Openembedded Core Discussions
 help / color / mirror / Atom feed
From: Stefan Herbrechtsmeier <stefan@herbrechtsmeier.net>
To: Jonathan Liu <net147@gmail.com>
Cc: "openembedded-core@lists.openembedded.org"
	<openembedded-core@lists.openembedded.org>
Subject: Re: [PATCH] cmake.bbclass: set ar/ld/nm/objcopy/objdump/ranlib/strip native paths
Date: Thu, 18 Jul 2013 11:46:10 +0200	[thread overview]
Message-ID: <51E7B962.4050108@herbrechtsmeier.net> (raw)
In-Reply-To: <CANwerB3M3p7q1iMHUcjdNdL68ozr0VDcDwor0afiSjZdx+24AQ@mail.gmail.com>

Am 18.07.2013 02:47, schrieb Jonathan Liu:
> On 17 July 2013 22:12, Stefan Herbrechtsmeier
> <stefan@herbrechtsmeier.net> wrote:
>> Am 17.07.2013 12:57, schrieb Jonathan Liu:
>>
>>> On 17/07/2013 7:49 PM, Jonathan Liu wrote:
>>>> On 17/07/2013 7:13 PM, Stefan Herbrechtsmeier wrote:
>>>>> Am 17.07.2013 10:40, schrieb Jonathan Liu:
>>>>>> The CMAKE_{AR,LINKER,NM,OBJCOPY,OBJDUMP,RANLIB,STRIP} cmake variables
>>>>>> are currently set to values like CMAKE_LINKER-NOTFOUND for native
>>>>>> recipes because the host paths are not searched. This is because the
>>>>>> CMAKE_FIND_ROOT_PATH_MODE_PROGRAM cmake variable is set to ONLY.
>>>>>>
>>>>>> To resolve this, explicitly set the variables using FIND_PROGRAM and
>>>>>> pass the CMAKE_FIND_ROOT_PATH_BOTH option so the host paths are also
>>>>>> searched.
>>>>> Can you please debug the issue. Regarding my current check the variables
>>>>> are set in CMakeFindBinUtils.cmake and this is included by
>>>>> CMakeDetermineCCompiler.cmake. The later set _CMAKE_TOOLCHAIN_LOCATION based
>>>>> on the CMAKE_C_COMPILER variable and this is used as search location for the
>>>>> find_program calls. Can you check the value of _CMAKE_TOOLCHAIN_LOCATION?
>>>>>
>>>> Within the context of the CMakeLists.txt file:
>>>> CMAKE_C_COMPILER is /usr/bin/gcc.
>>>> _CMAKE_TOOLCHAIN_LOCATION is not set to anything.
>>>>
>>>> If I add message(test) to
>>>> build/tmp/sysroots/x86_64-linux/usr/share/cmake-2.8/Modules/CMakeDetermineCCompiler.cmake,
>>>> it doesn't get printed out when running cmake on CMakeLists.txt (cmake is
>>>> build/tmp/sysroots/x86_64-linux/usr/bin/cmake).
>>>>
>>>> Regards,
>>>> Jonathan
>>> For:
>>> find_program(CMAKE_AR NAMES
>>> ${_CMAKE_TOOLCHAIN_PREFIX}ar${_CMAKE_TOOLCHAIN_SUFFIX} HINTS
>>> ${_CMAKE_TOOLCHAIN_LOCATION})
>>>
>>> It seems the ${_CMAKE_TOOLCHAIN_LOCATION} hint is not used when
>>> CMAKE_FIND_ROOT_PATH_MODE_PROGRAM is set to ONLY.
>> You are right. The CMAKE_FIND_ROOT_PATH is used also as base for the HINTS
>> paths.
>>
>> You can use _CMAKE_TOOLCHAIN_LOCATION to pass NO_CMAKE_FIND_ROOT_PATH to the
>> find_program calls:
>> set(_CMAKE_TOOLCHAIN_LOCATION "" NO_CMAKE_FIND_ROOT_PATH)
>>
>> This should work but I am not sure if it is a good solution.
>>
> I prefer setting documented variables rather than hacking the
> undocumented _CMAKE_TOOLCHAIN_LOCATION variable to add the
> NO_CMAKE_FIND_ROOT_PATH flag for FIND_PROGRAM.
I think it is a issue in cmake and the _CMAKE_TOOLCHAIN_LOCATION don't 
work with CMAKE_FIND_ROOT_PATH_MODE_PROGRAM set to ONLY.
I have open an issue in cmake to get some feedback. The could patch all 
find_program calls in CMakeFindBinUtils.cmake modules:
-  find_program(CMAKE_AR NAMES 
${_CMAKE_TOOLCHAIN_PREFIX}ar${_CMAKE_TOOLCHAIN_SUFFIX} HINTS 
${_CMAKE_TOOLCHAIN_LOCATION})
+  find_program(CMAKE_AR NAMES 
${_CMAKE_TOOLCHAIN_PREFIX}ar${_CMAKE_TOOLCHAIN_SUFFIX} PATHS 
"${_CMAKE_TOOLCHAIN_LOCATION}" NO_DEFAULT_PATH NO_CMAKE_FIND_ROOT_PATH)
+  find_program(CMAKE_AR NAMES 
${_CMAKE_TOOLCHAIN_PREFIX}ar${_CMAKE_TOOLCHAIN_SUFFIX})

Additionally there is a second issue as cmake fails to add the paths to 
the CMAKE_C*_COMPILER variables in CMakeDetermineC*Compiler modules and 
thereby couldn't set the _CMAKE_TOOLCHAIN_LOCATION. But this is a issue 
of oe as we should pass the full path of the compilers.



      reply	other threads:[~2013-07-18  9:46 UTC|newest]

Thread overview: 7+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2013-07-17  8:40 [PATCH] cmake.bbclass: set ar/ld/nm/objcopy/objdump/ranlib/strip native paths Jonathan Liu
2013-07-17  9:13 ` Stefan Herbrechtsmeier
2013-07-17  9:49   ` Jonathan Liu
2013-07-17 10:57     ` Jonathan Liu
2013-07-17 12:12       ` Stefan Herbrechtsmeier
2013-07-18  0:47         ` Jonathan Liu
2013-07-18  9:46           ` Stefan Herbrechtsmeier [this message]

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=51E7B962.4050108@herbrechtsmeier.net \
    --to=stefan@herbrechtsmeier.net \
    --cc=net147@gmail.com \
    --cc=openembedded-core@lists.openembedded.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox