I use a ruby gem that depends on tmail. tmail has native extensions. I
want to use the gem with jruby, so I thought I’d replace tmail, as and
where. However…
Googling around, I notice that jruby has a fix against tmail, 3116.
So, is it possible to install tmail for jruby, and if so, how do you do
it?
I use a ruby gem that depends on tmail. tmail has native extensions. I
want to use the gem with jruby, so I thought I’d replace tmail, as and
where. However…
Googling around, I notice that jruby has a fix against tmail, 3116.
So, is it possible to install tmail for jruby, and if so, how do you do
it?
We may only have fixed a bug relating to Rails that referenced tmail. If
it has a native extension we definitely wouldn’t support it. Can you
figure out what it’s using native code for?
We may only have fixed a bug relating to Rails that referenced tmail. If
it has a native extension we definitely wouldn’t support it. Can you
figure out what it’s using native code for?
After two glasses of good rioja blanco at 22:30?
It’s the usual C hieroglyphics. It’s parsing (it refers to scanning)
something - I guess a bag of bytes representing an email - with funky
refs to Japanese strings and ISO2022 (also Japanese?). It doesn’t look
very scary, about 600 lines. It basically seems to build and return a
couple of ruby objects.
However, I also spotted part of the build refs ENV[‘NORUBYEXT’], which
seems to allow a ‘echo Native extensions will be omitted’ build. Working
or red herring?
I don’t know anything about the Ruby/C interface, especially once you
get
down to rb_define_ funcs, so it would take me a while to get into the
details.
It’s damn ugly, though.
There’s a bunch of tests too, so it might be a candidate to de-C. Would
that be of interest?
There’s a bunch of tests too, so it might be a candidate to de-C. Would
that be of interest?
Yeah, I suppose we could look for references to it and first check if it
works without the C bits loaded, but if it seems like they’re needed or
helpful, it would definitely be a good project And if it’s not much
code, it may be trivial to port.
There’s a bunch of tests too, so it might be a candidate to de-C. Would
that be of interest?
Yeah, I suppose we could look for references to it and first check if it
works without the C bits loaded, but if it seems like they’re needed or
helpful, it would definitely be a good project And if it’s not much
code, it may be trivial to port.