Re: Building GMT - could somebody verify this problem?
On Fri, Feb 08, 2002 at 07:41:53PM +0100, Torsten Landschoff wrote:
> Fuck. Does not build anymore, the configure test for a bad sincos
> function fails. Could somebody check if that is the case on his
> system as well? Because, what is most interesting: If I run exactly
> the same test with exactly the same compiler options manually it
> works just fine.
I took a quick look at it seems that the configure script manages to
build a "conftest" binary that has:
RPATH $(libdir)
(Literally "$(libdir)", not whatever you would expect libdir to be.
It must be a quoting error.)
I'm guessing that the dynamic linker gets confused by this somehow
and gets stuck. strace shows it cpu cycling early in the dl startup
code.
Ah, I found it... the configure.in script sets LDFLAGS around lines 270-300,
and it uses single quotes where it should use double quotes.
Since it's a Debian package, you probably want to smack it on the head
until it stops setting rpath, anyway :) But changing the quotes should
work.
Ugh... a closer look shows that the configure.in is riddled with this
problem, it might need a careful search-and-replace.
--
Richard Braakman
Will write free software for money.
See http://www.xs4all.nl/~dark/resume.html
Reply to: