Picomite(VGA) V5.07.07 betas - bug fixes + focus on PIO

Author Message
matherp
Guru

Joined: 11/12/2012
Location: United Kingdom
Posts: 11615
Posted: 01:25pm 13 Jan 2023      

Picomite(VGA) V5.07.07b1

https://geoffg.net/Downloads/picomite/PicoMite_Beta.zip

Bug fix to ON ERROR SKIP (thanks Tom)

New functions

MM.INFO(EXISTS FILE fname$) ' see CMM2 manual for details
MM.INFO(EXISTS DIR dname$) ' see CMM2 manual for details

PIN(FLEVEL pio [,sm, DIR]) ' dir can be RX or TX. If specified gives the level of the specific fifo

PIO (SHIFTCTRL push_threshold [,pull_threshold] [,autopush] [,autopull] [,in_shiftdir] [,out_shiftdir] [,FJOIN_RX] [,FJOIN_TX]) 'additional optional parameters to create 8-word RX or TX FIFOs

PIO CLEAR now clears all the FIFOs for the pio specified
PIO INIT and PIO START clear the specific FIFOs for the pio and sm specified

New Command

NB: RX and TX are from the Basic code perspective

PIO INTERRUPT pio, sm [,RXinterrupt] [,TXinterrupt]

Use the value 0 for RXinterrupt or TXinterrupt to disable an interrupt
Omit values not needed

e.g

PIO INTERRUPT 1,0,,myTXinterrupt
PIO INTERRUPT 1,0,,0
PIO INTERRUPT 1,0,myRXinterrupt
PIO INTERRUPT 1,0,0
PIO INTERRUPT 1,0,myRXinterrupt,myTXinterrupt
PIO INTERRUPT 1,0,0,0


The RX interrupt triggers whenever a word has been "pushed" by the PIO code into the specified FIFO. The data MUST be read in the interrupt to clear it. The interrupt will re-trigger if a second word arrives and the interrupt only processes one.

The TX interrupt triggers whenever the specified FIFO has been FULL and the PIO code has now "pulled" it

PIO interrupts have priority have keyboard interrupts but before anything else. As with all interrupts interrupt conditions are processed one at a time.

NB: I am not convinced about the utility of the TX interrupt. Would it be better to have an interrupt on say half full  when it was previously more than half. This would allow the Basic code to keep the FIFO topped up and know how many writes were allowed NB: the enhanced PIO(FLEVEL ...) function makes it easy for the Basic code to know the FIFO status before writing
Volhout
Guru

Joined: 05/03/2018
Location: Netherlands
Posts: 5991
Posted: 02:03pm 13 Jan 2023      

Will try this weekend !! Looks very promising. I see a 3 phase sine wave generator on the horizon....

  matherp said  Picomite(VGA) V5.07.07b1

PIN(FLEVEL pio [,sm, DIR]) ' dir can be RX or TX. If specified gives the level of the specific fifo


I Guess this is a typo....

PIO(FLEVEL pio [,sm, DIR])

Regards,

Volhout
matherp
Guru

Joined: 11/12/2012
Location: United Kingdom
Posts: 11615
Posted: 02:06pm 13 Jan 2023      

  Quote  I Guess this is a typo....


No: it says you can specify just the pio, or all of the pio, statemachine and direction
thwill

Guru

Joined: 16/09/2019
Location: United Kingdom
Posts: 4370
Posted: 02:06pm 13 Jan 2023      

  matherp said  Picomite(VGA) V5.07.07b1 ...

Bug fix to ON ERROR SKIP (thanks Tom)


Thanks for your confidence Peter, I hope it proves to be justified.

  matherp said  New functions

MM.INFO(EXISTS FILE fname$) ' see CMM2 manual for details
MM.INFO(EXISTS DIR dname$) ' see CMM2 manual for details


Great, I'd been hoping for those for a while.

Could I push my luck and ask for:

       MM.INFO(EXISTS file_or_dir_name$)

Which both MMBasic for Windows and MMB4L already have, but not the CMM2 ... but perhaps that could have too ?

Best wishes,

Tom
Edited 2023-01-14 00:07 by thwill
matherp
Guru

Joined: 11/12/2012
Location: United Kingdom
Posts: 11615
Posted: 02:09pm 13 Jan 2023      

  Quote  Could I push my luck and ask for:


Isn't that effectively there with MM.INFO(filesize fname$) which returns -2 if it is a directory
Edited 2023-01-14 00:11 by matherp
thwill

Guru

Joined: 16/09/2019
Location: United Kingdom
Posts: 4370
Posted: 02:18pm 13 Jan 2023      

  matherp said  
  Quote  Could I push my luck and ask for:


Isn't that there with MM.INFO(exists file fname$) which returns -2 if it is a directory


Are you talking historically, in which case I think it is MM.INFO(FILESIZE fname$) which has that behaviour, or did, you may have changed it to report an error the last time I got on this hobby horse .

Or are you talking about what you've just done, in which case I'll have to take a look, though I'd still appreciate consistency between all the MMBasic platforms.

Maybe I should just let this slide until I've got my tests running on the PicoMite and then I will be able to quibble from a position based more on facts than supposition.

Thanks for the new additions,

Tom
thwill

Guru

Joined: 16/09/2019
Location: United Kingdom
Posts: 4370
Posted: 02:21pm 13 Jan 2023      

  matherp said  
  Quote  Could I push my luck and ask for:


Isn't that effectively there with MM.INFO(filesize fname$) which returns -2 if it is a directory


OK, you edited, and no, on some platforms that now throws an error.

This is the passing test code from when I checked about a week ago (but not with PicoMite):

Sub test_filesize_given_directory()
 If Mm.Device$ = "MMB4L" Then
   assert_int_equals(-2, Mm.Info(FileSize Mm.Info$(Directory)))
   assert_int_equals(-2, Mm.Info(FileSize Mm.Info$(Path)))
   assert_int_equals(-2, Mm.Info(FileSize "/"))
   assert_int_equals(-2, Mm.Info(FileSize "\"))
   assert_int_equals(-2, Mm.Info(FileSize "."))
   assert_int_equals(-2, Mm.Info(FileSize ".."))
   assert_int_equals(-2, Mm.Info(FileSize "A:/"))
   assert_int_equals(-2, Mm.Info(FileSize "A:\"))
   assert_int_equals(-2, Mm.Info(FileSize "C:/"))
   assert_int_equals(-2, Mm.Info(FileSize "C:\"))
   assert_int_equals(-2, Mm.Info(FileSize str.replace$(TMP$, "/", "\")))
   assert_int_equals(-2, Mm.Info(FileSize str.replace$(TMP$, "\", "/")))
   assert_int_equals(-2, Mm.Info(FileSize str.replace$("A:" + TMP$, "/", "\")))
   assert_int_equals(-2, Mm.Info(FileSize str.replace$("A:" + TMP$, "\", "/")))
   assert_int_equals(-2, Mm.Info(FileSize str.replace$("C:" + TMP$, "/", "\")))
   assert_int_equals(-2, Mm.Info(FileSize str.replace$("C:" + TMP$, "\", "/")))
 Else
   assert_filesize_fails(Mm.Info$(Directory))
   assert_filesize_fails(Mm.Info$(Path))
   assert_filesize_fails("/")
   assert_filesize_fails("\")
   assert_int_equals(Choice(Mm.Device$ = "MMBasic for Windows", -2, -1), Mm.Info(FileSize "."))
   assert_int_equals(Choice(Mm.Device$ = "MMBasic for Windows", -2, -1), Mm.Info(FileSize ".."))
   assert_filesize_fails("A:/")
   assert_filesize_fails("A:\")
   assert_filesize_fails("C:/")
   assert_filesize_fails("C:\")
   assert_int_equals(-2, Mm.Info(FileSize str.replace$(TMP$, "/", "\")))
   assert_int_equals(-2, Mm.Info(FileSize str.replace$(TMP$, "\", "/")))
 EndIf
End Sub

Sub assert_filesize_fails(f$)
 On Error Skip 1
 Local size% = Mm.Info(FileSize f$)
 assert_raw_error("Invalid file specification")
 On Error Clear
End Sub


Best wishes,

Tom
matherp
Guru

Joined: 11/12/2012
Location: United Kingdom
Posts: 11615
Posted: 02:29pm 13 Jan 2023      

The Picomite behaves the same as the CMM2. Sets the error message if a directory but does not throw an error
MMB4W is slightly different so should probably change but will still return -2 for a directory. It only errors if the filename ends in / or \
Edited 2023-01-14 00:33 by matherp
thwill

Guru

Joined: 16/09/2019
Location: United Kingdom
Posts: 4370
Posted: 02:34pm 13 Jan 2023      

  matherp said  The Picomite behaves the same as the CMM2. Sets the error message if a directory but does not throw an error
MMB4W is slightly different so should probably change but will still return -2 for a directory. It only errors id the filespace ends in / or \


OK, I had not noticed this. Can I officially go on the record as saying that setting MM.ERRMSG$ outside of throwing an ERROR completely breaks "The Principle of Least Surprise" ?

   

If I can be so bold, "least suprise" suggests that:

- If MM.ERRMSG$ <> "" then an error has been thrown ... and the CMM2 should not be setting it to "Success" which it sometimes does.
- If MM.ERRNO <> 0 then an error has been thrown ... I think this is always the case.
- At all times either (MM.ERRMSG$ <> "" AND MM.ERRNO <> 0) OR (MM.ERRMSG$ = "" AND MM.ERRNO = 0)

Best wishes,

Tom
Edited 2023-01-14 00:39 by thwill
matherp
Guru

Joined: 11/12/2012
Location: United Kingdom
Posts: 11615
Posted: 02:43pm 13 Jan 2023      

As with all this stuff I think it is only you who ever looks at errormsg or errno so completely irrelevant for most users
thwill

Guru

Joined: 16/09/2019
Location: United Kingdom
Posts: 4370
Posted: 02:46pm 13 Jan 2023      

  matherp said  As with all this stuff I think it is only you who ever looks at errormsg or errno so completely irrelevant for most users


That does appear to be the state of affairs.

Best wishes,

Tom
Pluto
Guru

Joined: 09/06/2017
Location: Finland
Posts: 411
Posted: 02:48pm 13 Jan 2023      

Tried to download Picomite(VGA) V5.07.07b1. When extracting: "Archive corrupt!"
JohnS
Guru

Joined: 18/11/2011
Location: United Kingdom
Posts: 4345
Posted: 02:51pm 13 Jan 2023      

  Pluto said  Tried to download Picomite(VGA) V5.07.07b1. When extracting: "Archive corrupt!"

Same for me.

John
JohnS
Guru

Joined: 18/11/2011
Location: United Kingdom
Posts: 4345
Posted: 02:56pm 13 Jan 2023      

  Volhout said  
  matherp said  Picomite(VGA) V5.07.07b1

PIN(FLEVEL pio [,sm, DIR]) ' dir can be RX or TX. If specified gives the level of the specific fifo


I Guess this is a typo....

PIO(FLEVEL pio [,sm, DIR])

Regards,

Volhout

In that it shouldn't be
PIN(...
but instead
PIO(...

Does look like a typo.

John
matherp
Guru

Joined: 11/12/2012
Location: United Kingdom
Posts: 11615
Posted: 03:38pm 13 Jan 2023      

It's amazing how you can't see your own errors but everyone else can  

PIO(FLEVEL pio [,sm, DIR]) ' dir can be RX or TX. If specified gives the level of the specific fifo

  Quote  Tried to download Picomite(VGA) V5.07.07b1. When extracting: "Archive corrupt!"


