Ctrl-x not function to leave console #166

ghost · 2016-01-06T18:36:04Z

When attempting to make local calls in the console, pressing ctrl-x merely prints ^x rather than leaving the console. At this point, the only way out is to kill pudb. An interesting behavior is that this only happens after issuing a command in the console; ctrl-x works fine until then.

Details:
pudb version: 2015.4.1
installed with pip
running in osx 10.11.2 in iTerm2, Terminal, and Neovim's terminal mode

The text was updated successfully, but these errors were encountered:

inducer · 2016-01-06T19:16:29Z

I can't reproduce this, so this likely has something to do with your setup. On Linux, with vte (gnome terminal), konsole, and xterm, this works just fine for me.

Can you try through both display drivers (raw/curses) in the pref screen and see if that helps?

ghost · 2016-01-06T19:25:06Z

Changing to curses seems to have fixed it. Thanks.

nedbat · 2016-06-10T00:39:36Z

I have the same reported behavior on Mac. I am using iTerm2.
^X gets me into the console, and another ^X will get me out of it. But if I type '1' to see the value 1 in the console, then ^X just echoes ^X, and I can't figure out a way to do anything but kill pudb.

nedbat · 2016-06-10T00:43:13Z

Switching to curses doesn't improve the console, and makes other things bad :(

inducer · 2016-06-10T01:50:03Z

^X used to have some flow control behavior. Can you turn that off in iTerm2?

inducer · 2016-06-10T01:50:30Z

(I think. Actually not 100% sure about that.)

nedbat · 2016-06-10T02:31:00Z

That doesn't sound right to me. The first ^X would have that effect then, but it doesn't. And the second ^X doesn't stop the flow, it just echoes ^X, and you can still type into the console window.

inducer · 2016-06-10T14:24:14Z

@asmeurer, you use (used?) iTerm2, I recall. Any clues?

asmeurer · 2016-06-10T17:01:05Z

It is working for me, but I am using the iTerm2 nightly build. But it also works for me in Terminal.app. What version of urwid are you guys using? I have 1.3.1.

nedbat · 2016-06-10T19:21:55Z

Full freeze output (only some of which is for pudb):

-e hg+ssh://[email protected]/ned/coveragepy@f4001cd0cd5210e67ba8ba2e15c9bc362a001e2e#egg=coverage
covtestegg1==0.0.0
decorator==4.0.9
eventlet==0.19.0
gevent==1.1.1
greenlet==0.4.9
mock==2.0.0
nose==1.3.7
pbr==1.10.0
pudb==2016.2
PyContracts==1.7.9
Pygments==2.1.3
pyparsing==2.1.4
six==1.10.0
unittest-mixins==1.0
urwid==1.3.1

nedbat · 2016-06-10T20:14:18Z

BTW, I tested using the Mac Terminal app also, and it behaves the same for me as in iTerm2.

asmeurer · 2016-06-10T20:46:58Z

Yeah me too, that is, it works in both.

In fact, I just tried using the guest user account on my Mac to install pudb fresh, using Terminal.app, so that everything is default settings, and I still can't reproduce this. Any guesses as to what might be causing it. I also tried fiddling with different $TERM settings, but it doesn't seem to matter.

inducer · 2016-06-10T20:51:11Z

@nedbat, whats your $TERM?

nedbat · 2016-06-10T21:23:39Z

xterm-256color

inducer · 2016-06-10T21:28:33Z

Is that what you have, @asmeurer?

asmeurer · 2016-06-10T21:30:41Z

Yeah, and I tried all the other ones in the Terminal settings in my guest account. Haven't yet found any configuration that reproduces this.

nedbat · 2016-06-11T00:43:10Z

If you've got some suggestion for where in the pudb code to add some debugging information extraction, I'm game to give it a try.

mlubimow · 2016-06-28T16:06:19Z

@asmeurer Yes but shell.py is lazy loaded, after its loaded problem is gone but its quite annoying to open shell every time user wants to debug something.

asmeurer · 2016-06-28T16:14:15Z

According to #191, importing readline at the top of debugger.py fixes the issue.

