At 18:25 08/08/27, U.Nakamura wrote:

>Maybe the ruby interpreters running "make test" and
>"cd test/ruby; ruby test_io.rb" are different.
>
>I guess that after "make install" the result will be
>same.

Many thanks for your advice.

This would explain some of the behavior, so I checked it.
First of all, "make test" uses miniruby, so there is a
difference there. But then I also discovered that "ruby -v"
returns "ruby 1.9.0 (2008-08-27 revision 15293) [i386-cygwin]"
whereas "svn up" says "At revision 18880". The file revision.h
doesn't seem to have been updated since January, and indeed
contains "#define RUBY_REVISION 15293". Any idea what went
wrong here?

Regards,    Martin.



>In message "[ruby-core:18396] problems with test_io.rb on cygwin"
>    on Aug.27,2008 18:14:28, <duerst / it.aoyama.ac.jp> wrote:
>> If I run "make test", there is a stall at test_io.rb
>> (with 6 dots output, i.e. working on the 7th test),
>> where enormous amounts of processing time are being burned.
>> 
>> If I use "cd test/ruby; ruby test_io.rb", I get something
>> similar, but with 11 dots output, i.e. working on the 12th
>> test, and only after some minutes 'cold' wait, i.e. where
>> no processing time is being burned.



#-#-#  Martin J. Du"rst, Assoc. Professor, Aoyama Gakuin University
#-#-#  http://www.sw.it.aoyama.ac.jp       mailto:duerst / it.aoyama.ac.jp