13:19 <odigity> (Ignore rubygems-update gem for a minute.) rubygems is a package in the ruby stdlib, and stdlib packages don't get loaded until you require them, right? Yet when I start with an empty ruby script and call require, it seems to be using the overridden version from the rubygems package by default. How is that possible?

13:19 <odigity> (as opposed to the default Kernel::require, which is not gem-aware)

13:24 <jhass> odigity: rubygems is implicitly required since Ruby 1.9, you can turn that off with ruby --disable-gems

13:24 <odigity> jhass, I see. Are there other stdlib packages that are always implicitly loaded?

13:24 <jhass> not that I'm aware of

13:25 <odigity> I just started IRB and ran `puts $LOADED_FEATURES.sort`. Definitely see rubygems there. :)

13:25 <jhass> you can see what's loaded with ruby -e 'puts $LOADED_FEATURES'

13:25 <jhass> ah, too slow :P

13:26 <jhass> so monitor apparently, probably as a dependency of rubygems

13:27 <odigity> just did it: https://gist.github.com/odigity/767d3a5ab136b4845a7f

13:27 <odigity> I ran it your way this time, to get rid of the IRB files

13:35 <jhass> ruby --disable-gems -e 'puts $LOADED_FEATURES' to see what's pulled in by rubygems

16:35 <pipework> Has anyone done work on making rubygems support gems with unicode names?

16:36 <jhass> mmh, is NTFS unicode by now?

16:37 <imperator> pipework, i can't remember if using non-ascii names causes problems or not

16:37 <imperator> jhass, not sure what you mean

16:38 <jhass> I could imagine some FSes will have trouble naming a directory ☃

16:38 <imperator> it's supported multibyte encodings since win2k

16:38 <imperator> the *shell* might have issues

16:38 <imperator> but not the OS

16:40 <imperator> personally, i think using non-ascii chars is a bad idea

16:41 <jhass> well, for what reason though if not filenames/paths?

16:44 <imperator> we already have one caps issue - Platform vs platform

16:45 <pipework> imperator: IT SEEMS TO.

16:45 <pipework> Oops, sorry for capslokc

16:48 <imperator> jhass, now imagine foo vs foó etc

16:49 <pipework> jhass: Gem names.

16:49 <pipework> Though, admittedly, most gem names show up in their require path.

16:49 <jhass> mmmh, yeah, would at least need to normalize according to https://en.wikipedia.org/wiki/Unicode_equivalence#Normalization

16:49 <jhass> pipework: whose don't?

16:49 <pipework> jhass: assholes' gems don't. :D

16:50 <jhass> pipework: not talking about lib/foo but the gems/foo part

16:50 <jhass> well, foo-0.1.0

16:50 <pipework> jhass: Ah, I see what you mean now.

16:51 <pipework> imperator: Well, ruby seems to handle a whole host of characters just fine when they're not ascii and you're changing their case.

16:52 <imperator> and now add in case insensitive filesytems :)

16:52 <pipework> I think it'd be reasonable to either only use characters which work or to be able to provide a mapping from one character to the intended 'upper case' in your gemspec.

16:53 <imperator> i can't wait for the cjkv gem names!