GitLab

Added STB5Dtop tags directory .. allows rhenium build with NTSC link in place to react as if the link were removed. Hence can be used PAL with appropriate *configure.

Detail:
Admin:


Version 0.28. Tagged as 'PortMan-0_28'

Added SmartCard_5V tag for STB55. Not added to STB5 - it does have a 5V/3V control line, but only for the second slot which we're not using.

Version 0.27. Tagged as 'PortMan-0_27'

Version 0.26. Tagged as 'PortMan-0_26'

Version 0.25. Tagged as 'PortMan-0_25'

Version 0.24. Tagged as 'PortMan-0_24'

Added button support for STB55. Non-trivial, as the buttons are implemented as a 2x4 matrix, using 4 of the LEDs as the column outputs.

Version 0.23. Tagged as 'PortMan-0_23'

Version 0.22. Tagged as 'PortMan-0_22'

Version 0.21. Tagged as 'PortMan-0_21'

Version 0.20. Tagged as 'PortMan-0_20'

Detail:
  Files modified: Resources.UK.Messages - Corrected typo in non-existent tag message
  Files added: Resources.STB.Tags
Admin:
  Updated from Kev's machine by Mike H.

Version 0.19. Tagged as 'PortMan-0_19'

Detail:
  The function which was returning a handle for a file in ResourceFS
  was passing &fh into a SWI, rather than fh. This would result in
  the inspected file being left open.
Admin:
  Tested by DEllis.

Version 0.18. Tagged as 'PortMan-0_18'

Summary:
  - Only rescans resource files if resourcefs handle changes
  - Disregards invalid tokens in tags file
Details:
  Since the tags file is cached when initialised, there's no point doing this
  if the actual data hasn't changed, so the resourcefs internal file handle
  is now recorded, which is a constant for a given file in resourcefs, only
  changing if the file itself changes.  This avoids unnecessary recaching of
  the data.

  There are 2 invalid entries in the STB3 messages file which musn't be
  cached.  These are now skipped.
Admin:
  Tested on a RiscPC with the hardware writes commented out.  There are 2
  BASIC programs in the test directory which exercise the module.

Version 0.17. Tagged as 'PortMan-0_17'

Service call handler now present in non-standalone builds so that new caching facility will read messages file is resourcefs changes.

Detail:
  To provide identical behaviour to previously, the messages file should
really be reparsed upon receipt of Service_ResrouceFSStarted.

Admin:
  Tested on a STB-400.


Version 0.16. Tagged as 'PortMan-0_16'

  - AddCallback in finalise corrected to RemoveCallback
  - RAM rule added to makefile for making ROM replacement

Detail:
  Customer I reported that the STB-3 was suffering from IRQ reentrancy problems.
Some of these were tracked down to PortManager calling MessageTrans_LookUp
(when IRQs were threaded), which can cause file operations after
Service_ResourceFSStarting has been issued, invalidating the MessageTrans FD.

  PortManager now caches the contents of its Tags file upon startup in a
linked list, which it searches instead of using MessageTrans.

  During testing, it was also found that crashes would often occur when
killing the module.  This was tracked down to an erroneous OS_AddCallback in
the finalisation handler.

Admin:
  Tested on STB-3.  Rentrancy problems during booting fixed.  The module
should be loaded _very_ early on in the boot sequence.  Preferrably a the
module should be launched from a utility which peeks the current softcopy of
the port bits, and programs the module with these as it's initial value for
the port.


Version 0.15. Tagged as 'PortMan-0_15'

Detail:

  * Only a single file exists: "Tags".  This has been copied from the
    Lazarus resources.  (I doubt that this describes the Customer L H/W
    adequately at all!)

Admin:

  * Required for Customer L demonstrator build.

Version 0.14. Tagged as 'PortMan-0_14'

Detail:
  Portman can't use bbe-generic - local mods made to makefile to ensure
  correct resources are exported.
Admin:
  Change only affects BBE.

Version 0.13. Tagged as 'PortMan-0_13'

Version 0.12. Tagged as 'PortMan-0_12'

Version 0.11. Tagged as 'PortMan-0_11'

Detail:
   Customer F.Tags:  Changed Chrontel_Power to 1:0:2 from 1:0:3
Admin:
   Tested by soft-loading from Zip drive. Works OK on NTSC TV.

Version 0.10. Tagged as 'PortMan-0_10'

Detail:
	On my oh-so-lovely straight-out-of-CVS aren't-I-A-Buildmaster-Genius
Customer F 4 build the smartcard and LED's were not working. It turns out that the
Port Manager settings for the LED's were wrong, hence preventing the
smartcard from also working. Kind Kev fixed this.

Admin:
	Tested on a Customer F 4 build out of CVS.
	Need some caffiene. Urgently.

Version 0.09. Tagged as 'PortMan-0_09'

Version 0.08. Tagged as 'PortMan-0_08'

Version 0.07. Tagged as 'PortMan-0_07'

Version 0.06. Tagged as 'PortMan-0_06'

Comment before Dynamic dependencies line at end was confusing AMU, causing it do add the line again, which in turn confused it at a later stage...

Detail
    Makefile was trying to get Messages file from Resources.<Locale>, but they
    were in Resources.<Machine> - added Resources.<Locale> and moved Messages
    in there.

    Changed to use LocalRes$Path for Messages.

    Added Resources.Peregrine and copied STB3 tags in there, to support STB22
    (as far as I know, it's the same, and what we've been using so far...).

    Added install phase to Makefile to support Expansion card builds.
Admin
    Built as part of ATM25 Expansion card build for Peregrine (STB22).

Version 0.05. Tagged as 'PortMan-0_05'

Version 0.04. Tagged as 'PortMan-0_04'

Version 0.03. Tagged as 'PortMan-0_03'

Added resources phase to Makefile (oops) and sends a service call round when we're starting and dying.

Version 0.02. Tagged as 'PortMan-0_02'

Version 0.01. Tagged as 'PortMan-0_01'