Re: [PATCH v3 0/11] MR2993: comdlg32: IFileDialog navigation bar control. Part 1.

[Vladislav Timonin (＠vt)]

On Tue Jun 13 19:10:31 2023 +0000, Esme Povirk wrote:
> I made an attempt at simplification (untested):
>     LIST_FOR_EACH_ENTRY_REV(crumb1, &info->crumbs, struct crumb, entry)
>     {
>         INT next_w = w + crumb1->full_w;
>         crumb1->current_w = crumb1->full_w;
>         if (next_w > max_crumbs_w)
>         {
>             if (crumbs_visible_n == 0)
>             {
>                 /* fit as much of the last crumb as possible */
>                 crumb1->current_w = max_crumbs_w - w;
>                 w = max_crumbs_w;
>                 crumbs_visible_n += 1;
>                 /* try to also fit the next crumb */
>                 continue;
>             }
>             else if (crumbs_visible_n == 1)
>             {
>                 INT min_crumb_width = (crumbs_visible_n == 1) ?
>                     MulDiv(56, info->dpi_x, USER_DEFAULT_SCREEN_DPI) : 1;
>                 min_crumb_width = min(min_crumb_width, crumb1->full_w);
>                 if (w + min_crumb_width <= max_crumbs_w)
>                 {
>                     /* fit as much of this crumb as possible */
>                     crumb1->current_w = max_crumbs_w - w;
>                     w = max_crumbs_w;
>                 }
>                 else
>                 {
>                     /* take space away from the last crumb to get this
> one to the minimum */
>                     struct crumb *last_crumb = (struct crumb *)list_tail(&info->crumbs);
>                     last_crumb->current_w = max(0, max_crumbs_w - min_crumb_width);
>                     crumb1->current_w = min_crumb_width;
>                     w = last_crumb->current_w + crumb1->current_w;
>                 }
>                 crumbs_visible_n += 1;
>             }
>             break;
>         }
>         w = next_w;
>         crumbs_visible_n += 1;
>     }
>
> It's still complicated, but it can use some bits of the loop and reduce
> the special cases for fewer than 2 items.

Nice, starting point calculations (`prev_x`) are missing, but after including them and giving it a quick test, does seem to work.