Re: [Boost-build] really dumb problem trying to use bjam to buildwith gcc compiler

From: Robert Ramey (ramey_at_[hidden])
Date: 2008-08-31 14:15:03

I ran bjam with -d5 got a huge output.  Here is the tail of it:

made+   update 
<p..\..\..\bin.v2\libs\serialization\test\test_const_pass.test\
gcc-3.4.4\debug>test_const_pass.test
made    update  all
...found 3341 targets...
...updating 1757 targets...
MkDir1 ..\..\..\bin.v2

    mkdir "..\..\..\bin.v2"

spawn: No such file or directory
Executing through .bat file
Executing command [cmd.exe] [/Q/C] [C:\Temp\jam3428-00.bat]
$ ls "C:\Temp\jam3428-00.bat"
C:\Temp\jam3428-00.bat
$ cat "C:\Temp\jam3428-00.bat"
mkdir "..\..\..\bin.v2"

Sleuthing around my system led me to the discover that my path
didn't contain an entry to \windows\system32 which contains mkdir.
The path has be reinitialized by installation of some unrelated software.
(what were these people thinking?sheesh!).

I might make a couple of suggestions for bjam.

a) perhaps an entry in "trouble shooting" section of the documentation.
(Though honestly I didn't look there)

b) perhaps severing or diminishing  the dependency between bjam and the path 
variable
by invoking the appropriate os api call directly rather than through the 
shell and
path.  Its very hard to predict what shell might be being used and of course
the path can be just about anything.

Thanks a lot for you help.  It resolved the questoin for me. Of course now
I can't figure out why it works with the msvc compiler. oh well

Robert Ramey

Jurko Gospodnetic wrote:
>   Hi Robert
>
>> When I try to run tests with bjam I get the following:
>>
>> ...
>> boost-test(RUN) "config/no_wchar_t_pass" :
>> "libs/config/test/no_wchar_t_pass.cpp"
>> ...patience...
>> ...found 3341 targets...
>> ...updating 1757 targets...
>> MkDir1 ..\..\..\bin.v2
>> spawn: No such file or directory
>>
>> Does anyone happen to know what I have to fix to make this work?  Is
>> this enough of a hint?
>
>   Hmm, try running with -d2 and see the exact commands executed.
> Perhaps that can give you a hint.
>
>   From this I do not really know where this 'spawn' message comes
> from.
>   Hope this helps.
>
>   Best regards,
>     Jurko Gospodnetic