Re: lynx-dev dev.10 progress

[Leonid Pauzner]

20-Nov-2002 05:56 Thomas Dickey wrote:
>> Try the attached patch. It reverts changes related to allocation of lines
>> in pool. Perhaps this will help to isolate the problem since the patch is
>> relatively small and straight forward. (Allocation of anchors left in pool).

> I tried something like this earlier this morning (am at work now), and it
> did fix the problem.  My guess is that one of the top-level variables is
> being nulled by the new code, and it was checked in the logic for deciding
> how to redisplay.  But narrowing it down to a small chunk is the first
> step in debugging the problem.


As a guess, try to memcpy/memset to a full range:

> @@ -2524,18 +2542,15 @@ PRIVATE HTLine * insert_blanks_in_line A
...
> -    memcpy(mod_line, line, LINE_SIZE(0));
> +    memcpy(mod_line, line, LINE_SIZE(1));

> @@ -2647,22 +2663,18 @@ PRIVATE void split_line ARGS2(
...
> +    /* can't wrap in middle of multibyte sequences, so allocate 2 extra */
> +    HTLine * line = (HTLine *)LY_CALLOC(1, LINE_SIZE(MAX_LINE)+2);

>      /*
> -     *  Set new line.
> +     *  Make new line.
>       */
> -    if (previous == (HTLine*)tmp_long_line[0])
> -       line = (HTLine*)tmp_long_line[1];
> -    else
> -       line = (HTLine*)tmp_long_line[0];
> -    memset(line, 0, LINE_SIZE(0));



; To UNSUBSCRIBE: Send "unsubscribe lynx-dev" to address@hidden