When I start pudb and go straight to the console and type

import sys
sys.modules['readline']

it's there, meaning it's already been imported. Is it the same for you?

asmeurer · 2016-06-28T16:22:12Z

OK, it seems that shell.py is run as soon as you run any command (above, running import sys runs it, which imports readline). Even so, I can't get this to reproduce before readline is ever imported (as verified by running the above as a script with pudb).

mlubimow · 2016-06-28T16:33:42Z

I don't have this issue on another machine as well, this is happening only on this system (El Capitan 10.11.5). Im not familiar with this project but maybe when readline is imported, other library (urwid? subprocess?) uses this library to read line instead of something else and this is why it works?
Hard to say for me.

Surely if we try to read line from console even once before library gets imported, screen gets corrupted.

nedbat · 2017-02-09T23:30:29Z

I just noticed that this problem seems to be fixed now? I have iTerm2 v3.0.12

inducer · 2017-02-09T23:52:12Z

Thanks for the update. It doesn't feel like we meaningfully got to the bottom of what was causing this--but I'm glad to hear it's no longer occurring for you. It'd be tempting to just blame it on some external circumstance and move on, but I'm not sure I'm confident it won't rear its head again at some point.

Let's just leave this open for now.

gopar · 2017-02-24T23:07:13Z

Having the same issue.

$TERM: xterm-256color
Pudb: 2016.2
Mac: 10.12.3

Using GNU Screen (Problem still happens with GNU Screen off)
Using default Mac terminal app
Installed via pip

Just tried iTerm 3.0.14. Still has the bug

asmeurer · 2017-02-25T03:32:39Z

Didn't think of this before, but maybe the version of readline has something to do with it. What does this give you (may not work if you don't have the developer tools installed)?

import readline
import subprocess
subprocess.call(["otool", "-L", readline._rl.__file__])

asmeurer · 2017-02-25T03:35:46Z

If that gives an error about no attribute '_rl' use

import readline
import subprocess
subprocess.call(["otool", "-L", readline.__file__])

nedbat · 2017-02-25T15:06:59Z

Ctrl-X works for me currently, and here's my readline data point:

>>> import readline
>>> import subprocess
>>> subprocess.call(["otool", "-L", readline._rl.__file__])
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
AttributeError: 'module' object has no attribute '_rl'
>>> subprocess.call(["otool", "-L", readline.__file__])
/usr/local/virtualenvs/stell/lib/python2.7/lib-dynload/readline.so:
	/usr/lib/libedit.3.dylib (compatibility version 2.0.0, current version 3.0.0)
	/usr/lib/libncurses.5.4.dylib (compatibility version 5.4.0, current version 5.4.0)
	/usr/lib/libSystem.B.dylib (compatibility version 1.0.0, current version 1213.0.0)
0
>>>

gopar · 2017-02-26T03:04:32Z

Here's the output

Python 2.7.13 (default, Dec 18 2016, 07:03:39)
[GCC 4.2.1 Compatible Apple LLVM 8.0.0 (clang-800.0.42.1)] on darwin
Type "help", "copyright", "credits" or "license" for more information.
>>> import readline
>>> import subprocess
>>> subprocess.call(['otool', '-L', readline._rl.__file__])
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
AttributeError: 'module' object has no attribute '_rl'
>>> subprocess.call(['otool', '-L', readline.__file__])
/usr/local/Cellar/python/2.7.13/Frameworks/Python.framework/Versions/2.7/lib/python2.7/lib-dynload/readline.so:
        /usr/local/opt/readline/lib/libreadline.7.dylib (compatibility version 7.0.0, current version 7.0.0)
        /usr/lib/libncurses.5.4.dylib (compatibility version 5.4.0, current version 5.4.0)
        /usr/lib/libSystem.B.dylib (compatibility version 1.0.0, current version 1238.0.0)
0
>>>

gopar · 2017-03-02T20:42:37Z

Not sure if it makes any difference but here's they py3 output. I don't have any issues using this with pudb. Any other thing I should try to get pudb working again in py2? Don't like using other debuggers beside this one \o/

