##// END OF EJS Templates
Reset the interactive namespace __warningregistry__ before executing code...
Reset the interactive namespace __warningregistry__ before executing code Fixes #6611. Idea: Right now, people often don't see important warnings when running code in IPython, because (to a first approximation) any given warning will only issue once per session. Blink and you'll miss it! This is a very common contributor to confused emails to numpy-discussion. E.g.: In [5]: 1 / my_array_with_random_contents /home/njs/.user-python2.7-64bit-3/bin/ipython:1: RuntimeWarning: divide by zero encountered in divide #!/home/njs/.user-python2.7-64bit-3/bin/python Out[5]: array([ 1.77073316, -2.29765021, -2.01800811, ..., 1.13871243, -1.08302964, -8.6185091 ]) Oo, right, guess I gotta be careful of those zeros -- thanks, numpy, for giving me that warning! A few days later: In [592]: 1 / some_other_array Out[592]: array([ 3.07735763, 0.50769289, 0.83984078, ..., -0.67563917, -0.85736257, -1.36511271]) Oops, it turns out that this array had a zero in it too, and that's going to bite me later. But no warning this time! The effect of this commit is to make it so that warnings triggered by the code in cell 5 do *not* suppress warnings triggered by the code in cell 592. Note that this only applies to warnings triggered *directly* by code entered interactively -- if somepkg.foo() calls anotherpkg.bad_func() which issues a warning, then this warning will still only be displayed once, even if multiple cells call somepkg.foo(). But if cell 5 and cell 592 both call anotherpkg.bad_func() directly, then both will get warnings. (Important exception: if foo() is defined *interactively*, and calls anotherpkg.bad_func(), then every cell that calls foo() will display the warning again. This is unavoidable without fixes to CPython upstream.) Explanation: Python's warning system has some weird quirks. By default, it tries to suppress duplicate warnings, where "duplicate" means the same warning message triggered twice by the same line of code. This requires determining which line of code is responsible for triggering a warning, and this is controlled by the stacklevel= argument to warnings.warn. Basically, though, the idea is that if foo() calls bar() which calls baz() which calls some_deprecated_api(), then baz() will get counted as being "responsible", and the warning system will make a note that the usage of some_deprecated_api() inside baz() has already been warned about and doesn't need to be warned about again. So far so good. To accomplish this, obviously, there has to be a record of somewhere which line this was. You might think that this would be done by recording the filename:linenumber pair in a dict inside the warnings module, or something like that. You would be wrong. What actually happens is that the warnings module will use stack introspection to reach into baz()'s execution environment, create a global (module-level) variable there named __warningregistry__, and then, inside this dictionary, record just the line number. Basically, it assumes that any given module contains only one line 1, only one line 2, etc., so storing the filename is irrelevant. Obviously for interactive code this is totally wrong -- all cells share the same execution environment and global namespace, and they all contain a new line 1. Currently the warnings module treats these as if they were all the same line. In fact they are not the same line; once we have executed a given chunk of code, we will never see those particular lines again. As soon as a given chunk of code finishes executing, its line number labels become meaningless, and the corresponding warning registry entries become meaningless as well. Therefore, with this patch we delete the __warningregistry__ each time we execute a new block of code.

File last commit:

r16665:c08bdde7
r18548:61431d7d
Show More
page.py
360 lines | 12.0 KiB | text/x-python | PythonLexer
# encoding: utf-8
"""
Paging capabilities for IPython.core
Authors:
* Brian Granger
* Fernando Perez
Notes
-----
For now this uses ipapi, so it can't be in IPython.utils. If we can get
rid of that dependency, we could move it there.
-----
"""
#-----------------------------------------------------------------------------
# Copyright (C) 2008-2011 The IPython Development Team
#
# Distributed under the terms of the BSD License. The full license is in
# the file COPYING, distributed as part of this software.
#-----------------------------------------------------------------------------
#-----------------------------------------------------------------------------
# Imports
#-----------------------------------------------------------------------------
from __future__ import print_function
import os
import re
import sys
import tempfile
from io import UnsupportedOperation
from IPython import get_ipython
from IPython.core.error import TryNext
from IPython.utils.data import chop
from IPython.utils import io
from IPython.utils.process import system
from IPython.utils.terminal import get_terminal_size
from IPython.utils import py3compat
#-----------------------------------------------------------------------------
# Classes and functions
#-----------------------------------------------------------------------------
esc_re = re.compile(r"(\x1b[^m]+m)")
def page_dumb(strng, start=0, screen_lines=25):
"""Very dumb 'pager' in Python, for when nothing else works.
