Geneve OS development discussion

[+InsaneMultitasker]

I will try to locate the keyboard XOP test code that I wrote as part of my attempt to code the keyboard "remap" routines.  I had tried to mimic the extended character set data entry method that uses ALT and the 3-digit character code, for example, typing ALT-0233.

 

Since the logo is displayed only at powerup and a warm reset, characters 0-31 could be used if a routine is available that will print them, or we need to devise an escape method that won't interfere with other printing. Something for us to think through.

[+InsaneMultitasker]

2 hours ago, mizapf said:

Is IBMGRF equal to codepage 437?

Yes.    https://en.wikipedia.org/wiki/Code_page_437 

Historically... the intent for MDOS in the 90s was to natively support what we called "ANSI Graphics" within the MDOS environment. IBMGRF was a stand-alone program and later was added to the OS.  VCLR was a fork of my terminal emulation code for MDOS use; @9640News has integrated and enhanced this code for XOP usage.

 

In the early 90s, Funnelweb only indirectly supported a subset of the extended characters.  This prompted me to create my ANSI-TOOLS program to generate compatible files for BBS usage.  Later, the McGoverns added full 8-bit character support which was quite an improvement, IMHO.  I really liked the Funnelweb development tools (editor/assembler/loader) back in the day.  but I digress.

 

[+mizapf]

1 hour ago, InsaneMultitasker said:

Yes.    https://en.wikipedia.org/wiki/Code_page_437

Footnote f of the table confirms the story about ß and β.

[+InsaneMultitasker]

GPL has been updated to v7.40 and is being tested.   @9640News and I will discuss whether or not to publish the source code when the new "GPL" file is released.  I haven't decided whether to update PLAYE.

 

For those waiting for the next OS update I will loosely paraphrase George R. R. Martin by saying, "An MDOS update is coming", and I expect it much sooner than the final novels of his fantasy series...

 

V7.40  12 March 2023
       - Added command line options GPL /R#V filename (CASE SENSITIVE)  @hloberg idea, thank you
         R - Rompage, with autostart or first entry
         # - speed 1-5
         V - 9938 Vr0 support for GPL VWTR (Scott Adam's Adventure 80 col)
         ? - help
       - Changed VR0 mask to support 80 column mode via GPL video routines
            use option V
       - Fixed speed issue identified by @mizapf; SPEED 1 did not work after
         starting the interpreter with either speed 4 or 5.
       - Cleanup in preparation for releasing the source code

 

 

 

[+9640News]

I am hoping we can release both GPL V7.40 and ABasic 4.09 this weekend.  I have been out of town for work for the past 8 days, and looks like I will not get home until Wednesday night.  Then, Friday is my local job/work and my daughter is flying in from Dallas in the evening for a late Christmas celebration and leaving Tuesday next week.

 

I should be able to find some free time in there somewhere.

 

MDOS 7.40 will come a bit later as Insanemultitasker and myself have a few things to resolve.  I discovered about 2 weeks ago some behavior in some new opcodes I added to the TIPI XOP I need to verify work as expected in some other test code.  I am hoping my test code is where the problem is and not the opcodes itself.

 

When MDOS 7.40 is released, there will be addendums to the GenProg manual as well as a new TIPI XOP section to cover the use of the routines.

 

CYA for MDOS 7.40 has already been updated, and hopefully, barring any significant changes in the future, it will be smart enough to update itself in future MDOS revisions.  This has been accomplished through a built-in table with memory locations and offsets for CYA to modify.

 

Later folks.  It has already been a 16+ hour day for me and need to get up in less than 8 hours.  At least the new product startup at the plant has been going well.  So well, the business manager wants to continue producing more  product before switching back to one of our other core products.

Later.

 

 

[+InsaneMultitasker]

On 3/13/2023 at 9:32 PM, 9640News said:

I am hoping we can release both GPL V7.40 and ABasic 4.09 this weekend

To upgrade your GPL to v7.40, replace the first file (typically GPL) with the attached file. 

 

See the readme files for details concerning the fixes, enhancements, and folder structure/MAKE process. 

 

@9640News has published the source to his github repository where it is now publicly available. 

 

TIFILES format:  GPL

Source and program files in TIFILES format (zipped): GPL0318B Source and bin distribution.zip

 

(The 80 column adventure cart is included because I was using it for test purposes; it will be removed from the next release package).

[+mizapf]

When I click on the GPL download link, why do I get a GPL.dsk file? This is not a sector dump but a TIFILES file.

[+InsaneMultitasker]

37 minutes ago, mizapf said:

When I click on the GPL download link, why do I get a GPL.dsk file? This is not a sector dump but a TIFILES file.

