Le Wed, 08 Apr 2009 08:16:14 +0200, Tetetest <tetetest_at_rambler.ru> a écrit:

> Hello Sukender,
>
> Wednesday, April 8, 2009, 2:08:19 AM, you wrote:
>
> S> Le Tue, 07 Apr 2009 23:57:14 +0200, Daniel Stenberg <daniel_at_haxx.se> a écrit:
>
>>> On Tue, 7 Apr 2009, Sukender wrote:
>>>
>>>> BTW, Do you think this could be wise to automatically define -DLIBxxx for
>>>> libxxx test, as I did in the CMake script?
>>>
>>> It sounds like a good idea, since quite clearly a bunch of those tests need
>>> them and thus it'll simplify things a little.
>>>
>
> S> Good. Can anyone do this, as I'm a total newbie to autotools? :)
>
> S> I fixed a few things and added Makefile.inc for tests. There are 3 warnings remaining on my machine; maybe you would fix them:
> S> http://www.cdash.org/CDashPublic/viewBuildError.php?type=1&buildid=17667
>
> S> I guess Bill still has errors with his build, though.
>
> Just tried to build on Windows (MSVS 2008, nmake):
>
> Linking C executable ..\lib558.exe
> lib558.c.obj : error LNK2019: unresolved external symbol _Curl_hash_destroy referenced in function _test
> lib558.c.obj : error LNK2019: unresolved external symbol _Curl_mk_dnscache referenced in function _test
> ..\lib558.exe : fatal error LNK1120: 2 unresolved externals
> LINK Pass 1 failed. with 2
> NMAKE : fatal error U1077: '"C:\Program Files\CMake 2.6\bin\cmake.exe"' : return code '0xffffffff'
> Stop.
> NMAKE : fatal error U1077: '"C:\Program Files\Microsoft Visual Studio 9.0\VC\BIN\nmake.exe"' : return code '0x2'
> Stop.
> NMAKE : fatal error U1077: '"C:\Program Files\Microsoft Visual Studio 9.0\VC\BIN\nmake.exe"' : return code '0x2'
> Stop.
>
> Looks funny because lots of other libXXX.exe targets linked ok.
>
> As for warnings: I can see only the warning about "_beginthreadex"
> (apparently, the compiler fails to find the prototype for this
> function).
>
> The proto of this function is in process.h file, so first of all
> make sure that HAVE_PROCESS_H is defined. Second, this function is
> only available in multithread builds (_MT is defined, /MD or /MT
> compiler switch supplied). CMake does supply /MD switch to compiler (I
> have just checked this), so the reason must be the absense of
> HAVE_PROCESS_H.
>
> By the way, HAVE_PROCESS_H is missing not only from CMake buildchain;
> there are no traces of it in auto* files, either.

Hi Tetetest,

I've added this to the ToDo list... (Root CMakeLists.txt - I'll commit later)

BTW, it you wish to submit your build output (errors/warnings), I suggest you turn on the "BUILD_DASHBOARD_REPORTS" option, and then clean your build, and then build the "Experimental" target. Your build errors will then be visible at http://www.cdash.org/CDashPublic/index.php?project=CURL . Moreover, that will allow you to see other's errors/warnings.

Thanks for the report and the clues about how to fix the problem.

-- 
Sukender
PVLE - Lightweight cross-platform game engine - http://pvle.sourceforge.net/