[Lcdproc] MtxOrb problem

[Michelle Dupuis]

Sounds great - send over the test program when you have a chance.

I noticed that the corrupt "]" character is at the end of a bar graph for a
CPU that is at 2% utilization.  So, it's not a problem of the bar graph
going too far to the right.  This looks more and more like a problem
relating the number of display lines!

What's strange is that without graphics, there were no scroll/overwrite
issues.  So something to do with using the graphics characters appears to be
causing a scroll / overwrite issue.  (The flickering reinforces that)

Thanks
MD 

-----Original Message-----
From: Ethan Dicks [mailto:ethan.dicks@gmail.com] 
Sent: Wednesday, October 04, 2006 10:27 PM
To: Michelle Dupuis
Cc: lcdproc@lists.omnipotent.net
Subject: Re: [Lcdproc] MtxOrb problem

On 10/5/06, Michelle Dupuis <support@ocg.ca> wrote:
> The MX212 is a current model (see
> http://www.matrixorbital.com/product_info.php?pName=mx212).  It is
actually
> an LK-202-USB combined with a drive bay insert.  The manual is available
> (http://www.matrixorbital.ca/manuals/LK_series/LK202-24-USB/) and I
believe
> that I do have 8 user definable characters.

I've grabbed the manuals but haven't opened them yet.  Thanks for the link.

> I turned off every client display except "SMP-CPU".  (I have a quad CPU so
4
> little graphs appear) to make testing easier.

I would suggest testing with one client screen at a time.

> This screen suffers from
> strange characters.  For example, the "]" at the right side of one of the
> four little bar graphs looks strange - as if it is overwritten with
another
> character.

Hmm... there's no such thing as "overwriting" with these displays
(they are character-cell-based, not pixel-based), so perhaps you are
seeing two chars fighting for the same spot.

> I should mention that my display is 20x2 character - and I think that has
> something to do with the problem.  The strange characters onscreen
combined
> with the rapid flickering MAY be due to scroll/overwriting of characters!

Indeed.  I haven't tested anything with a 2-line display in a long
time.  I'd really make sure the clients know to check for 2-line vs
4-line displays on the server.  So far, all of the clients I've
written myself in Perl _do_ check, but even my support isn't perfect
because I don't have any 2-line displays with me.

I _think_ the 'bigclock' was recently re-written to handle shorter
displays, but in the old days, it was 4 lines or nothing.

> (optical illusion).  Maybe the characters are fine but its alternating
> between two screens (or scrolling) more rapidly than I can detect?

That is entirely likely.

I have a simple h-bar testing client I wrote just for verifying that
the new MtxOrb stuff works.  I can send it to you if you like.  I'll
go check it and see if I can't make some 2-line stuff for it (right
now, it takes 3 lines... title bar, percent in digits (0%-100%), then
a full-width bar that happens to be 100 pixels on my display (5 pixels
per char, 20 chars).

You just fire up LCDd, then run this client and it counts from 0 to 100 at
1Hz.

-ethan

-ethan