exception traceback sometimes slow

[ocean-city]

BPO	4503
Nosy	@amauryfa, @pitrou

^{Note: these values reflect the state of the issue at the time it was migrated and might not reflect the current state.}

Show more details

GitHub fields:

assignee = None
closed_at = <Date 2009-03-13.23:56:38.042>
created_at = <Date 2008-12-03.10:31:50.382>
labels = []
title = 'exception traceback sometimes slow'
updated_at = <Date 2009-03-13.23:56:38.041>
user = 'https://bugs.python.org/ocean-city'

bugs.python.org fields:

activity = <Date 2009-03-13.23:56:38.041>
actor = 'pitrou'
assignee = 'none'
closed = True
closed_date = <Date 2009-03-13.23:56:38.042>
closer = 'pitrou'
components = []
creation = <Date 2008-12-03.10:31:50.382>
creator = 'ocean-city'
dependencies = []
files = []
hgrepos = []
issue_num = 4503
keywords = []
message_count = 6.0
messages = ['76823', '76824', '76826', '76827', '76836', '83548']
nosy_count = 3.0
nosy_names = ['amaury.forgeotdarc', 'pitrou', 'ocean-city']
pr_nums = []
priority = 'normal'
resolution = 'fixed'
stage = None
status = 'closed'
superseder = None
type = None
url = 'https://bugs.python.org/issue4503'
versions = ['Python 3.0']

[ocean-city]

Please try this code.

import zipfile
zipfile.PyZipFile("foobar")

I'm not sure this is problematic or not.

[pitrou]

What do you mean by "slow"? I can't reproduce under Linux.

[ocean-city]

Traceback (most recent call last):
  File "a.py", line 4, in <module>
    zipfile.PyZipFile("foobar")
  File "e:\python-dev\py3k\lib\zipfile.py", line 683, in __init__
    self.fp = io.open(file, modeDict[mode])
  File "e:\python-dev\py3k\lib\io.py", line 222, in open
    closefd)
  File "e:\python-dev\py3k\lib\io.py", line 615, in __init__
    _fileio._FileIO.__init__(self, name, mode, closefd)
IOError: [Errno 2] No such file or directory: 'foobar'
[33393 refs]

For example, some delay after

File "e:\python-dev\py3k\lib\zipfile.py", line 683, in __init__

this line. I don't know why, but maybe because current io module is not
written in C.

[amauryfa]

Is "e:" a network drive? This could explain the delay.

But it's true that the code that displays the exception has to call
readline() 683 times, and this code is written in python. There is a
project to rewrite io.py in C...

Also, you are using a debug build. Is the slowdown still noticeable with
a release build?

[ocean-city]

No, e: is not netword drive. It's another partition on hard drive. But
my machine is pretty slow.

Also, you are using a debug build. Is the slowdown still noticeable with
a release build?

I tried release build(VC6), and was much faster.

[pitrou]

I suppose we can close this now, it should be solved with the IO-C
rewrite in 3.1, and won't be backported to 3.0.