I just tried to download the link and there was no .DSK extension.  It is certainly a TIFILES format.

[+mizapf]

That is strange indeed. Does my Firefox fantasize a ".dsk" extension?

 

Edit: You are right, the Firefox in my Virtualbox does not add ".dsk". What the heck?

[+InsaneMultitasker]

55 minutes ago, mizapf said:

That is strange indeed. Does my Firefox fantasize a ".dsk" extension?

 

Edit: You are right, the Firefox in my Virtualbox does not add ".dsk". What the heck?

Did you by chance have a file called "gpl.dsk" in the target folder ? Maybe the dialogue was being too helpful. 

[+mizapf]

No, this may also be from the desktop environment (KDE/Plasma). It suggests the "dsk" suffix for the downloaded file. I remember that some years ago I registered a file type with dsk suffix (mainly for use with TIImageTool, wanted to launch it by clicking on the icon), but this does not seem to be the case here; at least this file type is not in the list anymore.

 

My settings said to store the downloaded file in my Downloads folder without asking, so it also added ".dsk" without asking. Still, I don't know where Firefox got this information from.

[+OLD CS1]

My Firefox does that stupid shyte, too.  For some reason, it has become fixated on .vgm, so every file I download without an extension gets .vgm appended.  It is just Stupid Programmer Tricks over at Mozilla, along with killing off the old download system.

[+Schmitzi]

Also, if I rename the file to "no-suffix" at the download prompt, Firefox will give it a .DSK

So I always have to go to the Explorer and rename it...

 

[+mizapf]

It's just above me why devs believe such things must be in the browser ...

 

Firefox changes file extensions:

 

https://support.mozilla.org/en-US/questions/1319659?page=2

https://support.mozilla.org/de/questions/1318447

 

There are so many things probably deemed to increase "comfort", but they only increase my level of adrenaline to dangerous heights.

[+InsaneMultitasker]

Maybe we can add similar functionality to Geneve OS. 

 

[+InsaneMultitasker]

The Geneve's DSR filename limit is 40 characters for the device+subdirectories+filename, the same limit imposed by the Myarc HFDC and carried through to other devices.  I have identified the buffer and DSR code locations to update to increase the DSR limit. This would mean the ability to use additional subdirectory levels and improved "compatibility" with devices like the TIPI.  The old limits would apply if ROMPAGE is in effect.

 

The command line interpreter and the parse utility XOP also adhere to the 40 character limit.  My concern/question is whether programs can recover from a parsed name that exceeds 40 characters; this depends on whether most programs rely upon the XOP to limit the length, or the program itself checks it.  I'm putting this out here for discussion and contemplation.

[+9640News]

58 minutes ago, InsaneMultitasker said:

The Geneve's DSR filename limit is 40 characters for the device+subdirectories+filename, the same limit imposed by the Myarc HFDC and carried through to other devices.  I have identified the buffer and DSR code locations to update to increase the DSR limit. This would mean the ability to use additional subdirectory levels and improved "compatibility" with devices like the TIPI.  The old limits would apply if ROMPAGE is in effect.

 

The command line interpreter and the parse utility XOP also adhere to the 40 character limit.  My concern/question is whether programs can recover from a parsed name that exceeds 40 characters; this depends on whether most programs rely upon the XOP to limit the length, or the program itself checks it.  I'm putting this out here for discussion and contemplation.

Any MDOS program requesting a filename at a prompt, I would expect to have a 40 character maximum filename input.  So, those type programs would not crash as you could not enter a longer filename, rather, if someone had a file at a path at exceeding 40 characters, they would not be able to type the filename in.  There is one program that would have an issue with longer filenames which was just debugged to handle the 40 char filepaths, GDM2K.  That is the only program I am aware that would have an issue.  The GPL version of that program would have the same issue as well.

 

So, I am not entirely sure about expanding the filepath/name.  Yes, it has its benefits, however, any TI or Geneve program as of this date is pretty much capped at the 40 char limit.

 

My initial raising the concern was the JSON capability would extend filenames to beyond 40 characters and @jedimatt42 responded writing to PI.CONFIG for a longer URL was a workaround solution to achieve the same results.  

 

Anyways, mixed thoughts on it at the moment.

 

 

[+InsaneMultitasker]

Just to clarify one point:  I am not considering any changes related to pathname length for the upcoming release.  I do see benefits beyond JSON, such as being able to back up a folder without bumping into the 40 character limit.  Regardless, we both agree the next step is to release MDOS 7.40 as there are fixes and enhancements that need to get out into the wild.

