Re: [BUG] sigsegv

[Sebastian Gniazdowski <sgniazdowski@xxxxxxxxx>]

It looks like the last field is incorrect:

(gdb) up
#1  0x00007ffff7e1f628 in histwgetfn (pm=0x508e10)
    at parameter.c:1241
1241                addlinknode(l, dupstring(h));
(gdb) p *l->list->last
Cannot access memory at address 0x7ffff5350068


On Tue, 25 Apr 2023 at 13:56, Peter Stephenson
<p.w.stephenson@xxxxxxxxxxxx> wrote:
>
> > On 25/04/2023 14:48 Sebastian Gniazdowski <sgniazdowski@xxxxxxxxx> wrote:
> > I have to say that current HEAD has many sigsegv errors, at least twp.
> > The reported in this email is an older problem with alt-/ history
> > complete, I've had it also with 5.8. I'm still waiting for the 5.9
> > only problem to occur, it's quite random and hard to reproduce.
> > To trigger the error, press alt-/ some times.
> >
> > #0  0x0000000000457868 in insertlinknode (list=list@entry=0x7ffff7abe368,
> >     node=0x7ffff55600d8, dat=0x7ffff556e8f0) at linklist.c:137
> > #1  0x00007ffff7bed6bf in histwgetfn (pm=<optimized out>) at parameter.c:1241
>
> The linked list management itself looks fairly harmless, but I wonder if there's
> something in the immediately surrounding code that could be fishy and messing
> up memory?
>
>         for (iw = he->nwords - 1; iw >= 0; iw--) {
>             h = he->node.nam + he->words[iw * 2];
>             e = he->node.nam + he->words[iw * 2 + 1];
>             sav = *e;
>             *e = '\0';
>             addlinknode(l, dupstring(h));
>             *e = sav;
>         }
>
> Some testing of strlen(he->node.nam) might not come amiss?
>
> Of course this is total guesswork.
>
> pws
>


-- 
Best regards,
Sebastian Gniazdowski