don't understand popen2

[Martin P. Hellwig]

Hi all,

I was doing some popen2 tests so that I'm more comfortable using it.
I wrote a little python script to help me test that (testia.py):

---------------------------------
someline = raw_input("something:")

if someline == 'test':
print("yup")
else:
print("nope")
---------------------------------

And another little thing that does it's popen2 stuff:

---------------------------------
import popen2

std_out, std_in = popen2.popen2("testia.py")

x=std_out.readline()
print(x)

std_in.writelines("notgood")

x=std_out.readline()
print(x)
---------------------------------

Now what I expected was that I got the return one the first line:
"something:" and on the second "nope", but instead of that I got:

something:
Traceback (most recent call last):
File "F:\coding\pwSync\popen_test\popen_test.py", line 8, in ?
std_in.writelines("notgood")
IOError: [Errno 22] Invalid argument

I played around a bit with flush, write and the order of first writing
and then reading, the best I can get is no error but still not the
expected output. I googled a bit copied some examples that also worked
on my machine, reread the manual and the only conclusion I have is that
I don't even understand what I'm doing wrong.

Would you please be so kind to explain my wrong doing?
(python 2.4 + win32 extensions on XPProSP2)

Thanks in advance!

--
mph

[gry]

Martin P. Hellwig wrote:

Hi all,

I was doing some popen2 tests so that I'm more comfortable using it.
I wrote a little python script to help me test that (testia.py):

---------------------------------
someline = raw_input("something:")

if someline == 'test':
print("yup")
else:
print("nope")
---------------------------------

And another little thing that does it's popen2 stuff:

---------------------------------
import popen2

std_out, std_in = popen2.popen2("testia.py")

x=std_out.readline()
print(x)

std_in.writelines("notgood")

x=std_out.readline()
print(x)
---------------------------------

Now what I expected was that I got the return one the first line:
"something:" and on the second "nope", but instead of that I got:

>>> something:
Traceback (most recent call last):
File "F:\coding\pwSync\popen_test\popen_test.py", line 8, in ?
std_in.writelines("notgood")
IOError: [Errno 22] Invalid argument >>>
I played around a bit with flush, write and the order of first writing
and then reading, the best I can get is no error but still not the
expected output. I googled a bit copied some examples that also worked
on my machine, reread the manual and the only conclusion I have is that
I don't even understand what I'm doing wrong.

Would you please be so kind to explain my wrong doing?
(python 2.4 + win32 extensions on XPProSP2)

help(sys.stdin.writelines)

Help on built-in function writelines:

writelines(...)
writelines(sequence_of_strings) -> None. Write the strings to the
file.

Note that newlines are not added. The sequence can be any iterable
object
producing strings. This is equivalent to calling write() for each
string>

You gave it a single string, not a list(sequence) of strings. Try
something like:
std_in.writelines(["notgood"])

[Donn Cave]

Quoth gr*@ll.mit.edu:
| Martin P. Hellwig wrote:
....
|> import popen2
|>
|> std_out, std_in = popen2.popen2("testia.py")
|>
|> x=std_out.readline()
|> print(x)
|>
|> std_in.writelines("notgood")
|>
|> x=std_out.readline()
|> print(x)
....
|> Traceback (most recent call last):
|> File "F:\coding\pwSync\popen_test\popen_test.py", line 8, in ?
|> std_in.writelines("notgood")
|> IOError: [Errno 22] Invalid argument
|> >>>
....

| You gave it a single string, not a list(sequence) of strings. Try
| something like:
| std_in.writelines(["notgood"])

Did you try it? For me, writelines(string) works fine. Since in
Python, a string is in a sense a sequence of strings, this doesn't
even really contradict the documentation -

| ... The sequence can be any iterable object producing strings.

Anyway, it seems unlikely he would get that INVARG error for this
reason. That's an error from the host operating system, not the
interpreter, and it mostly likely refers to the file descriptor.
Since it works for me, I guess his problem is basically this:

|> (python 2.4 + win32 extensions on XPProSP2)

Donn Cave, do**@drizzle.com

[Martin P. Hellwig]

gr*@ll.mit.edu wrote:
<cut>

You gave it a single string, not a list(sequence) of strings. Try
something like:
std_in.writelines(["notgood"])

I got this output then:

something:
Traceback (most recent call last):
File "F:\coding\pwSync\popen_test\popen_test.py", line 8, in ?
std_in.writelines(["notgood"])
IOError: [Errno 22] Invalid argument something:
Traceback (most recent call last):
File "F:\coding\pwSync\popen_test\popen_test.py", line 8, in ?
std_in.write("notgood")
IOError: [Errno 22] Invalid argument

It seems that it doesn't matter.

