[Octave-bug-tracker] [bug #32524] history command overwrites files

[Thorsten Meyer]

URL:
  <http://savannah.gnu.org/bugs/?32524>

                 Summary: history command overwrites files
                 Project: GNU Octave
            Submitted by: tmeyier
            Submitted on: Thu 17 Feb 2011 09:27:42 PM UTC
                Category: Interpreter
                Severity: 4 - Important
                Priority: 5 - Normal
              Item Group: Incorrect Result
                  Status: None
             Assigned to: None
         Originator Name: 
        Originator Email: 
             Open/Closed: Open
         Discussion Lock: Any
                 Release: dev
        Operating System: GNU/Linux

    _______________________________________________________

Details:

I have been using the command

history -w new_script.m

to write the interactive command history to a file as a start for a new octave
script, always assuming that this command would write the history once and
leave new_script.m as it is afterwards. After that I used to delete everything
except the series of recent commands I was interested in (in new_script.m) and
cleaned up the code only to find later that it looked like the raw command
history again (and I repeatedly thought I had been really stupid because I
forgot to save the changes).
It took me several weeks to realize, that what really happened was that octave
wrote the history again to this new_script.m file when I quit the octave
session.

So what 

history -w new_script.m

apparently does is to write the history to the file immediately and write it
to the file again, when the octave session is closed.
This could still be a feature. However, after issuing the above command, the
command

history_file

still gives
ans = /home/thorsten/.octave_hist

Interestingly, when I explicitly set the history filename with 

history_file some_other_filename

after the history -w first_file.m command, history is written to
some_other_filename when the octave session ends.

While debugging this, I found that the same happens in this example:

history_file -r myfile.m

I.e., octave appends the content of myfile.m to the history and when the
session quits, the complete history is written to myfile.m overwriting its
original content. 

Attached is a patch, that seems to solve the problem. However, I am quite
confused by the fact, that the history file is kept within octave in two
locations:
  in an internal variable history_file
and
  as a property of the command_history object (accessed by
command_history::file and command_history::set_file)

Is this really intended? And if yes, does my patch interfere with this cunning
scheme in an unwanted way?

regards

Thorsten
 



    _______________________________________________________

File Attachments:


-------------------------------------------------------
Date: Thu 17 Feb 2011 09:27:42 PM UTC  Name: fix_history_bug  Size: 1kB   By:
tmeyier

<http://savannah.gnu.org/bugs/download.php?file_id=22719>

    _______________________________________________________

Reply to this item at:

  <http://savannah.gnu.org/bugs/?32524>

_______________________________________________
  Message sent via/by Savannah
  http://savannah.gnu.org/