Re: history trouble

["Peter A. Castro" <doctor@fruitbat.org>]

On Sun, 29 Jan 2006, Thomas Richter wrote:

Greetings, Thomas,
   Can you verify something:
   Start a shell, enter a few commands, exit the shell, then look at the
history file and see if your commands were added to it?  If so, then
it's not a matter of your commands not being written to history, but that
they are being buffered and, upon exit, flushed to file.

Try this:
   setopt inc_append_history

> Hi Bart Schaefer <schaefer@brasslantern.com>,
> you wrote on Saturday, 2006-01-28 20:32:36 +0000:
>> On Jan 28,  8:04pm, Thomas Richter wrote:
>>
>> Possible reasons for the history not to be read are:
>>
>> - the NO_RCS option is set (the doc says this prevents history from
>>   being written, but it also suppresses reading it)
>
> This option is not set.
>
>> - the INTERACT option is not set
>
> This option is set.
>
>> - lockhistfile() fails because
>>   + a temporary lock file can't be created in the same directory
>>     as the HISTFILE
>
> How can I test this?
>
>>   + a link to that temp file can't be created
>
> ln -s $HISTFILE newname is ok
>
>>   + something else [possibly several somethings] has the file locked
>>     for more than about 10 seconds (but this shouldn't apply at shell
>>     startup, only for incremental history)
>
> no process has this file open
>
>> - the file can't be opened for reading
>
> cat $HISTFILE is ok
>
>> - some part of the file is corrupt
>
> no contains only one line with one word
>
>> - HISTSIZE is zero (in which case, I believe, the file is read but all
>>   the contents are discarded)
>
> HISTSIZE = 99999
>
>> My (new, ahem) guess is that the problem is with the lock file.
>
> wkr Thomas Richter

-- 
Peter A. Castro <doctor@fruitbat.org> or <Peter.Castro@oracle.com>
 	"Cats are just autistic Dogs" -- Dr. Tony Attwood