From: Rene Rivera (grafik666_at_[hidden])
Date: 2003-06-27 17:11:31
[2003-06-27] Ian Berthon wrote:
>Vladimir Prus wrote:
>>> set "PATH=;%PATH%"
>>> "bcc32" -v -q -v -tWD -v -tWC -tWR -tWC -WM- -L""
>>Hmm.. this -L"" is very strange. It should include the path to
>>libboost_filesystem.lib. And the code you've previously posted did include
>>proper <search> property. Are you absolutely sure you don't pick old
>>If yes, could you please archive the test project and send it to me. I'll
>>to reproduce with the CVS state.
>As a test I added an echo line to the 'link bind LIBRARIES' action at
>the bottom of the file and that got executed, so I'm fairly sure that
>I'm using the new Borland.Jam file. If you look at my original post, the
>empty -L switch was being passed to the compiler/linker then too BUT
>there was also an additional text-action which was writing a -L line
>into the response file. The attached test.zip file contains my Jamfile
>and the main.cpp file which is all that you should need to reproduce.
>Hope this helps,
It would have if the attachment was actually here ;-)
No worries though... I happen to have run into the same problem as you did
with the way I setup the borland toolset, with the --tool=borland option I
happen to be working on.
The problem was that the library paths for the toolset got set to "" empty
string values which translated to the -L"".
So I've done some changes to the borland.jam tool that might address your
problem. Could you try it out?
-- grafik - Don't Assume Anything
-- rrivera (at) acm.org - grafik (at) redshift-software.com
-- 102708583 (at) icq
Boost-Build list run by bdawes at acm.org, david.abrahams at rcn.com, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk