[Boost-build] no "sunpro.jam" when building 1.34 w/SunStudio 12 (C++ 5.9) on Solaris 9/Sparc

From: West, Larry (Larry_West_at_[hidden])
Date: 2007-06-19 00:18:15

This seems like it must be an error on my part, since it shows up on the
first "make": bjam errors out immediately saying it can't find
"sunpro.jam".

I downloaded Boost 1.34, untarred, cd boost_1_34_0, ran
$ ./configure --prefix=/sandbox/3rd/release/sf.net/boost/boost_1_34_0
--exec-prefix=/sandbox/3rd/bin/SunOS_CC/sf.net/boost/boost_1_34_0
--with-toolset=sunpro

(Those directories are non-existent at this point).  This took about 30
seconds and displayed the following:
----------
-n Building Boost.Jam with toolset sunpro...
tools/jam/src/bin.solaris/bjam
-n Detecting Python version...
2.2
-n Detecting Python root...
/opt/sfw
-n Unicode/ICU support for Boost.Regex?...
not found.
Generating Boost.Build configuration in user-config.jam...
Generating Makefile...
----------

The resulting Makefile is very short -- I've attached it and
user-config.jam.

When I run "make" (GNU make 3.81), I get the following output:
----------
./tools/jam/src/bin.solaris/bjam  --user-config=user-config.jam
sunpro.jam: No such file or directory
/sandbox/lwest/3rd/packages/sf.net/boost/boost_1_34_0/tools/build/v2/bui
ld/toolset.jam:27: in toolset.using
rule sunpro.init unknown in module toolset.
/sandbox/lwest/3rd/packages/sf.net/boost/boost_1_34_0/tools/build/v2/bui
ld/project.jam:825: in using
user-config.jam:5: in modules.load
/sandbox/lwest/3rd/packages/sf.net/boost/boost_1_34_0/tools/build/v2/bui
ld-system.jam:138: in load
/sandbox/lwest/3rd/packages/sf.net/boost/boost_1_34_0/tools/build/v2/ker
nel/modules.jam:261: in import
/sandbox/lwest/3rd/packages/sf.net/boost/boost_1_34_0/tools/build/v2/ker
nel/bootstrap.jam:132: in boost-build
/sandbox/lwest/3rd/packages/sf.net/boost/boost_1_34_0/boost-build.jam:9:
in module scope
Not all Boost libraries built properly.
----------

Note that the "sunpro.jam: No such file or directory" message is
actually going to stderr; the rest to stdout.

That's all the real info I have; I've added some output from "bjam -d9"
below, but I think the problem is more likely in the configure step.

I'm willing to try 'most anything or report back any values of interest,
though it might take me a day to get back to this.

Needless to say I'd appreciate any pointers.  My apologies for the
buckets of text and I hope it's easy to truncate them when replying.

Thanks
Larry West
San Diego, California

*** bjam -d9 output ***

As far as I can tell from looking at the voluminous output (253,726
lines) from adding the "-d9" flag to bjam, sunpro.jam is first
referenced by tools/build/v2/kernel/modules.jam almost at the end of
that output:
----------
/sandbox/lwest/3rd/packages/sf.net/boost/boost_1_34_0/tools/build/v2/ker
nel/modules.jam:93:>>>>|>>>>|>>>>|>>> if (1) 
list > 1 <
expand 'filename'
list > filename <
expand '$(module-name).jam'
expand 'module-name'
list > module-name <
expand '.jam'
list > .jam <
get module-name = sunpro
list > sunpro.jam <
expanded to sunpro.jam
/sandbox/lwest/3rd/packages/sf.net/boost/boost_1_34_0/tools/build/v2/ker
nel/modules.jam:95:>>>>|>>>>|>>>>|>>> set filename ?= sunpro.jam
list > sunpro.jam <
set filename = sunpro.jam
expand '.loaded'
list > .loaded <
----------

There are more lines of modules.jam hit after that (abbreviating):
modules.jam:98:>>>>|>>>>|>>>>|>>> set .loaded += sunpro
modules.jam:101:>>>>|>>>>|>>>>|>>> local suppress-test = build-system
modules.jam:104:>>>>|>>>>|>>>>|>>> set .loading += sunpro
modules.jam:107:>>>>|>>>>|>>>>|>>> set .untested += sunpro
modules.jam:110:>>>>|>>>>|>>>>|>>> poke sunpro : __name__ : sunpro
modules.jam:42:>>>>|>>>>|>>>>|>>>>| set __name__ = sunpro
modules.jam:111:>>>>|>>>>|>>>>|>>> poke sunpro : __file__ : sunpro.jam
modules.jam:42:>>>>|>>>>|>>>>|>>>>| set __file__ = sunpro.jam
modules.jam:116:>>>>|>>>>|>>>>|>>> IMPORT modules : no-test-defined :
sunpro : __test__
modules.jam:119:>>>>|>>>>|>>>>|>>> local module-target =
<module@>sunpro.jam
modules.jam:121:>>>>|>>>>|>>>>|>>> local search =
/sandbox/lwest/3rd/packages/sf.net/boost/boost_1_34_0 [...]
modules.jam:122:>>>>|>>>>|>>>>|>>> modules.peek  : BOOST_BUILD_PATH
modules.jam:122:>>>>|>>>>|>>>>|>>> set search ?=  /usr/share/boost-build
/sandbox/lwest/3rd/packages/sf.net/boost/boost_1_34_0/tools/build/v2
[...]

In the last few dozen lines of output, it appears to be searching for
<module@>sunpro.jam through a list of directories starting with
/usr/share/boost-build and then a bunch of sub-directories under my
boost_1_34_0.  This is where the stderr output is mixed in:
----------
expand '$(module-target)'
expand 'module-target'
list > module-target <
get module-target = <module@>sunpsunpro.jam: No such file or directory
ro.jam
list > <module@>sunpro.jam <
----------

And the next lines of modules.jam that are logged are:
modules.jam:123:>>>>|>>>>|>>>>|>>> set SEARCH on <module@>sunpro.jam =
/sandbox/lwest/3rd/packages/sf.net/boost/boost_1_34_0 [...]
modules.jam:124:>>>>|>>>>|>>>>|>>> set BINDRULE on <module@>sunpro.jam =
modules.record-binding
modules.jam:126:>>>>|>>>>|>>>>|>>> include <module@>sunpro.jam
modules.jam:190:>>>>|>>>>|>>>>|>>>>| set sunpro.__binding__ = sunpro.jam
modules.jam:129:>>>>|>>>>|>>>>|>>> RULENAMES sunpro

And then lines 133, 134, 140, 144, and 264 and it begins to wind down.

*** System details ***

$ uname -a
SunOS agdev1.sd.intuit.com 5.9 Generic_117171-07 sun4u sparc
SUNW,Sun-Fire-V210 Solaris

$ CC -V
CC: Sun C++ 5.9 SunOS_sparc 2007/05/03

$ bash --version
GNU bash, version 3.2.0(1)-release (sparc-sun-solaris2.9)
Copyright (C) 2005 Free Software Foundation, Inc.

$ make --version
GNU Make 3.81
Copyright (C) 2006  Free Software Foundation, Inc.
This is free software; see the source for copying conditions.
There is NO warranty; not even for MERCHANTABILITY or FITNESS FOR A
PARTICULAR PURPOSE.

This program built for sparc-sun-solaris2.9

*** End ***