mg_tst2.$b [1/1]

Somehow Saturday June 19, 1999 at 11:25:35 a conflict occurred
Renat Mamedov and ALL were arguing on the topic mg_tst2.$b [1/1]
I decided to add heat to this dialogue

RM> Posting this in zx.spectrum did not give any results,
RM> obviously it doesn’t work or works for cheers.

RM> This is a simple test to check the memory driver
RM> that I wrote for Homer2. Finds 48,128 ...
RM> extended: pentagon, pro, scorpion, atm.
RM> Cars with incomplete lines and gaps should also
RM> be identified and counted correctly...
The test seems to be correct, but after your test it completely demolishes both the resident
Honey Commander and ruins the ram-disk :((
There is a small suggestion for to this test: before changing anything in
pages, save the value of the RAM cell, test, and then restore the cells. And
in general, if you want, I can throw the source code of the memory test from TEST 4.21.

RM> Just run it and look at the screen.
Profi 1024. ->>>>>>>>>>>>>>>>>>>>> :))))

RM> ps. Can you see me or am I a myth? or does it not leave this address? :(
You will become a myth in 200 years, and not sooner :).

Long life and prosperity to you, Renat!

With best wishes, Musa Khaybullin.

June 20, 1999, Musa Khaybullin writes to Renat Mamedov:

[...]

MK> The test seems to be correct, but after your test it completely demolishes both the
MK> resident of Honey Commander and spoils the ram disk :(( There is a small

Test after one byte at the address #FFFF.

MK> suggestion for this test: first than to change anything in the pages,
MK> save the value of the RAM cell, protest, and then restore the cells. A

No need. When you start the game, be so kind as to play, and don’t
fiddle with the ramdisks and residents :)

MK> in general,
MK> if you want, I can throw in the sources of the memory test from TEST. 4.21.

Maybe my little hands are crooked, but not so much as to use other people's
procedures. The main thing here is not the implementation, but the technique.
And the technique is the following - the machine is determined, then the last byte is filled with zero in all memory, then the pages are sequentially sorted through,
when searching, the checked pages are marked in the driver as busy,
and in order to prevent intersections - Before marking the page, we check the
last byte for zero. Theoretically, this system is not breakable within a
megabyte. That’s what I was interested in in practice.

22 Jun 99 19:43, Renat Mamedov wrote to Musa Khaybullin:

RM> we check the last byte for zero. Theoretically, this system is not
RM> breakable within a megabyte. So I was interested in practice.

And in practice, on a faulty computer there will be a glucodrome.

Kirill Frolov. [ZX]

KF> But in practice, on a faulty computer there will be glucodrome.

Mr. Frolov! Reading your statements sometimes you doubt even your own mind... subj - checking the memory driver for the game...
Glitchy computers and crooked users don’t interest me,
although I did everything possible in my opinion so that problems would not arise.

And even more so, the fate of the owners of faulty machines does not concern me.
Although it may be original - checking the health of the computer with a game :)

With Best regards,
Renat

My on-board systems found that on Tuesday June 22, 1999 19:43, Renat
Mamedov wrote to Musa Khaybullin:

(_____________SKIP____________)

RM> No need. When you start the game, be so kind as to play, and not
RM> worry about ramdisks and residents. :)
Your game is unlikely to take up more than 450Kb, so only your test will ruin Honey Commander. And if the test worked smartly :), then everything would work perfectly and not conflict. Now, IMHO, there are too many residents
divorced for them to be ignored just like that.

MK>> in general, if you want, I can throw in the source code of the memory test from TEST 4.21.

RM> Maybe my little hands are crooked, but not so much as to use other people’s
RM> procedures. the main thing here is not the implementation, but the methodology.
RM> And the methodology there is the following - the machine is determined, then the entire memory
RM> is filled with the last byte with zero, then the pages are sequentially searched through, during the search, the checked pages are marked in the driver as occupied, and in order to prevent intersections - before marking the page

MK> Why reinvent the wheel :). IMHO, it is better to spend time on

this lisaped was invented back in the Stone Age :) there was simply no need to do it.
And despite the fact that this is a fairly reliable method, many
ignore it. So we see 512 (we have 384 pros (ru7 line and ru5 line)),
meg on 768, meg (!) on torn 512 and so on...

MK> something else, for example, improve the sound quality, make support for
MK> GS and
MK>, etc.

the sound, by the way, sucked. I am pleased with its quantity and direction
(in most cases it’s my squeaky voice) :)

And to support the GS - I don’t have it and don’t need it.
And in general, I fed this girl, I’ll dance with her (anecdote).
Pay the money...

MK>> only your test will destroy Honey Commander. And if the test
MK>> worked smart :), then everything would work perfectly and not
MK>> conflict. Now, IMHO, too many residents have divorced
MK>> for them to be easily ignored.

RM> Stop. Game. If you want to play, play. If you don’t want to, copy until
RM> you’re blue in the face in Khonich...
That’s what I do around the clock :)))))) Without rest or smoke breaks :)))
RM> In the end, no one is forcing you...
That’s for sure :))
RM> And regarding smart tests (I adopted this stone as my garden :)
RM> - look how yours The test finds memory on pros with torn
RM> memory. As far as I know (testers reported :) - there are problems there...
Describe which ones, please!

MK> Please describe which ones, please!

I willingly do this from the words of Alex Kalinushkin. He is the only one who
took this test as soon as Mona.

=== I quote the file TEXT.TXT ===

Numbers simm your test TEST4.20 HONEY
1 2 3 4
-+-------------------------------------------------------------
+ - - - 256 256 256
+ + - - 512 512 512
+ + + - 768 1024, unused 512 1024
+ + + + 1024 1024 1024
+ - + - 512 256 256
+ - + + 768 1024, unused 256 1024
+ - - + 512 1024, unused 256 1024
+ + - + 768 512 512
-+-------------------------------------------------------------

=== End of quote ===

He has 4 sims on his computer for 256k each, so he did various perversions with their
presence. The words 'your test' should be understood as the test of your humble servant...

Someone may exclaim that there are no such computers. This is not so.
For example, in professional 4.xx boards, it seems that there is a penultimate
configuration with two ru7 lines installed.

There are more exotic options - for example, I have twice seen computers with
two ru7 lines and another ru5 line (color CP/M)...
In this case, memorywe have 576k, but we see it differently everywhere :)

Ha! I just discovered that Subzhik doesn’t understand Turba 2+!
Where is this good ;-).

For this Au Revoire, CompoWellcome.

July 02, 1999, Musa Khaybullin writes to Renat Mamedov:

MK> Ha! I just discovered that Subzhik doesn’t understand Turba 2+!
MK> Where is this good ;-).

Cool! :) since I installed turbo-1 at the urgent request of the moderator,
there can be no question of turbo-2. Yes, we're already tired of these standards.
Drag a bunch of drivers behind you... Well, I don't have the desire/opportunity to tinker
with a soldering iron - it's okay, it'll just slow down. And great :)

Sincerely,
Renat