Jython

Avi,

This enhancement is indeed useful to me as well. Would you please post
the diff of your launch script?
Thanks a lot.

2009/3/12 Avi Flax <avif@...>:
> I'm a Groovy user, and I've found the ability to pass a "cp" option to
> the launch script for the classpath very useful.
>
> I've hacked my local version of the Jython startup script to accept
> "cp", and it works great for me.
>
> I'm wondering if there's any interest in adding this feature to the
> official launch script. I know I'd like it.
>
> If so, I'd try to help make the change happen before the release of
> 2.5, if possible. I can contribute a diff of my launch script, but
> someone would need to vet it, because my changes may have a negative
> impact on things I don't understand.
>
> Thanks,
> Avi
>
> --
> Avi Flax » Lead Technologist » Partner » Arc90 » http://arc90.com

Avi, Cheng,

you could also set the CLASSPATH environment variable before launching jython.

No promises about when, but if it is useful and does not break
anything, i might add it.

best wishes,
Oti.

On Thu, Mar 12, 2009 at 2:56 AM, Zhang Cheng <czhang.mlist@...> wrote:
> Avi,
>
> This enhancement is indeed useful to me as well. Would you please post
> the diff of your launch script?
> Thanks a lot.
>
> 2009/3/12 Avi Flax <avif@...>:
>> I'm a Groovy user, and I've found the ability to pass a "cp" option to
>> the launch script for the classpath very useful.
>>
>> I've hacked my local version of the Jython startup script to accept
>> "cp", and it works great for me.
>>
>> I'm wondering if there's any interest in adding this feature to the
>> official launch script. I know I'd like it.
>>
>> If so, I'd try to help make the change happen before the release of
>> 2.5, if possible. I can contribute a diff of my launch script, but
>> someone would need to vet it, because my changes may have a negative
>> impact on things I don't understand.
>>
>> Thanks,
>> Avi
>>
>> --
>> Avi Flax » Lead Technologist » Partner » Arc90 » http://arc90.com
>
> ------------------------------------------------------------------------------
> Apps built with the Adobe(R) Flex(R) framework and Flex Builder(TM) are
> powering Web 2.0 with engaging, cross-platform capabilities. Quickly and
> easily build your RIAs with Flex Builder, the Eclipse(TM)based development
> software that enables intelligent coding and step-through debugging.
> Download the free 60 day trial. http://p.sf.net/sfu/www-adobe-com
> _______________________________________________
> Jython-users mailing list
> Jython-users@...
> https://lists.sourceforge.net/lists/listinfo/jython-users
>

On Wed, Mar 11, 2009 at 21:56, Zhang Cheng <czhang.mlist@...> wrote:

> This enhancement is indeed useful to me as well. Would you please post
> the diff of your launch script?

Here you go, diff based on a Mac OS X install of 2.5b2:

99a100,106
> # ----- BEGIN Copied from Groovy by Avi Flax -----
> if [ "$1" = "-cp" -o "$1" = "-classpath" ] ; then
>     CP=$CP:$2
>     shift 2
> fi
> # ----- END Copied from Groovy by Avi Flax -----
>
215c222
<   java_args=("${java_args[@]}" -Xbootclasspath/a:"$CP")
---
>   java_args=("${java_args[@]}" -cp "$CP")

Please don't that I don't really understand what Xbootclasspath does,
so while this seems to work for me, YMMV. Hopefully Oti will help me
understand this.

On Thu, Mar 12, 2009 at 04:20, Oti <ohumbel@...> wrote:

> you could also set the CLASSPATH environment variable before launching jython.

True, but I personally find this awkward and kind of annoying. I also
like to add a shebang to some of my scripts so I can execute them
directly, and is most naturally a single line.

I really like to be able to add a line like this to the top of a script:

#! /usr/bin/env jython -cp "lib/*"

> No promises about when, but if it is useful and does not break
> anything, i might add it.

I hope we can make this happen!

Thanks,
Avi

-- 
Avi Flax » Lead Technologist » Partner » Arc90 » http://arc90.com

On Thu, Mar 12, 2009 at 12:06 PM, Avi Flax <avif@...> wrote:
> On Wed, Mar 11, 2009 at 21:56, Zhang Cheng <czhang.mlist@...> wrote:
>
> > This enhancement is indeed useful to me as well. Would you please post
> > the diff of your launch script?
>
> Here you go, diff based on a Mac OS X install of 2.5b2:
>
>  99a100,106
> > # ----- BEGIN Copied from Groovy by Avi Flax -----
> > if [ "$1" = "-cp" -o "$1" = "-classpath" ] ; then
> >     CP=$CP:$2
> >     shift 2
> > fi
> > # ----- END Copied from Groovy by Avi Flax -----

AFAICS, if both CLASSPATH and -cp options are used, the patch will
"merge" both, with -cp taking precedence. This is different from what
java -cp does (i.e completely overriding CLASSPATH)

Is this intentional? How does it compare with what other JVM languages do?

-- 
Leo Soto M.
http://blog.leosoto.com

On Thu, Mar 12, 2009 at 11:55, Leo Soto M. <leo.soto@...> wrote:

> AFAICS, if both CLASSPATH and -cp options are used, the patch will
> "merge" both, with -cp taking precedence. This is different from what
> java -cp does (i.e completely overriding CLASSPATH)
>
> Is this intentional? How does it compare with what other JVM languages do?

No, this wasn't intentional. This is just me hacking around to make
things more convenient for me. I definitely did not, and still don't,
fully understand this domain.

I think that the launch script should behave the same as the java
launcher. I'm hoping Oti will be able to help out with this!

