too many cat zombie processes

[goog]

the code is :

java = local["java"]
def seg(text):
    seged = (cat << text |java["-mx700m","-cp","/home/drill/segment/stanford-ner.jar:",
                               "edu.stanford.nlp.ie.NERServer","-port","9191","-client"])()
    seged = seged.split('\n',1)[1].strip()
    seged = seged.encode('utf8')
    return seged

I have thousands of short text to segment, so the seg function will be involved thousand times,but I found that there are too many cat processes in the current OS(ubuntu). how to solve it ?

[AndydeCleyre]

What do you mean when you say that there are too many processes -- too many for what? As in, so many that ___.

Maybe you don't need to use cat at all. Can you write in English what the purpose of this code is? seg takes a string text -- are you using cat to write that to a file, and your java command needs a file as input?

[goog]

it seems the java class needs a file as input.
the code snippet is a client handler, which involves a java class to handle the text, but the class will be a interactive shell and need to input a string as input, I don't want to write and read files repeatedly. before the code is finished there will be 2181 zombie processes

[goog]

does it exist a close function or smth like GC ? @tomerfiliba

[tomerfiliba]

first of all, why do you need to use cat to pass text to the java process? why not just:

(java["-mx700m","-cp","/home/drill/segment/stanford-ner.jar:", 
    "edu.stanford.nlp.ie.NERServer", "-port","9191","-client"] << text)()

that way there's second process involved. but you did uncover a bug, whereby the left hand side of the pipe was not waited on. i'm now pushing a fix for that. thanks for reporting.

but again, in your case, you most certainly don't need to use cat

[goog]

yes, u r right. and I referenced input-output-redirection lol.