There must be a difference between these other folks's environments and mine,
given the different behavior.
What I find so odd is that it hangs when invoked via clojure.java.shell/sh from
within the REPL, but returns immediately with another prompt when run from
bash. I've even copied the xdg-open on the system exhibiting this problem,
because it is a shell script, and added debug statements to it to print out
environment variables. They are nearly identical when invoked from bash vs.
when invoked from in the REPL, except for the value of LD_LIBRARY_PATH. I
thought that might cause the different behavior, but after tweaking things to
make them the same I still saw the hang from the REPL.
I'll try creating another user on that system to see if it happens there, but
if most people don't see this behavior, that is a good thing for when/if the
patch for CLJ-896 is committed, which is the reason I was testing this and
finding the odd behavior.
http://dev.clojure.org/jira/browse/CLJ-896
Thanks,
Andy
On Nov 10, 2012, at 12:26 PM, Aaron Cohen wrote:
> I'm pretty sure xdg-open ends up being a thin wrapper that delegates
> to your desktop environment's open program (gnome-open, kde-open,
> xce-open, etc)
> so probably that is where the difference lies.
>
> --Aaron
>
> On Nov 10, 2012, at 12:52 AM, Michael Gardner <gardne...@gmail.com> wrote:
>
>> I can't duplicate your results on my Debian wheezy box (Clojure 1.4, openjdk
>> 1.6.0_24, xdg-open 1.1.0 rc1); the call to (sh "xdg-open" …) returns
>> immediately. The only possibilities I can think of are that your version of
>> xdg-open blocks when run from a non-interactive shell, or that your version
>> of Java somehow handles subprocesses differently.
>>
>> The first possibility should be easy to test. And you should be able to test
>> the second by writing your own shell script that forks and then exits (at
>> least, I assume that's what xdg-open is doing).
>>
>> On Nov 9, 2012, at 19:26 , Andy Fingerhut <andy.finger...@gmail.com> wrote:
>>
>>> This issue may be specific to Linux, or even to a particular version of
>>> Linux that I am using (Ubuntu 11.10 32-bit desktop), although I doubt it
>>> is. If others try this out, I'd be curious to know what your results are,
>>> especially if you know why it is happening, and how it can be fixed.
>>>
>>> First, some behavior from a bash window on Ubuntu 11.10:
>>>
>>> % echo hi
>>> hi
>>> % xdg-open http://www.google.com
>>> %
>>>
>>> After pressing return for the xdg-open command above, my browser opens a
>>> tab to Google's home page. Back in the shell window, the next prompt
>>> appears immediately, even though the browser is still running and open to
>>> that page. That is what I expect to happen.
>>>
>>> However, if I do the following commands inside of a Clojure REPL (tested
>>> Clojure 1.4 and 1.5-beta1 so far):
>>>
>>> % rlwrap java -cp ~/lein/clojure-1.4.0/lib/clojure-1.4.0.jar clojure.main
>>> Clojure 1.4.0
>>> user=> (require '[clojure.java.shell :as sh])
>>> nil
>>> user=> (sh/sh "echo" "hi")
>>> {:exit 0, :out "hi\n", :err ""}
>>> user=> (sh/sh "xdg-open" "http://www.google.com";)
>>>
>>> The invocation of echo returns immediately, printing the result and the
>>> next REPL prompt. But when I invoke xdg-open, while the browser window
>>> appears and goes to the Google home page, back in the bash window I see no
>>> return value and no new REPL prompt.
>>>
>>> If I quit the browser, then back in the bash window I finally see the
>>> return value and a REPL prompt, as shown below:
>>>
>>> {:exit 0, :out "", :err ""}
>>> user=>
>>>
>>>
>>> What I'm wishing would happen is for the return value and REPL prompt to
>>> appear very soon after pressing return when invoking xdg-open with
>>> clojure.java.shell/sh.
>>>
>>> I added some debug print messages to a local copy of clojure.java.shell/sh,
>>> and it is stopping when waiting for the evaluation of either @out or @err
>>> in the final line of the function.
>>>
>>> If I do the same commands above on Mac OS X, with "open" instead of
>>> "xdg-open", it all works as I expect.
>>>
>>> Any clues?
>>>
>>> Thanks,
>>> Andy
>>>
>>> --
>>> You received this message because you are subscribed to the Google
>>> Groups "Clojure" group.
>>> To post to this group, send email to clojure@googlegroups.com
>>> Note that posts from new members are moderated - please be patient with
>>> your first post.
>>> To unsubscribe from this group, send email to
>>> clojure+unsubscr...@googlegroups.com
>>> For more options, visit this group at
>>> http://groups.google.com/group/clojure?hl=en
>>
>> --
>> You received this message because you are subscribed to the Google
>> Groups "Clojure" group.
>> To post to this group, send email to clojure@googlegroups.com
>> Note that posts from new members are moderated - please be patient with your
>> first post.
>> To unsubscribe from this group, send email to
>> clojure+unsubscr...@googlegroups.com
>> For more options, visit this group at
>> http://groups.google.com/group/clojure?hl=en
>
> --
> You received this message because you are subscribed to the Google
> Groups "Clojure" group.
> To post to this group, send email to clojure@googlegroups.com
> Note that posts from new members are moderated - please be patient with your
> first post.
> To unsubscribe from this group, send email to
> clojure+unsubscr...@googlegroups.com
> For more options, visit this group at
> http://groups.google.com/group/clojure?hl=en
--
You received this message because you are subscribed to the Google
Groups "Clojure" group.
To post to this group, send email to clojure@googlegroups.com
Note that posts from new members are moderated - please be patient with your
first post.
To unsubscribe from this group, send email to
clojure+unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/clojure?hl=en
