First, I like Intel's compiler, 'cause it really uses the power of two
my CPUs.
Second, I had some problems dynamically linking my DLLs to cygwin ruby
interpreter.
I don't like cygwin, because it's ambidgious: NT already has a POSIX
compatibility layer.
Last, I do not need any platform independence, 'cause I use Ruby on my
machine only.

# -----Original Message-----
# From: Yuri Leikind [mailto:YuriLeikind / scnsoft.com] 
# Sent: 11 ノタホム 2001 ヌ. 16:00
# To: Aleksei Guzev
# Subject: Re: [ruby-talk:16386] Ruby source
# 
# 
# 靤夫都
# 
# 靉侑和嬖 佻cygwin - ヒマヘミノフノメユナヤモム ツナレ
# ミメマツフナヘ -
# 
# ./configure && make && make install
# 
# cygwin 料 徨冕品 佻崛厠拱 - http://sources.redhat.com/cygwin/
# 
# I like Ruby, too
# 
# メユモモヒマムレル゛ホルネ メユツノメマノトマラ モヤチホマラノヤモム ツマフリロナ ! ネマメマロマ!
# 
# > I've compiled 1.6.4 for mswin32 almost succefully. Microsoft's 
# > compiler passed with some warnings. Intel's icl 4.5 issued a lot of 
# > errors in parse.y, because of the following fragment:
# >
# > #if SIZEOF_LONG_LONG > 0
# > typedef unsigned long long stack_type;
# > #elif SIZEOF___INT64 > 0
# > typedef unsigned __int64 stack_type;
# > #else
# > typedef unsigned long stack_type;
# > #endif
# >
# > I've removed the fragment except "typedef unsigned long 
# stack_type;" ( 
# > since I do want the stack be unsigned long :).
# >
# > I'm disappointed with a lot of warnings wich I prefer to call 
# > "errors": "the function must return a value". Intel's 
# compiler could 
# > not vectorize most of loops :(
# >
# > But... I like Ruby!
# 
# -- 
# Best regards,
# Yuri Leikind
# 
# Random signature:
# Women are wiser than men because they know less and understand more.
# 		-- Stephens
#