Build in msys2 shell fails

[Gene Smith]

On 08/19/2016 09:54 PM, Gene Smith wrote:
> On 8/19/2016 8:47 PM, Gene Smith wrote:
>> Installed msys2 on window 10 as recommended. The various "pacman" steps
>> were not quite as straight forward as described on the msys2 website.
>> However, I think I got it installed OK and it runs and all packages are
>> up to date (including the additional packages required by RSB). Also
>> installed the windows version of python2.
>>
>> However, when I run (in MSYS shell) a build with the same command line
>> as on cygwin or linux but with the added options --without-python and
>> --jobs=none, I immediately see this:
>>
>> error: failed to load Windows host support
>>
>> This is coming from builder file .../sb/options.py and indicates that it
>> thinks the "os.name" is "nt".
>>
>> I tried installing and using python3 and see the same thing.
>>
>> Also, running a MINGW64 shell does the same thing with both python
>> versions. (No path entries to cygwin code are present.)
>>
>> Am I doing something wrong? (Haven't tried the msys python package which
>> the documentation says doesn't work with RSB.)
>>
>> -gene
>
> Don't know if this will go all the way, but after doing this the build
> has now started. Building binutils now...:
>
> gene at hplt MINGW64
> /c/Users/gene/development/rtems/src/rtems-source-builder/rtems
> $ pacman -S  mingw-w64-x86_64-python2
>
> -gene

Well, it it ran for 55 minutes and got well into building gcc-newlib and 
then failed for some reason (maybe because using the wrong python?). 
Anyhow here is the top and bottom of the generated and very long log 
file (the error report file referenced contains the same info):

RTEMS Source Builder - Set Builder, 4.12 (cac72a2aea71 modified)
  Command Line: ../source-builder/sb-set-builder --log=l-arm_eabihf.txt 
--prefix=C:/Users/gene/development/bare --target=arm-eabi 
--without-python --jobs=none lang/gcc491
  Python: 2.7.11 (default, Mar  4 2016, 08:42:36)  [GCC 5.3.0 64 bit 
(AMD64)]
Build Set: lang/gcc491
config: devel/expat-2.1.0-1.cfg
package: expat-2.1.0-x86_64-w64-mingw32-1
script:  1: #!sh
script:  2: # ___build_pre in as set up in defaults.py
:
:
:
../../gcc-4.9.1/gcc/config/elfos.h:170:24: warning: invalid suffix on 
literal; C++11 requires a space between literal and string macro 
[-Wliteral-suffix]
        fprintf ((FILE), ","HOST_WIDE_INT_PRINT_UNSIGNED",%u\n",  \
                         ^
In file included from ../../gcc-4.9.1/gcc/cp/except.c:1013:0:
cfns.gperf: In function 'const char* libc_name_p(const char*, unsigned 
int)':
cfns.gperf:101:1: error: 'const char* libc_name_p(const char*, unsigned 
int)' redeclared inline with 'gnu_inline' attribute
cfns.gperf:26:14: note: 'const char* libc_name_p(const char*, unsigned 
int)' previously declared here
cfns.gperf: At global scope:
cfns.gperf:26:14: warning: inline function 'const char* 
libc_name_p(const char*, unsigned int)' used but never defined
make[2]: *** [Makefile:1058: cp/except.o] Error 1
make[2]: Leaving directory 
'/C/Users/gene/development/rtems/src/rtems-source-builder/bare/build/aeg4n2xwm1/build/gcc'
make[1]: *** [Makefile:3983: all-gcc] Error 2
make[1]: Leaving directory 
'/C/Users/gene/development/rtems/src/rtems-source-builder/bare/build/aeg4n2xwm1/build'
make: *** [Makefile:871: all] Error 2
shell cmd failed: sh -ex 
/C/Users/gene/development/rtems/src/rtems-source-builder/bare/build/aeg4n2xwm1/doit
error: building aeg4n2xwm1
   See error report: 
rsb-report-arm-eabi-gcc-4.9.1-newlib-2.1.0-x86_64-w64-mingw32-1.txt
Build Set: Time 0:55:00.755452