Re: [Nano-devel] solving a history bug in nano [patch]

[Tito]

On 05/29/2016 03:18 PM, Benno Schulenberg wrote:
> On Sun, May 29, 2016, at 15:01, Tito wrote:
> > it will be hard to test it for me because I have to revive one of my
> > 2 old nokia n900 find the maemo sdk virtualbox image i used, but i will
> > try and post to the list if I succeed.
>
> Thanks.
>
> > Push the change if you are sure that the previous code is correct,
> > on the other hand i would dare to say that if nobody complained in over
> > seven years maybe it is the case to remove the find history feature
> > completely as it is not used?  ;-)
>
> :)
>
> Oh, it is used intensively, at least by me.  I regularly search alternatingly
> for two, three, four strings -- being able to just hit <Up><Up> <Up> is then
> a great help.  Apparently people seldom search first for some string and then
> for a shorter version and then want to search for the longer one again.
>
> (And since I know that tabbing works at the search prompt, I use that
> too sometimes.)
>
> Benno
Hi,
after a lot of fiddling this are my results, nano 2.5.3:
1) original source: search history does not work but nano works
2) with your patch: as soon as you do a search nano segfaults,

   ^W ahahaha <Enter>  boom

   other features (writing, saving to a file) work.

So there must be something wrong....

I suppose the reason for my patch was that
size_t is basically an unsigned integer so
(size_t)-1 looked bad and suspicious and it
somehow fixed the segfault. That was enough
for me at that time because i was not aware
of the search history feature.
Ideas about how to debug it even If I don't
use the N900 nowadays I would like to fix
it.

Ciao,
Tito