Please try again
Edited 2023-01-14 01:44 by matherp
matherp
Guru

Joined: 11/12/2012
Location: United Kingdom
Posts: 11615
Posted: 04:23pm 13 Jan 2023      

Forgot to mention also in b1

The PWM command is modified as follows:

PWM channel, frequency, [dutyA] [,dutyB] [,phase] [,deferredstart]

Phase is a new parameter that causes the waveforms to be centred such that a wave form with a shorter duty cycle starts and ends equal times from a longer one. Use 1 to enable this mode and 0 (or omit) to run as normal

The new parameter "deferredstart" when set to 1 prepares the PWM channels as before but does not start the output running. This is to avoid any undesirable startup artefacts

There is a new sub command

PWM SYNC [channel0offset] [,channel1offset] .......[channel7offset]

This initiates the PWM on channels where a deferred start was defined or just syncs exisitng running channels. However, the power comes in the ability to offset the channels one to another (defined as a percentage of the time period as per the duty cycle - can be a float)
You can use an offset of -1 to omit a channel from the synch
Edited 2023-01-14 02:44 by matherp
JohnS
Guru

Joined: 18/11/2011
Location: United Kingdom
Posts: 4345
Posted: 06:33pm 13 Jan 2023      

Download now OK, thanks.

John
Bleep

Guru

Joined: 09/01/2022
Location: United Kingdom
Posts: 817
Posted: 08:46pm 13 Jan 2023      

  matherp said  PIO (SHIFTCTRL push_threshold [,pull_threshold] [,autopush] [,autopull] [,in_shiftdir] [,out_shiftdir] [,FJOIN_RX] [,FJOIN_TX]) 'additional optional parameters to create 8-word RX or TX FIFOs

NB: RX and TX are from the Basic code perspective


Hi Peter,
I've just tried using the FJOIN_TX, this is for my Basic code sending out data to the FiFo for transmission by the PIO, however I've found I actually have to use:-
f0=Pio(shiftctrl 0,0,0,1,0,0,1,0) 'Set Auto Pull at 32bits (0) & Extend Tx  FiFo

Which would seem to me to be using FJOIN_RX, so either I'm not understanding the TX RX direction, or I'm doing something wrong. If I actually set FJOIN_TX my code locks up, presumably because there is no FiFo for it to write to.
Regards, Kevin.
Volhout
Guru

Joined: 05/03/2018
Location: Netherlands
Posts: 5991
Posted: 11:31pm 13 Jan 2023      

37us per word on fifo is pushing the limit of what mmbasic on fifo can do.
Interrupt gives you freedom for your main code
phil99

Guru

Joined: 11/02/2018
Location: Australia
Posts: 3319
Posted: 11:49pm 13 Jan 2023      

Edit
Ignore, this new Pico has faulty GP0. Does not drive low, just goes high Z. Adding a pull-down resistor enables normal DOUT to work, adequate on PWM. Has 16MB flash yet cost less than regular Pico, now I know why!
On an old Pico it works well.

Playing with the new PWM commands.
PWM channel 0 phase shift seems wrong. This three phase sequence works perfectly if using PWM channels 1,2,3 but not with PWM channels 0,1,2

> LIST
' Three phase motor drive - not quite right

SetPin gp0,pwm  'CH 0a
SetPin gp1,pwm  'CH 0b
SetPin gp2,pwm  'CH 1a
SetPin gp3,pwm  'CH 1b
SetPin gp4,pwm  'CH 2a
SetPin gp5,pwm  'CH 2b

PWM 0, 50, 33.33, -66.66, 1, 1
PWM 1, 50, 33.33, -66.66, 1, 1
PWM 2, 50, 33.33, -66.66, 1, 1

PWM sync 0, 100/3, 200/3
>

> LIST
' Three phase motor drive - perfect

SetPin gp2,pwm  'CH 1a
SetPin gp3,pwm  'CH 1b
SetPin gp4,pwm  'CH 2a
SetPin gp5,pwm  'CH 2b
SetPin gp6,pwm  'CH 3a
SetPin gp7,pwm  'CH 3b

PWM 1, 50, 33.33, -66.66, 1, 1
PWM 2, 50, 33.33, -66.66, 1, 1
PWM 3, 50, 33.33, -66.66, 1, 1

PWM sync 0, 0, 100/3, 200/3
>

Edited 2023-01-14 15:02 by phil99