On Mar 30, 2008, at 5:36 PM, Jordan Breeding wrote:
On Mar 30, 2008, at 12:57, Pierce T. Wetter III wrote:
On Mar 29, 2008, at 1:02 PM, Laurent Sansonetti wrote:
Hi Pierce,
The RDoc generation is a little bit broken in the testing branch, due to missing write barriers in the string class. This will be fixed in 0.2, and might already be fixed in trunk (though I didn't test it yet).
Now the error you get doesn't seem to be a GC problem, but related to something else. Could you remove .ext/rdoc or checkout a clean testing snapshot (or trunk) and see if you reproduce the same error message again?
Removing gc-stub.o make that part work, I get a different error now with the rdoc part after removing .ext/rdoc, but if its known to be broken, that's fine.
Pierce
I am going to guess (because it was the problem I was having before) that you had most likely built with one set of architectures, then did a clean, and then tried to build with a different set of architectures. The problem was that the Makefile system wasn't removing gc-stub.o when cleaning. It should be fixed on trunk so the next time that trunk gets moved to testing it should be fixed.
Yeah, I think the problem was that I did the configure wrong, then read the correct line in the read me and started over after a make clean. Pierce