@bean

Sad do say, but still the same solid freeze. I just do : reboot, run owb (do nothink in it), run tunenet, choice mp3, press play and after 1-2 seconds solid freeze of OS. Music not start to play, It's just start to think, and then freeze. Can do any tests for you to help catch the problem.

more about: freezes happenes even if i do reboot and run tunenet only.
Sometime it start to plays, some time freeze, sometime it show depacking info of module/mp3 in the tunenet (but nothink to play), sometime not.

Also what i noticed: before the freeze, you choice for example mp3. And just do nothink. Tunenet not start to plays, but, mouse cursor works, and when i move over the buttons of tunenet, it looks like works. Then, if i press on any place of any app (for example to drag the window of the tunenet, or drag the window of the owb if i run it) - then freeze.

But at moment, have not any noticed about "opening" crashes. Only these freezes.

@orgin

Only why i use Dopus sometime, it's becouse of "2 panel manager" which easy and logical to copy files from archive (i.e. unpack), and just do copy from some place to another (i still miss "Easy" copy stuff in the Filler. For example will be cool to just run 2 filler window and in one filler going inside the archive, select need it files, and just copy it to another window (or just drag&drop it)).

Filler run/exit programm correctly(current diropus4 are not, even if dopus have "async" flags, it's still not works like it must. Filler always run anythink (even bad one) and back control back), i like that i can in one little window see almost all my files and directoryes. Also will be cool to have "configurable keys" for all the actions, but.. it's all depend on your time and wish :) And the design and layot of filler looks logical/compact/modern for me. I can resize it as i want and so on (dopus a bit like "stone" here, you can control of course size of window, but only by setup, that is suck, becouse some time jsut need do fast resize to see somethink and filer won here again).

I made this kind of post on the amigaworld.net, but not all the ppls visit AW, so i also duplucate it here. Imho it can be interesting to someone:

I made a little video related to os4 for some friends but i think that video can be interesting to many other ppls, so, just make a topic for that.

Video can be splitted on 2 parts. First part (up to the 4:30 of time) showing 68k scene diskmags which works on os4.1 now becouse of new wrapper/libs done by Salas00 for the old ptreplay and medplayer librarys. You can check this topic for detect how it was done and which old 68k progs now works with it.

For firstly i run 3 (4,6,7) issues of russian diskmag called "poweramiga". It's a bit lame by the content, but good looking by gfx/music. Issues 4,5,6 uses "ptreplay.library". Issue 7 uses medplayer.library. Then next i run Excess8 and Excess9 diskmags by MadWizards, and after all the poland diskmag called Seven Days by LazyMinds.

The next part of video (started from the 4:30 of time), can be called as "os4 stability when you load it heavy". I just run seven days1 diskmag on own screen, then run Uprough's EP-0003 (lowtech EP) on port of which i works right now (this nice minimal music it's from this ep), then:
- Filler
- Wookiechat and going to irc.amigaworld.net
- DirOpus4
- OWB (going to amigaworld.net)
- ScummVM (running FullThrottle game)
- tunenet (play mp3 on it)
- FallingUP (a nice game ported by IAmSonic)
- Ltris game
- PIG game

Then battery of the camera start to be "low" and i stop the record. But for sure i can run somethink else over all of this.

Must to say, that some of these progs "hold" network connection all the time, some of these progs (many) use AHI (you can heard all of this at the same time).

So the result: for the os without memory protection, it's already _very good_. Some oses _with_ memory protection not have this kind of stability. I really enjoy by current state of os4 :)

Of course it's all mean runing only "good" programms. Which are writed good. Becouse all of this can be just die becouse of only one running bad programm. But in generall, when we works on somethink, we all the time uses stable programms, so for the reality it's make a sense.

ps. The video done on peg2/1gz/1gb of ram/radeon9250/latest os4.1 with latest quickfix.

@bean
Yeah, of course i am in interest about deep tests/beta testing. Write me at kas1e <blah> yandex.ru (kas1e, not kasle).

About your questions : xx50 version works just fine all the time without any problems. So answers is:

1. full path are different. For example i tryed 2 times: work:music/mp3/name_of_mp3_name.mp3 and work:mag1/data/music/mod_name.mod.

2. tryed also PT and AIFF. The same. Sometime crash, some time solid freeze (now i think that it's 99% open() kind fucntions).

3. i try different files. Just looks like problem not with "depacking" of file itself, or kind, but just with open it for the begining to works with that.

4. old tunenets works perfectly for me all the time and very stable. The problems start happnes with only new latest releases.

5. I tryed with skinfx, with docky, without skinfx, without docky, with graphical look, without - always the same bug.

6. In general i have always setuped default unit and unit 0 at 32 channels, etc.

Right now i just put old tunenet back (xx50) when write message to you, open any file, it work/plays/fine.

@bean
Yeah, just when i read your post, i go to diropus, remove xmp plugin at all, then run, choice mp3, and have TuneNet Opener crash.

I got an answer from the os4 devels:

INTB_VERTB (it has nothing to do with the vblank of the gfx board) doesn't work on SAM and Peg2 machines until rtg.library 41.4330 (10.3.2009).

It mean that it's not in the public at moment. So solution for me was a writing a software interrupt emulator of vblank. To anyone in interest here is os4 binary for tests, and here is the source. Also a list of results from differetn modes (different HZ). (you can detect - count is bigger when HZ is bigger =>more smooth fps).

All of this mean, that if someone want to use vblank interrupts, just use software one if you want to support all.

Have strange situation here. I tryed to port some os3 programm, which heavily related on hardware interrupt INTB_VERTB. The functions just setup inerrupt server by the AddIntServer(INTB_VERTB,myint), and then in Code of interrupt just do somethink , which by logica of VERTB will be happenes when interrupt happenes.

