN)

Why does the host affect the library prefix for binaries meant for the target platform?

> Some of the code is C++11. I'm not sure if that is an issue.

It is.  We build on compilers that barely support C++98.

> Also, it skips the step where it determines the compiler by using the
> force compiler macro. I don't see any other cmake module using this
> functionality.
[snip]
> +set(CMAKE_SYSTEM_NAME "GHS-MULTI")
> +set(CMAKE_SYSTEM_PROCESSOR "ARM")
...
> +include(CMakeForceCompiler)
> +CMAKE_FORCE_CXX_COMPILER("cxarm" "GhsMultiArmCXX")

Those APIs are meant for use in a CMAKE_TOOLCHAIN_FILE.

The cmGlobalGhsMultiGenerator::EnableLanguage method should be able to take care of those.  This is the first time we've had a generator that locks to a specific target platform (and implicitly cross-compiles).

Is CMAKE_SYSTEM_VERSION always hard-coded on a given system or should the user be able to choose it in some way?

Thanks,
-Brad

This message contains confidential information and is intended only for the recipient. If you are not the named addressee you should not disseminate, distribute or copy this e-mail. Please notify the sender immediately if you have received this e-mail by mistake and delete this e-mail from your system. Finally, the recipient should check this email and any attachments for the presence of viruses. The company accepts no liability for any damage caused by any virus transmitted by this email.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: 0001-Added-some-support-for-a-Green-Hill-MULTI.patch
Type: application/octet-stream
Size: 50910 bytes
Desc: 0001-Added-some-support-for-a-Green-Hill-MULTI.patch
URL: <http://public.kitware.com/pipermail/cmake-developers/attachments/20141117/eecbbd2c/attachment-0001.obj>