Python 3.6.0 (default, Dec 24 2016, 08:01:42) 
[GCC 4.2.1 Compatible Apple LLVM 8.0.0 (clang-800.0.42.1)] on darwin
Type "help", "copyright", "credits" or "license" for more information.
>>> import readline
>>> import subprocess
>>> subprocess.call(["otool", "-L", readline._rl.__file__])
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
AttributeError: module 'readline' has no attribute '_rl'
>>> subprocess.call(["otool", "-L", readline.__file__])
/usr/local/Cellar/python3/3.6.0/Frameworks/Python.framework/Versions/3.6/lib/python3.6/lib-dynload/readline.cpython-36m-darwin.so:
        /usr/local/opt/readline/lib/libreadline.7.dylib (compatibility version 7.0.0, current version 7.0.0)
        /usr/lib/libncurses.5.4.dylib (compatibility version 5.4.0, current version 5.4.0)
        /usr/lib/libSystem.B.dylib (compatibility version 1.0.0, current version 1238.0.0)
0
>>>

matiasbastos · 2018-12-03T20:33:26Z

Sorry but still having this issue :(. It gets solved doing what @asmeurer said (#166 (comment)). Im on py 2.7.10, pudb 2018.1 and iTerm2 3.2.6.... I'm missing something?

The issues with using a custom dict subclass to eval/exec are fixed in Python 3. Without this, local functions in the shell cannot access global variables (for instance, inside of list comprehentions). See the discussion on issues inducer#166 and inducer#103.

The issues with using a custom dict subclass to eval/exec are fixed in Python 3. Without this, local functions in the shell cannot access global variables (for instance, inside of list comprehensions). See the discussion on issues inducer#166 and inducer#103.

scirelli · 2019-06-13T17:18:31Z

I'm still seeing this issue.

asmeurer · 2019-06-13T18:42:23Z

@scirelli what version of pudb and Python are you using?

scirelli · 2019-06-14T00:00:48Z

@scirelli what version of pudb and Python are you using?

pudb: 2019.1
Python: 3.7.3

asmeurer · 2019-06-14T01:06:06Z

Maybe my diagnosis of the source of the problem was incorrect. Which makes sense because it was never really clear to me how the globals() dict behavior could cause issues with urwid.

Can you give me more info about your system. What OS is it, and what Python are you using (system Python, python.org, anaconda...)? Also do the exact steps from #166 (comment) reproduce the issue for you?

scirelli · 2019-06-14T02:10:28Z

Can you give me more info about your system. What OS is it, and what Python are you using (system Python, python.org, anaconda...)? Also do the exact steps from #166 (comment) reproduce the issue for you?

Mac, Python.org, just debugging using the cli python -m pudb <script>

All I have to do to reproduce is press <Ctrl+x>, try to execute a statement, then try to switch back to source with <Ctrl+x> again. It won't work and I'll start seeing control codes ^X etc. printed. They are printed outside of pudb. It's like pudb is not refreshing, typing enough characters and it'll scroll the previous text (pudb UI) up the screen. The only thing I can do is kill Python/pudb. I think if I hit enter enough times...not sure what causes it...but it'll try to execute everything I typed. I'll provide more info tomorrow, I'm not in front of that computer right now.

I was not having this problem a few weeks ago. I don't think this started happening until I updated to 2019.1. I was on 2018.1. I'll try to revert back to see what happens.

scirelli · 2019-06-14T13:28:42Z

@asmeurer
Recorded a video https://youtu.be/DZW94qwyrPE
macOS Mojova 10.14.5

Backspace wasn't working either.

I install pudb 2018.1 and it works fine. Looks like it's just something in the new version.

asmeurer · 2019-06-14T17:16:04Z

I'll check if I can reproduce this later (I need to install the python.org Python 3 on my Mac). It's also good to see that you are using a virtualenv. That may be relevant.

scirelli · 2019-06-26T14:58:27Z

I see it happening in 2018.1 when I hit the tab key to auto complete a variable name.

asmeurer · 2019-07-09T02:34:31Z

I was able to reproduce this with the python.org Python 3 on my Mac.

