-
Notifications
You must be signed in to change notification settings - Fork 17
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Provide ruby.exe so it's a drop-in replacement for MRI #29
Comments
Having a 64-bit jruby.exe launcher would also be valuable: |
@headius This became a problem again, see ruby/setup-ruby#18 (comment) I'll copy the comments here for convenience:
Could we have a |
Here are the executables provided on MRI built by RubyInstaller2:
So having a |
Is this only affecting Windows? Most of the time we do not want to depend on the native launcher when the bash script works. If we could get rid of the native launcher we would, but it's needed on a few platforms (Darwin, Solaris) that don't support putting scripts into shebang lines. |
FWIW I would be fine with JRuby publishing |
Yes, it's only on Windows. The Bash script should also work with Bash on Windows, but currently it doesn't deal with quotes correctly. My current workaround is copying the jruby Bash script to ruby, and adding ruby.bat but it's not perfect and it feels hacky I have to mess with the launchers manually: Could we maybe fix whatever in jruby.exe assumes the executable to be named |
A trivial fix would be to have the launcher build and install both Someone's going to have to get the launcher updated to support our |
Another reason is the latest jruby snapshot currently generates this for
Which is https://github.com/ruby/jruby-dev-builder/runs/448860418 |
We do not generate those stubs... it's almost certainly something that has changed in RubyGems. |
This is rubygems issue rubygems/rubygems#2873 (a 3.0 specific regression caused by rubygems/rubygems#2119). It was fixed by rubygems/rubygems#2876, which is present in the latest rubygems but it was never backported to rubygems 3.0 (although it's been scheduled for backport to rubygems 3.0.7). @hsbt Do you still intend to release rubygems 3.0.7 including rubygems/rubygems#2876? |
If we can get RubyGems 3.0.7 I can update it for JRuby 9.2.10, due out VERY SOON. |
I will release RG 3.0.7 in this week. |
Thanks for releasing so quickly @hsbt! Unfortunately there's a regression in 3.0.7, could you have a look at rubygems/rubygems#3139? |
Also released 3.0.8 https://blog.rubygems.org/2020/02/19/3.0.8-released.html @deivid-rodriguez Thanks for ping me. |
Thanks for the quick turnaround 💜! Unfortunately I don't think it made it into the jruby release, but it's good to have this fixed on the 3.x line :) |
Since we are spinning 9.2.11 I will look at 3.0.8 and check that there's no other regressions. |
FWIW, I pushed a quick fix for that issue to JRuby master, and looks like it made it in 9.2.10: jruby/jruby@4f5b1d6. Still would be better to use the fix from upstream of course. |
See jruby/jruby#6042 (comment)
Currently it's not obvious how to make
ruby ...
work on Windows.On Unix this is usually done easily with
ln -s jruby ruby
, but on Windows the only way I could find is to create aruby.bat
with:(I also tried
mklink
on Windows, that doesn't work)It would be nice if this could be simplified and e.g. JRuby already ships with that
ruby.bat
andruby
files.gem
,rake
, etc already conflict with MRI binaries so the only way to use a Ruby is to add it to PATH anyway. Not having something forruby ...
means there is a the danger to use MRI for that one, but not for other commands such asgem
andrake
, which is never intended.The text was updated successfully, but these errors were encountered: