On Sat, Feb 17, 2018 at 11:13:00PM +0100, andre timmermans wrote:
>    Wish list:
>    - Extension in the config file to specify rom name path for hosfs & .hdf
>    files. This would make it easier to switch to different OS versions
>    (different ROMs, !Boot and compatible applications).

Hi Andre, thanks for the testing, on this matter I have an idea for a more 
comprehensive change to the way models are configured. But this will be 
in several releases time as there is prep work to support it first.

>    - Possibility to access the other folders/discs of the host system without
>    risk of modifying the file extension of the files (i.e. using mimemap like
>    Win95FS or LanManFS). I make sense for documents to be readable on both
>    sides without having constantly add and remove the ",xxx" filtype to the
>    filename's extension.

This I suspect will never happen, it's difficult enough to create a 
filesystem that's RISC OS compliant without the extra headache of being 
compatible with a variety of host OSes requirements too. This is mainly 
due to my extremely poor experiences with the highly configurable hostfs 
provided with red squirrel and various versions of Virtual Acorn where 
they ended up not even being compatible with each other.

>    Issues:
>    - ESC key not working

I have noticed this too occasionally, but have not been able to reproduce 
recently. When I did have it failing, it worked for a while then 
failed later. As such if you have any advice on how to reproduce it, 
please let me know.
>    - when switching back to windowed mode the mouse on the RISC OS side
>    cannot reach the top (I think it only occurs if the RISC OS resolution is
>    not the same on exit of full-screen mode qs on entry),

Oh well spotted, this is not a new bug, but caused by needing to implement 
part of 'follows host mouse' even when not using that (fullscreen mode 
does not use 'follows host mouse'. There will be a fix for this in the 
next version.

>    - sound is easely interrupted by disk accesses.

Annoyingly this is associated with the other fix of reducing the 'lag' in 
the sound you mentioned. I need to do a proper fix by triggering the sound 
interupts when data in consumed rather than just on a timer. However this 
is not going to be fixed by the next release version. There is a 
workaround (and you are allowed to be annoyed by me suggesting it) in that 
if you run it on a faster processor the timers smooth out and you get 
stutter free sound, it's not a great solution :)


Peter Howkins

Rpcemu mailing list

Reply via email to