I also got a traceback which sent me to post-mortem mode:

Traceback (most recent call last):
  File "/Users/aaronmeurer/Documents/pudb/pudb/__init__.py", line 152, in runscript
    dbg._runscript(mainpyfile)
  File "/Users/aaronmeurer/Documents/pudb/pudb/debugger.py", line 468, in _runscript
    self.run(statement)
  File "/Library/Frameworks/Python.framework/Versions/3.7/lib/python3.7/bdb.py", line 585, in run
    exec(cmd, globals, locals)
  File "<string>", line 1, in <module>
  File "test.py", line 3, in <module>
    import pudb
  File "test.py", line 3, in <module>
    import pudb
  File "/Library/Frameworks/Python.framework/Versions/3.7/lib/python3.7/bdb.py", line 88, in
trace_dispatch
    return self.dispatch_line(frame)
  File "/Users/aaronmeurer/Documents/pudb/pudb/debugger.py", line 189, in dispatch_line
    self.user_line(frame)
  File "/Users/aaronmeurer/Documents/pudb/pudb/debugger.py", line 410, in user_line
    self.interaction(frame)
  File "/Users/aaronmeurer/Documents/pudb/pudb/debugger.py", line 378, in interaction
    show_exc_dialog=show_exc_dialog)
  File "/Users/aaronmeurer/Documents/pudb/pudb/debugger.py", line 2147, in call_with_ui
    return f(*args, **kwargs)
  File "/Users/aaronmeurer/Documents/pudb/pudb/debugger.py", line 2400, in interaction
    self.event_loop()
  File "/Users/aaronmeurer/Documents/pudb/pudb/debugger.py", line 2360, in event_loop
    keys = self.screen.get_input()
  File
"/Library/Frameworks/Python.framework/Versions/3.7/lib/python3.7/site-packages/urwid/raw_display.py",
line 336, in get_input
    keys, raw = self.parse_input(None, None, self.get_available_raw_input())
  File
"/Library/Frameworks/Python.framework/Versions/3.7/lib/python3.7/site-packages/urwid/raw_display.py",
line 436, in get_available_raw_input
    codes = self._get_gpm_codes() + self._get_keyboard_codes()
  File
"/Library/Frameworks/Python.framework/Versions/3.7/lib/python3.7/site-packages/urwid/raw_display.py",
line 510, in _get_keyboard_codes
    code = self._getch_nodelay()
  File
"/Library/Frameworks/Python.framework/Versions/3.7/lib/python3.7/site-packages/urwid/raw_display.py",
line 644, in _getch_nodelay
    return self._getch(0)
  File
"/Library/Frameworks/Python.framework/Versions/3.7/lib/python3.7/site-packages/urwid/raw_display.py",
line 554, in _getch
    return ord(os.read(self._term_input_file.fileno(), 1))
TypeError: ord() expected a character, but string of length 0 found

(also annoyingly, the exception viewer did not have the option to save the traceback to a file)

asmeurer · 2019-07-09T02:37:56Z

Bisected the problem, and unsurprisingly, this is the bad commit:

commit b0fc02060bee885a7846384a50a5a5c20d577d9f
Author: Aaron Meurer <[email protected]>
Date:   Mon Mar 11 16:29:25 2019 -0600

    Only avoid SetPropogatingDict in the builtin shell in Python 2

    The issues with using a custom dict subclass to eval/exec are fixed in
    Python 3. Without this, local functions in the shell cannot access global
    variables (for instance, inside of list comprehensions).

    See the discussion on issues #166 and #103.

So it seems my theory that this issue should only occur in Python 2 was wrong. There is something about the system Python that is at issue here, as it doesn't do it for me with the Anaconda Python.

asmeurer · 2019-07-09T02:40:08Z

Maybe it has to do with the python.org Python (and also the system Python) using libedit instead of readline. Which means it could be related to #337.

asmeurer · 2019-07-09T02:42:06Z

It's definitely readline. If I remove the readline imports in shell.py, the issue goes away. #337 doesn't fix it however.

This fixes inducer#166. I have no idea why, but if readline is really libedit (like for a system or python.org install on Mac), it causes the built in shell to break the terminal inputs, but only when imported at the top level. If the import is deferred, as far as I've seen, it doesn't happen.

aseutin · 2019-11-05T21:42:09Z

I'm having this same problem using the latest version of pudb and iterm2 on a Mac (I'm also using the curses shell in pudb). To reproduce the issue I just have to run pudb main.py, then hit ctrl-x or "!" to enter the python shell, then type some gibberish and click enter. Backspace stops working and ctrl-x appears in the prompt. It seems like ctrl-x does actually work (even though it doesn't appear to be working visually) because hitting ctrl-X then 'n' then will actually exit the shell and go to the next line as expected. Once the UI updates the issue will be fixed for the rest of the session and the shell works as expected

golharam · 2020-01-17T21:08:27Z

I'm just now running into this.
MacOS Mojava
Python 3.6.5

pudb was working fine for my under Python 2.7. I only noticed this when moving to Python 3. I"m using Terminal.app.