35 minutes ago, 9640News said:

So, I am not entirely sure about expanding the filepath/name.  Yes, it has its benefits, however, any TI or Geneve program as of this date is pretty much capped at the 40 char limit.

A path length greater than 40 characters is possible if the device allows it, e.g., TIPI.  I know there are programs that allow longer than 40 character paths outside of BASIC/XB but they might be more of an exception.  Most of us take care not to exceed 40, so I'm not really sure.  This too is a reason I wanted to at least have some discussion about it here, without the pressure of making it happen for the upcoming release.

[+9640News]

38 minutes ago, InsaneMultitasker said:

A path length greater than 40 characters is possible if the device allows it, e.g., TIPI.  I know there are programs that allow longer than 40 character paths outside of BASIC/XB but they might be more of an exception.  Most of us take care not to exceed 40, so I'm not really sure.  This too is a reason I wanted to at least have some discussion about it here, without the pressure of making it happen for the upcoming release.

Hopefully, others will chime in with their thoughts!!!

Speak up people!!!

[+dhe]

I'm not sure how much folks care about docs....

 

It seems odd that the number of subdirectories is controlled not by say X deep, but by total length of the string.

 

mkdir a.s.d.c.a.d.K.c.a.x = Ok.

 

but

 

mkdir DIRECTORY.DIRECTORY.DIRECTORY.FILENAME < Done.

 

Then, if someone has a perfectly good path like the first example, and then someone goes in and changes K to directoryK you might explode everything further down the line.

 

Looks like dragons to me.

 

[+InsaneMultitasker]

On 4/17/2023 at 4:39 PM, dhe said:

It seems odd that the number of subdirectories is controlled not by say X deep, but by total length of the string.

Agreed. Single character subdirectories will maximize the depth, though at the expense of knowing what in the world you are looking at. Unless you like single-letters.  

 

On 4/17/2023 at 4:39 PM, dhe said:

Then, if someone has a perfectly good path like the first example, and then someone goes in and changes K to directoryK you might explode everything further down the line.

This is certainly a possibility with the current maximum length.  You can inadvertently lock yourself out of files doing this; hmm, this could double as a simple protection method for those who don't know the file system. 

 

 

[+OLD CS1]

On 3/21/2023 at 12:57 PM, mizapf said:

It's just above me why devs believe such things must be in the browser ...

 

Firefox changes file extensions:

 

https://support.mozilla.org/en-US/questions/1319659?page=2

https://support.mozilla.org/de/questions/1318447

 

There are so many things probably deemed to increase "comfort", but they only increase my level of adrenaline to dangerous heights.

A lot of devs are not concerned with the user experience writ large, only with what they perceive to be a good (or proper) experience.  We need more Tom Smykowskis in this world.  Some devs I would just like to throw into a well and stare down at them saying, "It listens to the user request or else it gets the hose, again."

[Reciprocating Bill]

1 hour ago, OLD CS1 said:

A lot of devs are not concerned with the user experience writ large, only with what they perceive to be a good (or proper) experience.  We need more Tom Smykowskis in this world.  Some devs I would just like to throw into a well and stare down at them saying, "It listens to the user request or else it gets the hose, again."

Maybe interface choices through skinning would be helpful?   

[+InsaneMultitasker]

On 3/13/2023 at 9:32 PM, 9640News said:

MDOS 7.40 will come a bit later as Insanemultitasker and myself have a few things to resolve.  I discovered about 2 weeks ago some behavior in some new opcodes I added to the TIPI XOP I need to verify work as expected in some other test code.  I am hoping my test code is where the problem is and not the opcodes itself.

 

When MDOS 7.40 is released, there will be addendums to the GenProg manual as well as a new TIPI XOP section to cover the use of the routines.

Short update for the group:

 

@9640News and I have reconciled (sync'd) our OS code base. For my part, I leveraged Git desktop to identify differences.  With exception of the CYA updates, we found one video XOP change for ABASIC that had been forgotten about.  I'm looking forward to the ?W work that is happening in the TIPI-9640 dev topic, as that could make Git a more feasible option for some Geneve OS development.

 

Additionally, I have started to pull together source for a few OS-specific items such as WIPE, the CRC detection utility, and PFM utilities. Maybe drivers can be added to this list, like mouse, ANSI and Windows.

 

We've drawn the line in the sand and are now working to understand whether there is anything truly needed before we release 7.40.  Getting close now...

[+InsaneMultitasker]

I was driving the other day and we were behind this white Chevy.  The plate caught my attention; I chuckled at the thought (albeit unlikely) that I might be following the author of our Geneve's Infocom interpreter.