@LiveForIt:
Are you sure about that? I can't think of any examples where a pointer with sign bit set means anything special.

In many DOS calls, pointers have to be converted into BPTRs by ( >> 2). If the original pointer is > 2GB, it has a sign bit which is maintained during the shift. Now there may be some functions in DOS which don't check that the upper bits are zero and which would fail for that reason. I'm only guessing.

On the other hand there are many structures defined like this:
Quote:

Now fh_Pos and fh_End are set to -1 when undefined (when the file is not open). To use the whole 4GB address range, we would have to replace those int32s with a uint32 and a separate "invalid" indicator. Otherwise we would have to limit the available address space to (4 GB - a few values).

Quote:

...means that Ramlib (the process that loads programs into memory) couldn't load a library of some sort. You could identify the library by downloading Snoopy from OS4Depot and running it before you try to start DvPlayer, then save the output to a file and inspect it (or post it here). The interesting part is "Ramlib" FAIL = xxx".

But it's probably going to end up the same way: I would suggest reinstalling DVPlayer to make sure that all libs are where they are meant to be.

Sicky, for Christ's sake don't post full crash logs into a forum. Only the stack trace is useful information, all the rest is just noise.

The most important thing (that you have not posted) is what the yellow alert said (the "01000009" or whatever).

Just what I was going to ask.

You probably know, but the serial debug output (used by IExec->DebugPrintF() and GR) uses the port settings from U-Boot (or CFE in the X-1000).

The serial Prefs settings do not matter for debug output. On the other hand, if you use "print >SER:" etc, then that will use the serial prefs settings.

Some obvious causes:

1. If you cd to a drawer in a Shell, then try to delete the drawer, the Shell has a lock on the drawer while you are cd'ed to it.
2. If you have a window open on WB, in Filer or DOpus, there is a lock on the drawer.

Of course, if you are cleaning out a partition, it's quicker and safer to Format Quick the partition.

@LiveForIt:

Quote:

That's not the best way to reply when someone offers help. He's unlikely to bother again.

You didn't say where it crashed. That might help.

I would suggest trying to find what is asking for CONSOLE:. It sounds like an assign somewhere.

Click on WB background, type RAmiga/F to bring up the Find command.
Select your system partition only, type "#?" in the "Pattern" gadget and "CONSOLE:" in the "Contents" gadget. That should find any references to the string "CONSOLE:" in your assigns or prefs.
If that doesn't find it, then search the partition where things like the browsers are stored, etc.

(All strings without the quotes, of course).

I can't speak for most of the testers, but I don't listen to music on my main system (the A1-XE). I use it as a development platform, compiling and testing system components, so I am frequently installing new OS components (drivers) and rebooting. I can't leave a CD running during that time.

In any case, if there's a radio or record playing, I stop working and listen to it, so it slows me down. I'm a single app at a time guy.

Someone who really stresses the system the way you do, would be a valuable beta tester. Wanna come and join us?

Yes, it's embarrassing that it has gone so long without being noticed by a betatester. Oh, well...

Can anyone reproduce this using only OS4 components, ie no DOpus or Tunenet?

For example, using multiview to play a sound sample?

[edi]
Never mind, I just did myself:
Load a CD, copy a track from the CD: into RAM:
Open Multiview, double-click the track and then press the Play button.
While that is playing, C:Type a large doc file to the Shell.

Sound goes mad (broken up).

I'll write a report about it.
[/edit]

[edit2]
FWIW, I get exactly the same effect with Update2, so it's nothing new.
[/edit2]

[edit3]
Correction, it's OK with aclean Update 2, I was checking a modified version with beta updates...
[/edit3]

You can actually use DMA with the on-board IDE controller, the problem is that you can't use it at the same time as the on-board Ethernet controller or you get DMA transfer errors.

Use a PCI card for either of them and you're sweet. Most people used an SiI0680 PATA card for their HDs, but these days you're better off with a SATA card, eg SiI3112, 3512 or 3114. All are supported by OS4 drivers.

The "DMA fix" or "DMA hack" as I prefer to call it, has a side effect - it means that you can't use a graphics card in the first (66 MHz) PCI slot. Many people prefer to leave their boards without the "DMA hack" for this reason.

