Closed brainsucker closed 12 years ago
I can't work with this. Please send a pull request for this change as well as your suggestion for a FAQ.
Alright, I have to apologize. I was in a hurry and wrote crap. Its actually really line 583. But it's: "(['-o', so_name.inspect] if sane)," which I changed to "(['-o', so_name.inspect])," Not what i wrote before. The thing is that the output parameter is conditional. I don't know if this is intentional or simply a mistake. Although the output parameter is missing, compilation will still work but with the standard output destination which is on windows "a.exe" and of course not what RubyInline is looking for. As I'm not really confident neither with RubyInline nor with Github I won't be able to provide a decent pull request. I just felt that this should mentioned somewhere as I didn't came across this issue or the solution anywhere and not everybody should have to investigate it himself.
I'm using RubyInstaller 1.9.3p125 at Windows 7 x64 just for the sake of completeness.
No diff... Seriously... Come on.
No real problem description, No diff. I can't work with this. It's been 20 days with no response. Closing.
Besides requiring devkit i had to change at line 583 in inline.rb this "("-L#{RbConfig::CONFIG['libdir']}" if sane)" to that ("-L#{RbConfig::CONFIG['libdir']}"). It kept naming the output file in "...ruby_inline\ruby-1.9.1" "a.exe" (standard output if no output destination is provided) until I did so. Therefore the .so RubyInline looked for wasn't found. I don't know if this is a general solution or if it just works for me. But perhaps someone can look into this. And that devkit should be requiered in combination with RubyInstaller should perhaps also be mentioned somewhere in a faq just for ease.