[syslinux] 5.10 regression (from 5.01) MENU INCLUDE broken.

[Ady]

> On 09/05/2013 01:55 AM, Gene Cumm wrote:
> > On Wed, Sep 4, 2013 at 2:34 PM, Piotr Karbowski
> > <piotr.karbowski at gmail.com> wrote:
> >
> >> With syslinux (extlinux) it will jump to last INCLUDE'ed menu always, with
> >> 5.01 it will display list of elements, then allow me to choose which one I
> >> want to enter.
> >>
> >> Even if the include part contain a single label + menu label, it will
> >> display only it with newer syslinux.
> >
> > What do the INCLUDEd config files look like?  Perhaps an overriding DEFAULT?
> >
> 
> ESC does not bring me to the proper menu.
> 
> Attached config files.
> 
> if extlinux.conf have only one include 'sysrcd-2.4.1.conf' it does not 
> display the 'first' menu with 5.10 but goes into the sysrescd-2.4.1.conf 
> config. Or whatever is included as last.
> 
> -- Piotr.
> 
 
1. Delete all unneeded:
 *** 
label -
  menu label 
  menu disable
 *** 
 Optionally, replace them with an empty line.

2. Avoid duplicated LABEL (e.g. more than one "LABEL live"; change 
the label names so to make them unique, at least within each cfg file 
if not all across).

3. Correct "Ur	 vesamenu.c32".

4. By using multiple UI directives in INCLUDEd cfg files, together 
with several TIMEOUT and ONTIMEOUT directives (also in INCLUDEd cfg 
files), you are elevating the chances of "unexpected" behaviors. 
Using these directives in INCLUDEd files is indeed allowed, but you 
are more prone to some "rare" result. This might not be a real 
problem, but simplifying the cfg files might help debugging your 
situation.

5. Since you are repeatedly using UI directives in each INCLUDEd cfg, 
your MENU EXIT entries won't work, just as the ESC key doesn't return 
to the parent menu.

I would suggest simplifying your cfg files by commenting out all 
non-essential lines / directives and re-testing. If it is still not 
working as expected, perhaps you could try the latest 5.xx-preNN 
testing pre-release.

I'm not saying your original report about a possible regression is 
incorrect. I'm just suggesting that you could/should make a more 
"adequate" usage of Syslinux directives, so to confirm whether there 
is indeed a regression. Whichever the case, please report back.

Best Regards,
Ady.