SpeedUI update to 2005 released:-)

[Raymond Del Tondo]

Hello all,

I just released an updated version (2005) of SpeedUI,
the fast and only modular user interface accelerator for the HP 48G series,
and the only one which still gets updates after more than 15 years.

Here's the link: https://www.hpmuseum.org/guest/deltondo/sui_2005.zip

Be sure to first read the Readme.txt , then Relnotes.txt , then the other information slices.

What has changed over the years? Some places experienced kinda fine tuning and bug fixing.
But one of the most important things is: User mode should now work:-)

If you still find a bug, please contact me via mail or pn - thanks.
If not, even better.
In either case, enjoy:-)

[Sandhri]

Thanks Raymond! I've been waiting for this update. I have one question. Are any modules unchanged or should I delete them all and install everything again?

[SammysHP]

Ohh... awesome! Just when you think that something seems to be dead finally, it stands up and says Hi.

I think there were some documentation insufficiencies in the documentation for new users. When I install the new version I will take some notes what could be added.

[Marco Polo]

(06-03-2020 08:47 PM)Raymond Del Tondo Wrote:  Hello all,

I just released an updated version (2005) of SpeedUI,
the fast and only modular user interface accelerator for the HP 48G series,
and the only one which still gets updates after more than 15 years.

Here's the link: https://www.hpmuseum.org/guest/deltondo/sui_2005.zip

Be sure to first read the Readme.txt , then Relnotes.txt , then the other information slices.

What has changed over the years? Some places experienced kinda fine tuning and bug fixing.
But one of the most important things is: User mode should now work:-)

If you still find a bug, please contact me via mail or pn - thanks.
If not, even better.
In either case, enjoy:-)

Great!!!
I am testing right now and User mode seems to work flawlessly. THANK YOU!!!!

One question: what are libraries FC, FM and TO for?

[Raymond Del Tondo]

(06-03-2020 09:50 PM)Sandhri Wrote:  Thanks Raymond! I've been waiting for this update. I have one question. Are any modules unchanged or should I delete them all and install everything again?

Since the last release was some years ago, and several code shifts and parameter changes happened, I'd recommend replacing all SpeedUI components.

Some of the libraries which rely on code in another library have a built-in version check.
Iow there are checks whether a function call into another library goes to a compatible release, which means in most cases: The same version as the caller.
This is a rudimentary check only, but serves the intended purpose.

[Raymond Del Tondo]

(06-04-2020 05:02 AM)SammysHP Wrote:  Ohh... awesome! Just when you think that something seems to be dead finally, it stands up and says Hi.

Hi:-)

(06-04-2020 05:02 AM)SammysHP Wrote:  I think there were some documentation insufficiencies in the documentation for new users. When I install the new version I will take some notes what could be added.

Yes, please!

There's a great chance that first contact to SpeedUI installation doc could be too complicated or irritating for some users,
although I tried to create a relatively simple entry in the SpeedUI Start Here pdf.

[Raymond Del Tondo]

(06-04-2020 12:35 PM)Marco Polo Wrote:  Great!!!
I am testing right now and User mode seems to work flawlessly. THANK YOU!!!!

One question: what are libraries FC, FM and TO for?

Hope you like it:-)
FC is my Flag Catalog replacement. If installed, you can call it via cmd line FC, the menu key FC, or via the [Flags] menu key in the Modes form.

FM is the font management library, which is an extension the the original UFL font management library. My version can handle up to four UFL fonts.
I see this isn't in the docs so far, and should be added.

TO is not exactly a part of SpeedUI. Actually it's my personal Tools library,
consisting of some very basic tools. Somehow it slipped into the package:
tEVAL Timed EVAL
SYS Alias for SYSEVAL
MON My full screen memory monitor (should update the info screen, which shows 1993;-)
PGLIB Purge library given by decimal id. Also does detach beforehand
BLANK$ Fast ML version. This thing eases creation of low memory situations.
->XLIB Takes ROM ID and index number, and creates XLIB Pointer
ASC-> Decode from ASCed transfer format. Outdated, not needed anymore, except for files only available in ASC fmt.
->ASC Code into ASC transfer format. Outdated, not needed anymore.
SWAPLIB takes a LID and tries to swap ROM Part Table entries of the first two libs with same LID. Effect is activating the other one.
LIBADR Return address of library, if existent in ROM Part Table
ADR2PTR Takes a Bint and returns a pointer to the address given in the Bint

For most users these tools will be useless. Amongst other tools, I used some of them while developing and debugging SpeedUI.
Relatively sure they will not be in the next release of SpeedUI.

[Marco Polo]

(06-04-2020 11:19 PM)Raymond Del Tondo Wrote:  

(06-04-2020 12:35 PM)Marco Polo Wrote:  Great!!!
I am testing right now and User mode seems to work flawlessly. THANK YOU!!!!

One question: what are libraries FC, FM and TO for?

Hope you like it:-)

Thanks for the infos.
In the meanwhile, i found a nice candy: when adding UOM to a number the typing of "_" automatically switches to alpha mode. Very useful :-)

The User keyboard works fine but i found a small glitch: when using user mode on "." key and having flag -61 cleared (USR USR locks user mode), the user mode remains active and "1US" annunciator ON.
It remains active until i press any other key (which is executed in user mode) EXCEPT ".", then it is disengaged.
If "." is pressed, the user mode and "1US" annunciator remain active.
This happens even with no user keys defined (<< 0 DELKEYS >> executed before testing).
No other key seems to trigger the same behaviour.
Anyway, no crashes so far

[SammysHP]

One thing I noticed with the previous and the current version: If you install CF lite and UI full, an equation on the stack will be shown as .. (after the screen goes mostly empty for a short time). The documentation does not mention this behavior with pretty print.

[Raymond Del Tondo]

(06-05-2020 01:17 PM)SammysHP Wrote:  [..]CF lite and UI full[..]. The documentation does not mention this behavior with pretty print.

Actually this combination (and the combination CF full and UI lite) weren't tested.
The lite versions were made mainly to show that SpeedUI can be had within the 32K limit.
Anyway, thank you for the hint. This will have to be documented, as well as the display annunciator glitch and the UI candy reported by Marco Polo;-)

[Raymond Del Tondo]

So here's the first update for the update:-)

I found (and hopefully eliminated) the expression display bug.
Many thanks to SammysHP for finding it.

The updated archive is available for download, see first post.
The only binary that was changed is CF_Lite.LIB .
Please replace a previous CF_Lite with the updated version from 10th of June, 2020.
All other files except the release notes remain unchanged.

If you want more details, please read on:
The bug mentioned above only occurs when CF_Lite and UI_Full are installed.

The "Pretty Print" feature sits inside the UI.LB full version.
In many cases this mechanism works nearly self-contained,
but there are situations where the "Pretty Print" evaluator calls a
"decompile anything, and as fast as possible" function, which is located in the full version CF.LIB.

That is the ML decompiler, mentioned elsewhere. Only things to know about that thing:
MUCH faster than the HP 48 built-in general decompiler, but needs more than 10k of space.

In the lite version of CF.LIB , the ML decompiler simply doesn't exist.
However there was no replacement code, and thus the caller called the next entry inside the lib, and that was the bug.

As indicated, I inserted a replacement code for the ML decompiler code call in the lite version of CF.LIB ,
so now it should work. If you encounter any problems, please don't hesitate to contact me.

The 1US display glitch reported by Marco Polo has not been addressed yet,
but it's noted, and thus may be corrected in a later release;-)

[Marco Polo]

(06-10-2020 10:46 PM)Raymond Del Tondo Wrote:  The 1US display glitch reported by Marco Polo has not been addressed yet,
but it's noted, and thus may be corrected in a later release;-)

Thank you again for your efforts.

I just would like to point out that the 1US issue is not just a display glitch but, at least on my Emu48 calculator, the behaviour is identical to USR locked, but only for "." key.

May this key is handled in a differenta way?!?!?

[SammysHP]

Just found a small graphical glitch:

In the (built-in) flag browser and the scroll indicator set to scrollbar, the selection does not touch the scrollbar. If you scroll down and the whole screen moves one line up, the selection extends up to the scrollbar. Now if you move up again the small area next to the scrollbar is not updated.







Installed libraries: CF_full, UI_full, B0, B1, B3, SBUC, UF, UFLF