-- 
Avi Flax » Lead Technologist » Partner » Arc90 » http://arc90.com

The problem with adding a -cp command line option is that it conflicts with
the already existing -c (execute single command) option. "-cprint" is a
perfectly valid python command line option (try it with CPython, it doesn't
work with Jython at the moment - this might be a bug). We have the -J
command line option space, would -J-cp be ok instead?

On Thu, Mar 12, 2009 at 4:59 PM, Avi Flax <avif@...> wrote:

> On Thu, Mar 12, 2009 at 11:55, Leo Soto M. <leo.soto@...> wrote:
>
> > AFAICS, if both CLASSPATH and -cp options are used, the patch will
> > "merge" both, with -cp taking precedence. This is different from what
> > java -cp does (i.e completely overriding CLASSPATH)
> >
> > Is this intentional? How does it compare with what other JVM languages
> do?
>
> No, this wasn't intentional. This is just me hacking around to make
> things more convenient for me. I definitely did not, and still don't,
> fully understand this domain.
>
> I think that the launch script should behave the same as the java
> launcher. I'm hoping Oti will be able to help out with this!
>
> --
> Avi Flax » Lead Technologist » Partner » Arc90 » http://arc90.com
>
>
> ------------------------------------------------------------------------------
> Apps built with the Adobe(R) Flex(R) framework and Flex Builder(TM) are
> powering Web 2.0 with engaging, cross-platform capabilities. Quickly and
> easily build your RIAs with Flex Builder, the Eclipse(TM)based development
> software that enables intelligent coding and step-through debugging.
> Download the free 60 day trial. http://p.sf.net/sfu/www-adobe-com
> _______________________________________________
> Jython-users mailing list
> Jython-users@...
> https://lists.sourceforge.net/lists/listinfo/jython-users
>

On Thu, Mar 12, 2009 at 13:32, Tobias Ivarsson <thobes@...> wrote:

> The problem with adding a -cp command line option is that it conflicts with
> the already existing -c (execute single command) option. "-cprint" is a
> perfectly valid python command line option (try it with CPython, it doesn't
> work with Jython at the moment - this might be a bug). We have the -J
> command line option space, would -J-cp be ok instead?

How about we just use "classpath"? "cp" is convenient but "classpath"
is fine, and even better in some ways.

-- 
Avi Flax » Lead Technologist » Partner » Arc90 » http://arc90.com

On Thu, Mar 12, 2009 at 04:20, Oti <ohumbel@...> wrote:


> No promises about when, but if it is useful and does not break
> anything, i might add it.
>

Oti: should I add an issue to the bugtracker to keep track of this possible
enhancement?

Also I don't know if you saw Tobias' point that there might be a conflict
between "cp" and the existing "c" option, because of, for instance,
"-cprint". So I suggested we just name the new option "classpath".

Thanks,
Avi

-- 
Avi Flax » Lead Technologist » Partner » Arc90 » http://arc90.com

On Fri, Mar 20, 2009 at 9:24 AM, Avi Flax <avif@...> wrote:
> On Thu, Mar 12, 2009 at 04:20, Oti <ohumbel@...> wrote:
>
>>
>> No promises about when, but if it is useful and does not break
>> anything, i might add it.
>
> Oti: should I add an issue to the bugtracker to keep track of this possible
> enhancement?
Please do add this to the bug tracker.

> Also I don't know if you saw Tobias' point that there might be a conflict
> between "cp" and the existing "c" option, because of, for instance,
> "-cprint". So I suggested we just name the new option "classpath".
If we do decide to go with classpath, I'd prefer to use the gnu style
--classpath (we already support --help and several other -- style
options) since knowing the difference between -cprint and -classpath
is still slightly annoying from a parsing perspective (it could be
interpreted as -c lasspath).

-Frank

On Fri, Mar 20, 2009 at 2:40 PM, Frank Wierzbicki <fwierzbicki@...:

> On Fri, Mar 20, 2009 at 9:24 AM, Avi Flax <avif@...> wrote:
> > On Thu, Mar 12, 2009 at 04:20, Oti <ohumbel@...> wrote:
> >
> >>
> >> No promises about when, but if it is useful and does not break
> >> anything, i might add it.
> >
> > Oti: should I add an issue to the bugtracker to keep track of this
> possible
> > enhancement?
> Please do add this to the bug tracker.
>
> > Also I don't know if you saw Tobias' point that there might be a conflict
> > between "cp" and the existing "c" option, because of, for instance,
> > "-cprint". So I suggested we just name the new option "classpath".
> If we do decide to go with classpath, I'd prefer to use the gnu style
> --classpath (we already support --help and several other -- style
> options) since knowing the difference between -cprint and -classpath
> is still slightly annoying from a parsing perspective (it could be
> interpreted as -c lasspath).
>

Double dashes solves this nicely. That could even work for --cp, although by
convention double dash arguments should be spelled out words.

/Tobias

On Fri, Mar 20, 2009 at 9:52 AM, Tobias Ivarsson <thobes@...> wrote:
> Double dashes solves this nicely. That could even work for --cp, although by
> convention double dash arguments should be spelled out words.
Yeah, I'd prefer we stick to the whole-word convention for --

-Frank

On Fri, Mar 20, 2009 at 2:24 PM, Avi Flax <avif@...> wrote:
> On Thu, Mar 12, 2009 at 04:20, Oti <ohumbel@...> wrote:
>
>>
>> No promises about when, but if it is useful and does not break
>> anything, i might add it.
>
> Oti: should I add an issue to the bugtracker to keep track of this possible
> enhancement?

Avi, the bug is fine as a reminder - thanks!

> Also I don't know if you saw Tobias' point that there might be a conflict
> between "cp" and the existing "c" option, because of, for instance,
> "-cprint". So I suggested we just name the new option "classpath".

Yes, i saw that, and the following posts of Tobias and Frank as well!

Oti.