Hey Steve,
Well, at least with the testing I did using DATA / READ A$, I was able
to put low codes in the hidden lines with no issues. I printed A$ and
it simply skipped the low codes during print, however when I printed
LEN(A$), it showed the expected length, including the low characters.
Perhaps there are modes where it works and other where it has issues.
How did you test this when you saw the bad behavior?
Ken
On 6/5/18 9:13 AM, Stephen Adolph wrote:
Ken for some reason I did not get your note, but I saw it in John's
response.
Are you sure?
If the program attempts to execute the hidden BASIC lines as BASIC, as
you demonstrate, then codes <32 will kick off major problems. I
believe I confirmed that this AM by manually changing a line to
include low codes.
But I could be wrong ;) -- I rarely feel the need to question you Ken,
so go easy on me! ;)
On 6/5/18, Ken Pettit <[email protected]
<mailto:[email protected]>> wrote:
> Hey Steve,
>
> A minor edit to your summary. For the hidden lines, you don't
need to
> avoid codes < 32 decimal, other than 00h, quote and comma as
identified,
> though they are harder to edit if you do use codes < 32.
>
> Ken
>
> On 6/5/18 4:31 AM, Stephen Adolph wrote:
>> so, As a summary, I think this is what is new here:
>>
>> * Ken has demonstrated that XIP ML can be embedded in a basic
program
>> that can run anywhere
>> * John has illuminated the fact that there are hidden basic lines
>> * by using hidden lines, one can create a very compact hybrid
ML/BASIC
>> program that can safely be edited in BASIC
>> * hidden lines can be executed in BASIC or not
>> *****if not executed then RAW ML can be embedded - avoid code 00
>> *****if executed
>> ***********hide the ML in strings - avoid " and codes <32 decimal
>> ***********hide the ML in data structures - avoid comma, and
<32 decimal
>>
>>
>> We've already had Basic with embedded ML, we've just never had
it with
>> next to no overhead - both in time and memory - and with the
ability
>> to run in place WITHOUT absolute addressing.
>>
>>
>>
