Boost Testing :
From: David Abrahams (dave_at_[hidden])
Date: 2005-05-26 08:51:49
"Victor A. Wagner Jr." <vawjr_at_[hidden]> writes:
> At 08:19 2005-05-25, David Abrahams wrote:
>>"Victor A. Wagner Jr." <vawjr_at_[hidden]> writes:
>> > At 13:18 2005-05-23, Hartmut Kaiser wrote:
>> >>Victor A. Wagner Jr. wrote:
>> > I don't see any change, and I don't understand this "manifest"
>> stuff either.
>>Me neither, but Microsoft assures me that it's neccessary to do it
>>when I build DLLs. I have no idea why it should be failing. The
>>command I was given by Microsoft supposedly extracts the manifest from
>>the DLL and then processes it somehow.
>> > it almost looks like the test framework is attempting to invoke
>> > the ........manifest file directly
>>Hum, yeah it does. Is this Boost.Build v1 or v2?
>>Are you running the build under the NT command prompt or under cygwin?
> regression.py is running the build(s)
>From a regular Windows version of Python, and not from a Cygwin build?
>>Can you show the command that triggers the error above?
> I thought it preceeded it (from the "linker output" at the website):
> mt -manifest
> I'm kinda puzzled about the ;#2
Me too, but Martyn told me I needed it there.
> the actual log from the build is at ftp://rudbek.com/pub/boost as it
> says in the RudbekAssociates link.
But now the errors are gone because of my patch, below. If you undo
this in your local copy you should be able to see the errors again.
No need to do a complete test run.
RCS file: /cvsroot/boost/boost/tools/build/v1/msvc-tools.jam,v
retrieving revision 1.48
retrieving revision 1.49
diff -b -d -u -b -u -r1.48 -r1.49
--- msvc-tools.jam 22 May 2005 10:45:44 -0000 1.48
+++ msvc-tools.jam 25 May 2005 19:18:13 -0000 1.49
@@ -129,7 +129,7 @@
with-command-file vc-Link $(<) : $(sources) $(NEEDLIBS) ;
- if $(target-type) != LIB
+ if $(target-type) = PYD
MANIFEST on $(target) = $(VC_MANIFEST) ;
OUTPUTRESOURCE on $(target) = $(VC_OUTPUTRESOURCE) ;
-- Dave Abrahams Boost Consulting www.boost-consulting.com