Segmentation fault: Ruby 1.8.5 Under VC++ express 2005

Hi,
Have anybody tried to compile the 1.8.5 code with VC++ express 2005
SP1 and Windows platform SDK 2003 R2? I have pass the compile. And
when it run miniruby, coredump happens.
---------------->8-----------------------------------------------
miniruby ./win32/mkexports.rb:9: [BUG] Segmentation fault
ruby 1.8.5 (2007-01-31) [i386-mswin32_80]
---------------->8----------------------------------------------

Both the stable version and the snapshot version have this problem. I
am running the compilation on XP SP2.

Any help will be greatly appreciated!

On 2/9/07, zm [email protected] wrote:

am running the compilation on XP SP2.

Any help will be greatly appreciated!

I haven’t seen that, specifically, but I haven’t tried in over a year
to compile with VS2005 (and have even less incentive to do so now that
my primary machine is a Mac). I wouldn’t be surprised if something was
linked in that used a different C runtime, which is the number one
problem with even thinking about using VS2005.

I need to bug the folks at MS again and get them talking with people
who actually care about this problem any longer.

(Curt, if you’re reading this, I’d say that MinGW is your best option
at this point. I’ve heard nothing from them since RubyConf.)

-austin

On Feb 9, 7:53 pm, “zm” [email protected] wrote:

am running the compilation on XP SP2.

Any help will be greatly appreciated!

I have manage to “bypass” the error. The problem seems like whenever
the miniruby trying to call a “popen” kind of things it will die.
The 9th line of the mkexports is actually
IO.foreach("|dumpbin -symbols " + objs.join(’ ')) do |l|

I generate a static file with the command line of dumpbin, and let
IO.foreach read this file, then the compilation go through.

Again when do nmake test, it will fail at the execute quote ...

Has anybody got the same issue here?

On 2/9/07, Austin Z. [email protected] wrote:

I need to bug the folks at MS again and get them talking with people
who actually care about this problem any longer.

(Curt, if you’re reading this, I’d say that MinGW is your best option
at this point. I’ve heard nothing from them since RubyConf.)

Yeah… I’m here. I was planning to go MinGW less something happenned
along
the way that presented a better option.

Curt