Only moves forward, same interface as page(), except for pager_cmd and
mode."""
out_ln = strng.splitlines()[start:]
screens = chop(out_ln,screen_lines-1)
if len(screens) == 1:
print(os.linesep.join(screens[0]), file=io.stdout)
else:
last_escape = ""
for scr in screens[0:-1]:
hunk = os.linesep.join(scr)
print(last_escape + hunk, file=io.stdout)
if not page_more():
return
esc_list = esc_re.findall(hunk)
if len(esc_list) > 0:
last_escape = esc_list[-1]
print(last_escape + os.linesep.join(screens[-1]), file=io.stdout)
def _detect_screen_size(screen_lines_def):
"""Attempt to work out the number of lines on the screen.
This is called by page(). It can raise an error (e.g. when run in the
test suite), so it's separated out so it can easily be called in a try block.
"""
TERM = os.environ.get('TERM',None)
if not((TERM=='xterm' or TERM=='xterm-color') and sys.platform != 'sunos5'):
# curses causes problems on many terminals other than xterm, and
# some termios calls lock up on Sun OS5.
return screen_lines_def
try:
import termios
import curses
except ImportError:
return screen_lines_def
# There is a bug in curses, where *sometimes* it fails to properly
# initialize, and then after the endwin() call is made, the
# terminal is left in an unusable state. Rather than trying to
# check everytime for this (by requesting and comparing termios
# flags each time), we just save the initial terminal state and
# unconditionally reset it every time. It's cheaper than making
# the checks.
term_flags = termios.tcgetattr(sys.stdout)
# Curses modifies the stdout buffer size by default, which messes
# up Python's normal stdout buffering. This would manifest itself
# to IPython users as delayed printing on stdout after having used
# the pager.
#
# We can prevent this by manually setting the NCURSES_NO_SETBUF
# environment variable. For more details, see:
# http://bugs.python.org/issue10144
NCURSES_NO_SETBUF = os.environ.get('NCURSES_NO_SETBUF', None)
os.environ['NCURSES_NO_SETBUF'] = ''
# Proceed with curses initialization
try:
scr = curses.initscr()
except AttributeError:
# Curses on Solaris may not be complete, so we can't use it there
return screen_lines_def
screen_lines_real,screen_cols = scr.getmaxyx()
curses.endwin()
# Restore environment
if NCURSES_NO_SETBUF is None:
del os.environ['NCURSES_NO_SETBUF']
else:
os.environ['NCURSES_NO_SETBUF'] = NCURSES_NO_SETBUF
# Restore terminal state in case endwin() didn't.
termios.tcsetattr(sys.stdout,termios.TCSANOW,term_flags)
# Now we have what we needed: the screen size in rows/columns
return screen_lines_real
#print '***Screen size:',screen_lines_real,'lines x',\
#screen_cols,'columns.' # dbg
def page(strng, start=0, screen_lines=0, pager_cmd=None):
"""Display a string, piping through a pager after a certain length.
strng can be a mime-bundle dict, supplying multiple representations,
keyed by mime-type.
The screen_lines parameter specifies the number of *usable* lines of your
terminal screen (total lines minus lines you need to reserve to show other
information).
If you set screen_lines to a number <=0, page() will try to auto-determine
your screen size and will only use up to (screen_size+screen_lines) for
printing, paging after that. That is, if you want auto-detection but need
to reserve the bottom 3 lines of the screen, use screen_lines = -3, and for
auto-detection without any lines reserved simply use screen_lines = 0.
If a string won't fit in the allowed lines, it is sent through the
specified pager command. If none given, look for PAGER in the environment,
and ultimately default to less.
If no system pager works, the string is sent through a 'dumb pager'
written in python, very simplistic.
"""
# for compatibility with mime-bundle form:
if isinstance(strng, dict):
strng = strng['text/plain']
# Some routines may auto-compute start offsets incorrectly and pass a
# negative value. Offset to 0 for robustness.
start = max(0, start)
# first, try the hook
ip = get_ipython()
if ip:
try:
ip.hooks.show_in_pager(strng)
return
except TryNext:
pass
# Ugly kludge, but calling curses.initscr() flat out crashes in emacs
TERM = os.environ.get('TERM','dumb')
if TERM in ['dumb','emacs'] and os.name != 'nt':
print(strng)
return
# chop off the topmost part of the string we don't want to see
str_lines = strng.splitlines()[start:]
str_toprint = os.linesep.join(str_lines)
num_newlines = len(str_lines)
len_str = len(str_toprint)
# Dumb heuristics to guesstimate number of on-screen lines the string
# takes. Very basic, but good enough for docstrings in reasonable
# terminals. If someone later feels like refining it, it's not hard.
numlines = max(num_newlines,int(len_str/80)+1)
screen_lines_def = get_terminal_size()[1]
# auto-determine screen size
if screen_lines <= 0:
try:
screen_lines += _detect_screen_size(screen_lines_def)
except (TypeError, UnsupportedOperation):
print(str_toprint, file=io.stdout)
return
#print 'numlines',numlines,'screenlines',screen_lines # dbg
if numlines <= screen_lines :
#print '*** normal print' # dbg
print(str_toprint, file=io.stdout)
else:
# Try to open pager and default to internal one if that fails.
# All failure modes are tagged as 'retval=1', to match the return
# value of a failed system command. If any intermediate attempt
# sets retval to 1, at the end we resort to our own page_dumb() pager.
pager_cmd = get_pager_cmd(pager_cmd)
pager_cmd += ' ' + get_pager_start(pager_cmd,start)
if os.name == 'nt':
if pager_cmd.startswith('type'):
# The default WinXP 'type' command is failing on complex strings.
retval = 1
else:
fd, tmpname = tempfile.mkstemp('.txt')
try:
os.close(fd)
with open(tmpname, 'wt') as tmpfile:
tmpfile.write(strng)
cmd = "%s < %s" % (pager_cmd, tmpname)
# tmpfile needs to be closed for windows
if os.system(cmd):
retval = 1
else:
retval = None
finally:
os.remove(tmpname)
else:
try:
retval = None
# if I use popen4, things hang. No idea why.
#pager,shell_out = os.popen4(pager_cmd)
pager = os.popen(pager_cmd, 'w')
try:
pager_encoding = pager.encoding or sys.stdout.encoding
pager.write(py3compat.cast_bytes_py2(
strng, encoding=pager_encoding))
finally:
retval = pager.close()
except IOError as msg: # broken pipe when user quits
if msg.args == (32, 'Broken pipe'):
retval = None
else:
retval = 1
except OSError:
# Other strange problems, sometimes seen in Win2k/cygwin
retval = 1
if retval is not None:
page_dumb(strng,screen_lines=screen_lines)
def page_file(fname, start=0, pager_cmd=None):
"""Page a file, using an optional pager command and starting line.
"""
pager_cmd = get_pager_cmd(pager_cmd)
pager_cmd += ' ' + get_pager_start(pager_cmd,start)
try:
if os.environ['TERM'] in ['emacs','dumb']:
raise EnvironmentError
system(pager_cmd + ' ' + fname)
except:
try:
if start > 0:
start -= 1
page(open(fname).read(),start)
except:
print('Unable to show file',repr(fname))
def get_pager_cmd(pager_cmd=None):
"""Return a pager command.
Makes some attempts at finding an OS-correct one.
"""
if os.name == 'posix':
default_pager_cmd = 'less -r' # -r for color control sequences
elif os.name in ['nt','dos']:
default_pager_cmd = 'type'
if pager_cmd is None:
try:
pager_cmd = os.environ['PAGER']
except:
pager_cmd = default_pager_cmd
return pager_cmd
def get_pager_start(pager, start):
"""Return the string for paging files with an offset.
This is the '+N' argument which less and more (under Unix) accept.
"""
if pager in ['less','more']:
if start:
start_string = '+' + str(start)
else:
start_string = ''
else:
start_string = ''
return start_string
# (X)emacs on win32 doesn't like to be bypassed with msvcrt.getch()
if os.name == 'nt' and os.environ.get('TERM','dumb') != 'emacs':
import msvcrt
def page_more():
""" Smart pausing between pages
@return: True if need print more lines, False if quit
"""
io.stdout.write('---Return to continue, q to quit--- ')
ans = msvcrt.getwch()
if ans in ("q", "Q"):
result = False
else:
result = True
io.stdout.write("\b"*37 + " "*37 + "\b"*37)
return result
else:
def page_more():
ans = py3compat.input('---Return to continue, q to quit--- ')
if ans.lower().startswith('q'):
return False
else:
return True
def snip_print(str,width = 75,print_full = 0,header = ''):
"""Print a string snipping the midsection to fit in width.
print_full: mode control:
- 0: only snip long strings
- 1: send to page() directly.
- 2: snip long strings and ask for full length viewing with page()
Return 1 if snipping was necessary, 0 otherwise."""
if print_full == 1:
page(header+str)
return 0
print(header, end=' ')
if len(str) < width:
print(str)
snip = 0
else:
whalf = int((width -5)/2)
print(str[:whalf] + ' <...> ' + str[-whalf:])
snip = 1
if snip and print_full == 2:
if py3compat.input(header+' Snipped. View (y/n)? [N]').lower() == 'y':
page(str)
return snip