The problem is it: i have never happenes VERTB interrupt. I tested the same binary on other os4 oses (on m1, on other peg2 with _the same gfx card!_ ) - it works just fine. But for my setup, looks like it never works.

The binary (68k ones) for all in interest are here , source here. So, anyone can test it, and check:

if count = 0 always - then it mean VBLANK interrupt does not works for you. If count = "some different values", then it mean VBLANK interrupt works for you. I am in interest to see on how many aos4.x it works, and on which is not.


Anyway, can anybody point me what i can change and what i can try to remove/add, to make it works for me ? Somebody says that when you start to works with "prefs:screens" for promote screens => can be cause of problem. But for me it looks very strange. Becouse if i have in my radeon.info INTERRUPT=YES, then in any case interrupt must be happenes. Right ?

@bean
Tryed new version, The same freeze (and some time crash) on the begining. So, how it looks like:

1. reboot
2. run owb
3. run tunenet -> open file -> choice mp3 -> press play = solid freeze.

Some time, it's crashes on task called Tunenet Opener. Maybe it's the same bug which make solid freezes, maybe not. Anyway, when it was "crash", i catch it by GR and save logs, you can found it here. There is 2 directoryes, in which i colled crash logs from one time, and from second time (just to be sure about).

@bean

No have any crashlogs for you right now, but my expirience with 8654 on peg2/os4.1:

1. Frize from first time when try to play mp3. Before OWB was loaded and directory opus. Then i just run new tunenet, open mp3, press play - freeze of all the os

2. reboot, run tunenet only, then try to play with "radio". Just do some resizing, jump from station to station (i think 10-20 times) => freeze of whole os :)

Very possible that it's somethink wrong was added for "play" routine and becouse of it it was the same freeze on 2 different thinks (visually, by coding part i think it's the same fucntion ?). With 8650 (which i use before) - no problems at all. I can load tons of progs, etc , etc.

I will try to test it more deeply ..

@rigo
Goog to hear about update. Programm itself are very nice one. Very like feature about jumping to the fucntions just by list of names in different window and "jump" on the need it line by number (very usable thinks).

But as usually happenes, "updates" it's somethink like "when it's done" :) So maybe copy+paste fix will be easy to release before the next "massive" update ? Just a new binary with only this fix. And maybe not for public, just for some ppls who read this topic :)

@rigo
Then can you please just (maybe) upload a fix for the latest public version, which just only fix copy+paste ? (very necessary one, and just not works => without unpossible work with software in "full mode"). For me it looks like i just mark somethink, press "ctrl+c", and have freeze of os.

@Rigo
It is possible to get from you latest version (non-public one) for betatesting ?

@cygnusEd

Out of main topic, but AmiCygnus related : Have you any progress with "speedup/rendering" thinks on the amicygnix itself ? :) If yes, can you write what you already to do f compare with the latest AmiCygnix from the os4depot ?:)

@Snuffy
Some old 68k programms do almost in 50% of the times some DSI which are not so necessary to see for me anymore. So, it's good that i just can disable GR (even by this dirty way), and have working programms visually.

@Elwood

I know what i do and for what :)

@Kotler

Indeed :) thanks

You sure that i can't ? I know that for morphos you can off/on this kind of options for somethink like that GR errors. For sure it's possible to off somehow, and i just think about arexx programm which will be automatically press on "ignore dsi errors button", but it's ugly solution of course. I just think some kind of 1/0 value to some file somethere is present to on/off gr windowses.

As i know it's possible to swith-off showgin of Gr/DSI window ? It's sometime a bit annoing, becouse some programms works fine, but sometime do DSI. So just want to disable it for some time. How i can do it ?

@salass00

Thanks a lot again ! It works ! The good think, that i can setup any mode now and choice for example 400x300, which looks better than 640x256 , but looks good.

Btw, have os4 any kind of utility for "move" the screen ? I.e. for my monitor, with p96, all my modes (almost everyone) over the 'bounds'. I think it's becouse of monitor, but it possible to "resize" a bit screens visually by some programm ?


@TSK
thanks for ideas, i just tryed Salas00 ones and it works as it must for me. What about programms, you can try these music disks which works fine with new ptreplay.library which salas00 done some days ago (replaicement for old one, which worked before over the paula, this one works over the ahi, so any programm which use ptreplay.library can now works fine, when you replace old one or new one + ptplay.library). The list of music disks which you can try:

RedPack4,5,6
Embrion2

The all was 640x256, but now it can be any mode with prefs/screen. Or for example you can try these little prods:

wheelpz info (1,2)
MWI Zosia (1,2,3,4)

It has no music at moment (becouse used some deep hardware players), but it gfx/etc works, and it also was 640x256 on half of screen but works ok with prefs/screen.

@tboeckel

Akways strange to see answers like somethink "impossible". Everythink is possible. About what i think right now:

1. fix the binary. Found the screenmode values in the binary itself, and patch it by hardcore way. Of course solution far from perfect, but why not.

2. Writing a wrapper, which will be as resident /commoditie in the memory, and catch asknowlegment to the 640x256 modes, and when it happenes, change it on the 640x480.

3. If aos4 can handle 640x256 mode, with just redirection of 640x256 to 640x480, then, very possible to do the same, but just redirect to the 640x256 mode, which i have native on P96 screenmode.

And it only about i think (i know nothink about, but it possible already). I just think, that some kind of utility whch can do some of this already present on os4. Becouse on os3, some kind of this utilty was done (can't remember this name).

ps. For example on Morphos, this mode handled correctly. I just choice 640x256 screens, and it shows in full screen, no like on os4 on half of screen.