Version history for TTXHelper
---------------

Early versions will have an 'a', 'b', or 'c' suffix. This does NOT appear in
the module version string - go by the module date and the file datestamp...
Additionally, very early versions will say the full version number at the END
of the module code. Load it into an editor...


0.00  2004/02/07

      Initial version of the module. It doesn't do an awful lot really. :-)



0.00a 2004/02/08

      "ScanFlash" (easy), "ConvertToISO" (easier), and "ExtractLine".



0.00b 2004/02/09

      I have to stay up ALL night to turn the water on every 45 minutes. We
      are on a pump (from a well; we use it for bathing/washing, we don't
      drink it). The pump is liable to freeze when it is -7C outside, so I
      turn the water on in the assumption that the operating of the pump will
      destroy any ice forming inside the pump. It isn't covered by our
      insurance for what seems, to me, to be a pile of excuses...

      Anyway, it's time for some serious coding.

      "UseInterface", "ReadBytes", "WriteBytes", "ReadRegister", 
      "WriteRegister", "ReadChapter", "ClearTSDP".
      I'm not entirely convinced that the read/write register commands work
      as they should. You can poke data directly into memory, and read it
      back, but it seems the lower (setup) registers return odd results...
      Please let me know if this is what you encounter.

      "AutoDetect" searches for a teletext device on the three major
      interfaces that are supported.



0.00c 2004/02/10

      The rest of "AutoDetect"...

      Stupid! Stupid! Stupid! Always ALWAYS remember that SWI calls in
      modules occur in SVC mode, hence R14 is buggered up, and the tail
      effect of calling the corrupted R14 can be... interesting. :-)
      Also, only a total asswipe (I sheepishly raise my hand) would insert
      debugging statements in the MIDDLE of a frickin' conditional, viz:
        CMP    something, whatever
        MOVLT  return, some_value
        SWILT  256 + '!'
        BLT    my_exit_code
      I'm not even going to point out the error, it is so obvious that I feel
      I should get in the bath right now (yes, fully clothed) and do
      something useful, like, drown... :-) Duh! Duh! Duh! Duh! Duh!

      The documentation is NOT correct when it says R0 is 4 for RiscTV's
      teletext. It will attempt to match the RiscTV hardware to a known
      'Ground Control' decoder type (usually the UTA-3). You can tell it is
      a RiscTV device because R1 will be 2.



0.01  2004/02/11

      Wrote test software. Fixed a few bugs that came to light. Implemented
      two "undocumented" SWIs:

        SWI &16FC6 (no SWI name; uses unofficial SWI chunk)
          Entry:
            R2 = Language to set character set to; 0-7 ISO, 8-15 ANSI
          Exit:
            -

        SWI &16FC7 (no SWI name; uses unofficial SWI chunk)
          Entry:
            R0 = Character, in teletext character set
          Exit:
            R0 = Character, converted to ISO/ANSI character set
                 Control codes and mosaic status ignored.

      The SWIs above are included to facilitate character set testing. The
      SWIs may be renumbered, or completely removed, in future versions of
      the TTXHelper module. DO NOT rely on their presence...

      A 'tested' initial version, several bugfixes.

      All SWIs except "TTXHelper_ConvertToANSI" are now implemented.

      Emailed an allocation request for an official SWI chunk.



0.01a 2004/02/12

      No allocation yet - I cannot pick up email *that* often to check! :-)

      Fixed bug with mosaic testing in ISO conversion, now it does The Right
      Thing instead of blanking lower case letters.

      Now calls its own SWIs for read/write operations. The code is 'safer'
      than 'loading the stack' to set up return addresses; though that method
      did work just fine (and is still used in a few cases). If nothing else,
      it saves three instructions in the executable (but probably eats a few
      thousand in the SWI despatch routine!). Can't win 'em all...



0.01b 2004/02/13 [~3am in the morning]

      Added three *commands:

        TTXHelper_Licence
          Just tells you that you should read the licence file to find out
          about using/distributing the module, and gives the URL for my
          teletext software.

        TTXHelper_Probe
          Attempts to auto-probe for an attached teletext receiver. If it
          finds one, it'll report on it.
          Note that the order of scan is system->PtrIIC->RiscTV (assuming
          drivers are available for each).
          Note also that it'll report the FIRST device found, so if you have
          multiple teletext receivers on multiple devices, it won't probe
          them...

        TTXHelper_View <filename>
          A simple switch-to-MODE-7-and-look viewer for saved teletext
          frames. This view method reads the first '24 rows' of the file, so
          it should work with any teletext software that puts the data first,
          without any kind of header.
          Alternatively, you might find sprites of Alyson Hannigan look
          pretty when interpreted as teletext data. Or, failing that, how
          about some executable or other? TTXHelper isn't fussy, as long as
          it can read 1000 bytes of data (that's the teletext standard 24
          rows with 40 coloumns...), it'll display practically anything to
          you. Interpreted as teletext, of course! :-)



0.01c 2004/02/13 [~8pm at night]

      *TTXHelper_View now does not allow you to view teletext frames in a
      taskwindow. It *is* possible using !GraphTask, but then, you could just
      load up !Teletext...
      This command is intended for (true) command-line use.

      Tidied the output of *TTXHelper_Probe slightly.

      *TTXHelper_Licence is now a command. The command prints out the same
      stuff you'd see if you tried:
        *Help TTXHelper_Licence
      I have done this because RISC OS requires *Help for help (only) text,
      and you tying *TTXHelper_Licence would be met with:
        File 'TTXHelper_Licence' not found
      which is not so helpful...
      [okay, it isn't 100% identical, it skips the "Help on keyword" bit]



0.02  2004/02/17

      Picked up official SWI chunk allocation yesterday, so recompiled with
      the correct SWI. Nothing else altered, but version number changed as
      this module may be released (and "0.01c" is potentially confusing)...

