history-search-backward function disappears suddenly

[ViktigLemma]

Sometimes while using ipython I lose the following functions:

readline_parse_and_bind "\e[A": history-search-backward
readline_parse_and_bind "\e[B": history-search-forward

Instead I get that the arrows just go through command history without considering what I've already written.

I haven't been able to see the pattern in when it happens, but it never happens in the beginning of a session. Here is a workflow example where it just happened:

===============================================================
In [47]: Correlator(lizt,ran='no',kind='RNA',sumit='yes',p=0.05,msatyes='yes')
---------------------------------------------------------------------------
TypeError Traceback (most recent call last)

/home/jorgsk/phdproject/5primeUTR/<ipython console> in <module>()

TypeError: Correlator() got an unexpected keyword argument 'msatyes'
> <ipython console>(1)<module>()

ipdb> x
*** NameError: name 'x' is not defined
ipdb> q

In [48]: X
---------------------------------------------------------------------------
NameError Traceback (most recent call last)

/home/jorgsk/phdproject/5primeUTR/<ipython console> in <module>()

NameError: name 'X' is not defined
> <ipython console>(1)<module>()

ipdb> x
*** NameError: name 'x' is not defined
ipdb> q

In [49]: execfile('transcriptMain.py')

==============================================================

In "In [47]" the history-search-backward is working. In "In [49]" it is not working. All I did was go to work with the debugger and make some faulty key-presses. If I reproduce it again I'll try and pay more attention to what is happening.

This has happened to me 5-6 times the last months (I'm not programming every day). I'd say it happens perhaps ever 5th session.

[ViktigLemma (jorgsk)]

I tried to reproduce these steps exactly with a restart if ipython, but I could not reproduce it.

[Fernando Perez (fdo.perez)]

OK, closing for now since it seems really hard to reproduce. Please do reopen it if you find a way to trigger it reliably, so we can track it down (I've never seen the problem myself).