Other than that, I have nothing useful to contribute.

asmeurer · 2020-01-17T22:43:52Z

Can you check if #350 fixes the problem?

golharam · 2020-02-24T23:00:48Z

Yep, that fixed it

JamoBox · 2020-03-31T11:33:25Z

@asmeurer I also ran into this issue using Python 3.8.0 and Kitty terminal on MacOS.

I've just tried out your patch and can confirm it fixes the issue for me.

Xepe · 2020-11-14T01:18:41Z

I am having the same issue, I can't exit from the command line using ctrl x after writing something

asmeurer mentioned this issue Jun 28, 2016

OSX - going into console and typing causes draw issues (console is losing focus?) #191

Closed

inducer closed this as completed in #236 Apr 4, 2017

asmeurer mentioned this issue Mar 11, 2019

Names not defined in list comprehension in the IPython shell #103

Open

asmeurer mentioned this issue Mar 11, 2019

Only avoid SetPropogatingDict in the builtin shell in Python 2 #330

Merged

asmeurer reopened this Jun 14, 2019

asmeurer mentioned this issue Jul 9, 2019

fix issue with libedit messing up terminal state #337

Open

asmeurer mentioned this issue Jul 9, 2019

Move the readline import out of the top (module) level and add Jedi completion #350

Merged

asmeurer mentioned this issue Jul 17, 2019

Variable printing in command line messes up ui #352

Open

asmeurer closed this as completed in #350 Mar 31, 2020

Ctrl-x not function to leave console #166

Ctrl-x not function to leave console #166

Comments

ghost commented Jan 6, 2016

inducer commented Jan 6, 2016

ghost commented Jan 6, 2016

nedbat commented Jun 10, 2016

nedbat commented Jun 10, 2016

inducer commented Jun 10, 2016

inducer commented Jun 10, 2016

nedbat commented Jun 10, 2016

inducer commented Jun 10, 2016

asmeurer commented Jun 10, 2016

nedbat commented Jun 10, 2016

nedbat commented Jun 10, 2016

asmeurer commented Jun 10, 2016

inducer commented Jun 10, 2016

nedbat commented Jun 10, 2016

inducer commented Jun 10, 2016

asmeurer commented Jun 10, 2016

nedbat commented Jun 11, 2016

mlubimow commented Jun 28, 2016

asmeurer commented Jun 28, 2016

asmeurer commented Jun 28, 2016

mlubimow commented Jun 28, 2016 • edited Loading

nedbat commented Feb 9, 2017

inducer commented Feb 9, 2017

gopar commented Feb 24, 2017 • edited Loading

asmeurer commented Feb 25, 2017

asmeurer commented Feb 25, 2017

nedbat commented Feb 25, 2017

gopar commented Feb 26, 2017

gopar commented Mar 2, 2017 • edited Loading

matiasbastos commented Dec 3, 2018 • edited Loading

scirelli commented Jun 13, 2019

asmeurer commented Jun 13, 2019

scirelli commented Jun 14, 2019 • edited Loading

asmeurer commented Jun 14, 2019

scirelli commented Jun 14, 2019

scirelli commented Jun 14, 2019 • edited Loading

asmeurer commented Jun 14, 2019

scirelli commented Jun 26, 2019

asmeurer commented Jul 9, 2019

asmeurer commented Jul 9, 2019

asmeurer commented Jul 9, 2019

asmeurer commented Jul 9, 2019

aseutin commented Nov 5, 2019

golharam commented Jan 17, 2020

asmeurer commented Jan 17, 2020

golharam commented Feb 24, 2020

JamoBox commented Mar 31, 2020

Xepe commented Nov 14, 2020

mlubimow commented Jun 28, 2016 •

edited

Loading

gopar commented Feb 24, 2017 •

edited

Loading

gopar commented Mar 2, 2017 •

edited

Loading

matiasbastos commented Dec 3, 2018 •

edited

Loading

scirelli commented Jun 14, 2019 •

edited

Loading

scirelli commented Jun 14, 2019 •

edited

Loading