--
mph

[Martin P. Hellwig]

Donn Cave wrote:
<cut>

Anyway, it seems unlikely he would get that INVARG error for this
reason. That's an error from the host operating system, not the
interpreter, and it mostly likely refers to the file descriptor.
Since it works for me, I guess his problem is basically this:

|> (python 2.4 + win32 extensions on XPProSP2)

Donn Cave, do**@drizzle.com

Thank you for your suggestion.

I shuffled it a bit around and executed it on my BSD box and there
indeed it works, if I use that exactly program (well not the path var of
course) on NT it has no error but it still not gives the expected
results. So I guess that pipes,std-in, std-out & std-err work that
different on NT then on other POSIX systems.

The lucky thing is that the code where I am exercising for must run on a
BSD system so my immediately problem is dealt with, but I'm still
curious how to get this working on NT.

I posted my new code and the results under my sep.

--
mph

popen_test.py------------------------------------
#! /usr/local/bin/python
import popen2

std_out, std_in = popen2.popen2("F:\coding\pwSync\popen_test\testia. py")
std_in.writelines("test\n")
std_in.flush()
std_in.close()
x=std_out.readlines()
print(x)
-------------------------------------------------

testia.py----------------------------------------
#! /usr/local/bin/python
someline = raw_input("something:")

if someline == 'test':
print("yup")
else:
print("nope")
-------------------------------------------------
results on NT:
F:\coding\pwSync\popen_test>popen_test.py
[]
results on BSD:
%./popen_test.py
['something:yup\n']
%

[Sion Arrowsmith]

Martin P. Hellwig <mh******@xs4all.nl> wrote:

std_out, std_in = popen2.popen2("F:\coding\pwSync\popen_test\testia. py")

^^
Your problem is, I suspect, nothing to do with popen2(), which is
supported by the fact that the only thing other than OS different
between this and your working BSD version is the path:

"F:\coding\pwSync\popen_test\testia.py"

'F:\\coding\\pwSync\\popen_test\testia.py'

Try:
std_out, std_in = popen2.popen2("F:/coding/pwSync/popen_test/testia.py")
or:
std_out, std_in = popen2.popen2("F:\\coding\\pwSync\\popen_test\\tes tia.py")
(and please avoid the abuse of raw strings for Windows paths).

--
\S -- si***@chiark.greenend.org.uk -- http://www.chaos.org.uk/~sion/
___ | "Frankly I have no feelings towards penguins one way or the other"
\X/ | -- Arthur C. Clarke
her nu becomeþ se bera eadward ofdun hlæddre heafdes bæce bump bump bump

[Fredrik Lundh]

Sion Arrowsmith wrote:

"F:\coding\pwSync\popen_test\testia.py" 'F:\\coding\\pwSync\\popen_test\testia.py'

this might make it more obvious that something's not quite right with that
string literal:

print "F:\coding\pwSync\popen_test\testia.py"

F:\coding\pwSync\popen_test estia.py

</F>

[Kent Johnson]

Sion Arrowsmith wrote:

Try:
std_out, std_in = popen2.popen2("F:/coding/pwSync/popen_test/testia.py")
or:
std_out, std_in = popen2.popen2("F:\\coding\\pwSync\\popen_test\\tes tia.py")
(and please avoid the abuse of raw strings for Windows paths).

Why do you consider that abuse of raw strings? It's very easy and
convenient to copy a path from Windows Explorer and paste it directly
into a raw string, no editing required.

Kent

[Sion Arrowsmith]

Kent Johnson <ke**@kentsjohnson.com> wrote:

Sion Arrowsmith wrote:

(and please avoid the abuse of raw strings for Windows paths).

Why do you consider that abuse of raw strings?

I consider it abuse because it's not what they were invented for.
I consider discouraging it to be a good thing in order to reduce
the chances of people being bitten by, for instance:

os.listdir(r"C:\")

File "<stdin>", line 1
os.listdir(r"C:\")
^
SyntaxError: EOL while scanning single-quoted string

and getting all confused on this group.

--
\S -- si***@chiark.greenend.org.uk -- http://www.chaos.org.uk/~sion/
___ | "Frankly I have no feelings towards penguins one way or the other"
\X/ | -- Arthur C. Clarke
her nu becomeþ se bera eadward ofdun hlæddre heafdes bæce bump bump bump

[Kent Johnson]

Sion Arrowsmith wrote:

Kent Johnson <ke**@kentsjohnson.com> wrote:

Sion Arrowsmith wrote:

(and please avoid the abuse of raw strings for Windows paths).

Why do you consider that abuse of raw strings?

I consider it abuse because it's not what they were invented for.
I consider discouraging it to be a good thing in order to reduce
the chances of people being bitten by, for instance:

os.listdir(r"C:\")

File "<stdin>", line 1
os.listdir(r"C:\")
^
SyntaxError: EOL while scanning single-quoted string

and getting all confused on this group.

Judging from the traffic here, people are far more often bitten by *not*
using raw strings for paths containing \ than they are by using them for
paths ending with \. So perhaps we should encourage raw strings for
Windows paths. ;)

Kent