mu-editor/mu

Column select, spaces, unstable, crash.

Opened this issue · 6 comments

MvGulik commented
  1. New file.
  2. Add single line with only spaces. (~64 spaces should do) (same with tabs)
  3. Initiate column select (CTRL+Mouse-Select) on that lines (at or near beginning of line).
  4. Extend selection with mouse towards the end of the line.
  5. Tends to crash at some point.
  • On multiple blank lines. Extending the selection over more than one line first, seems to bypass this issue. (as long as multiple lines are kept selected)

Terminal output:
Logging to /home/<user>/.cache/mu/log/mu.log (not included, did not contain relevant/related data)
free(): invalid next size (fast)
Aborted (core dumped)

Process 6851 (mu-editor) of user 1000 dumped core.
Stack trace of thread 6851:
#0  0x00007f4c3f5c7e97 __GI_raise (libc.so.6)
#1  0x00007f4c3f5c9801 __GI_abort (libc.so.6)
#2  0x00007f4c3f612897 __libc_message (libc.so.6)
#3  0x00007f4c3f61990a malloc_printerr (libc.so.6)
#4  0x00007f4c3f620f60 _int_free (libc.so.6)
#5  0x00007f4c2b8587f0 n/a (/home/<user>/test_venv/lib/python3.6/site-packages/PyQt5/Qt/lib/libQt5XcbQpa.so.5)
#6  0x00000000036283e0 n/a (n/a)
  • Suspected issue source: QScintilla (at least v2.11.3 & 2.11.4)

(Linux Mint 19.3 cinnamon)

(See also matkuki/ExCo#33 => Other effected code editors)

pip list:

Package          Version   
---------------- ----------
appdirs          1.4.3     
backcall         0.1.0     
certifi          2020.4.5.1
chardet          3.0.4     
colorzero        1.1       
cycler           0.10.0    
decorator        4.4.2     
docutils         0.16      
eric-ide         20.4      
gpiozero         1.5.1     
guizero          1.1.0     
idna             2.9       
ipykernel        5.2.1     
ipython          7.13.0    
ipython-genutils 0.2.0     
jedi             0.17.0    
jupyter-client   6.1.3     
jupyter-core     4.6.3     
kiwisolver       1.2.0     
Markdown         3.2.1     
matplotlib       2.2.2     
mu-editor        1.0.3     
nudatus          0.0.4     
numpy            1.18.2    
parso            0.7.0     
pexpect          4.8.0     
pgzero           1.2       
pickleshare      0.7.5     
pigpio           1.46      
Pillow           7.1.1     
pip              20.0.2    
pkg-resources    0.0.0     
prompt-toolkit   3.0.5     
ptyprocess       0.6.0     
pycodestyle      2.4.0     
pyflakes         2.0.0     
pygame           1.9.6     
Pygments         2.6.1     
pyparsing        2.4.7     
PyQt5            5.14.1    
PyQt5-sip        12.7.2    
PyQtChart        5.14.0    
PyQtWebEngine    5.14.0    
pyserial         3.4       
python-dateutil  2.8.1     
pytz             2019.3    
pyzmq            19.0.0    
QScintilla       2.11.4    
qtconsole        4.3.1     
requests         2.23.0    
semver           2.9.1     
setuptools       39.0.1    
six              1.14.0    
tornado          6.0.4     
traitlets        4.3.3     
urllib3          1.25.9    
wcwidth          0.1.9
MvGulik commented

Matkuki of Ex.Co editor has found and added additional information about this issue.
(see github link in op)

carlosperate commented

Thanks for the report @MvGulik!

Can't replicate in macOS with v1.1.0-alpha.2, but sounds like this could be a C/C++ issue in Qscintilla on Linux: matkuki/ExCo#33 (comment)

We still need to reproduce in Mu on Linux to make sure it's the same problem, but I'd be surprised if we can't reproduce (at least in debian-based distros).

@MvGulik are you planing to report this as well in the Qscintilla mailing list?

MvGulik commented

@MvGulik are you planing to report this as well in the Qscintilla mailing list?

Considering eric6' s developer already opened a related thread on it, and also seems to have had more Qscintilla mailing list contacts. I figured I first give him to opportunity to follow-up on it. On top of that, he has better experience and skill for this. (kinda think you all have in this case)

But ... if need be, yes I will.
Although, as I'm not having any Qscintilla experience(code-wise), I'm not sure if I can provide what was asked at the Qscintilla mailing list. (at least not yet)

(Mmm, I see eric6's dev is at least not on holiday ... but has not reacted to my update-post yet)

MvGulik commented

I'm not sure if I can provide what was asked at the Qscintilla mailing list.

But than again, Matkuki's reports might do the trick too.
Will forward those (including links) in a days or so to the Qscintilla mailing list. (if you like me to wait with that, just let me know)

MvGulik commented

Actually ... it turns out matkuki and eric6' s dev already posted a follow-up post on the Qscintilla mailing list.
https://www.riverbankcomputing.com/pipermail/qscintilla/2020-April/001451.html

...

Easier update checking links: (months without activity return as 404)

carlosperate commented

Thanks @MvGulik, looks like at the moment Qscintilla dev is still in the "trying to replicate" phase. We'll keep an eye.