These are chat archives for jruby-gradle/jruby-gradle-plugin

6th
Oct 2014
Schalk W. Cronjé
@ysb33r
Oct 06 2014 04:52
Sure
Schalk W. Cronjé
@ysb33r
Oct 06 2014 17:00

@rtyler looked at you #63 merge - it’s good, but I did think of an enhancement we can make.

script ‘rspec’, usePath : true

That will understand that the script is in the path and automatically add the -S. If not set the old behaviour will apply. What do you think?

R. Tyler Croy
@rtyler
Oct 06 2014 17:01
I think that might be a little confusing
because rspec isn't really in the path
it's one of the Ruby binstubs
inside build/tmp/jrubyExec
Schalk W. Cronjé
@ysb33r
Oct 06 2014 17:04
maybe a different property name?
R. Tyler Croy
@rtyler
Oct 06 2014 17:04
sure
Schalk W. Cronjé
@ysb33r
Oct 06 2014 17:06
I sort of think that we need to keep script mandatory, but we need to accommodate the use-case of scripts inside jruby-complete or in the unpacked gems to work without the kludge that we had to do prior to your commit
rtyler @rtyler nods
R. Tyler Croy
@rtyler
Oct 06 2014 17:07
binstub: true would work for me
Schalk W. Cronjé
@ysb33r
Oct 06 2014 17:07
or alternatively if we detect that -S was passed then we don’t enforce a check for a script that exists
mmm… this suggestion of mine might actaully work better for people used to ruby
rtyler @rtyler raises his hand :)
Schalk W. Cronjé
@ysb33r
Oct 06 2014 17:14
Then we can annotate script with @Input as to ensure that changes to the value causes a rebuild. internally in the exec method we can detect whether script is absolute and if it is we can check whether it exists.
R. Tyler Croy
@rtyler
Oct 06 2014 17:20
ah fancy
I'm open to changes that make my life easier
I hit that over the weekend
I was trying to have gradle start up a Pry-based repl
SUCKS
Schalk W. Cronjé
@ysb33r
Oct 06 2014 17:36
Did you want to to take input from the keyboard during execution of a task?
R. Tyler Croy
@rtyler
Oct 06 2014 17:37
yeah, pretty much exactly the bug/problem that guy hit