You are right about the USB problem. However, anyone handy with a soldering iron can add the resistors for you - it only takes a few minutes.

@ChrisH:
Yes, I should have included "nvgetvar sii3114ide_conf". Thanks for adding that.

I think it's OK to have xxxide_conf undefined - the driver should be able to interrogate all drives and determine which is what - but I'm not sure, I've always defined it on my machines. What I'm worried about is the addressing of the ports - I know it's wrong and I also suspect that there is an interrupt problem on the early driver versions. We'll have to help 'carvedeye' through with that.

It's certainly OK to leave xxxide_xfer undefined, in fact it's a good idea if you move drives around as I do.

Please be patient. We've all been through this process so many times and some people just won't do their homework first.

Run Media Toolbox ("MTB"), select your driver (sii3114ide.device) and click "Start".
Now look at the column marked "Id, LUN".
What does the list of Units say? I would expect to see:
0,0 Fixed hard disk (your old disk)
0,1 Fixed hard disk (your new disk) (maybe this says 0,2)
0,3 Removable CD-ROM

Does your sii3114ide_conf reflect this configuration? In the above example it would be 1102 (1012 if it says LUN 2).
Does MTB bring up a requester suggesting that you let it fix something? If so, say yes and let MTB fix it for you.
Do you want to boot from this new HD? If so, select the HD from the list and click "Edit RDB/ReInstall". Look at "AmigaOne Boot Code", bottom of window. Does it say "l:slb_v2, 29036 bytes, ver 1.19"? If not, install it. Don't touch anything else in this window. Click "OK, accept changes" or "Cancel" if nothing needed changing.
Click "Edit partitions and filesystems". Make sure that every partition in the list has "Automount" set (except a "SWAP" partition). If you want to boot from one or more of them, set "Bootable" and a priority.
If you have two disks, each with bootable partitions, the lower Unit number will always take precedence, even if the higher Unit number has a partition with a greater priority. You can also get funny situations where the Kickstart modules are loaded from one device but the system is loaded from another, making a real mixup of library versions etc. It's safer to have only one disk with "bootable" partitions enabled.

Forget about sii3114ide_xfer for the moment, just set it to "0000" or don't define it at all. You can trim it afterwards.

@TearsOfMe

MinFontSize is disabled here and always has been.

Whether a program is reentrant or not depends on the way it is written. Although most AOS programs are inherently reentrant (meaning you can have two or more independent copies running at once), there are some limitations.

For instance, if the program locks a global resource (like the WB screen) while it is running, the other instances of it have to wait. Poor example, I know, but it only needs one Lock to screw things up. Some programs were not written with reentrancy in mind, some were deliberately written to hold onto system resources while they do their work.

It is an exaggeration to say that AmigaOS was completely reentrant. There are always some system tools that will take control of a resource while they do their work. Likewise, some calls to Exec, DOS or Intuition will lock out all others until the resource is freed again. It's necessary to make sure that system structures can not be changed by someone else while you are reading or changing them yourself.

There is no dependency on reentrancy for which version of printf() you use. You can use the C runtime library printf(), you can use DOS' Printf(), Exec's DebugPrintF() or Utility's RawDoFmt(). Normally your program will have Exec and DOS open, so you could use any of them. The C library "printf()" and DOS' "Printf()" will automatically open a console window for you if you are running from a WB icon. DebugPrintF() will usually write to the debug (serial) port, while RawDoFmt() just makes a printable string for you, like fprintf().

I can't see the point of the "PRintf()" definition in your example. You may as well use DOS' "Printf()" or the C library's "printf()".

@Slayer

Yeah, there's a bug in the driver with its addressing of the ports. It seems to number the ports 0, 2, 1, 3. He may have to experiment to find out how his drive is actually "connected".

@nbache:
Yes, it's funny that the site failed to load as soon as that story was posted on Tuesday. My computer must have a bias away from certain owners.

@ChrisH:
AFAIK no such filter has actually been implemented. Besides, this is the quasi-government-owned Australian Broadcasting Corporation, It can't possibly be offensive

I've tried every OWB version from 3.28 to 3.32, they all do the same thing, load to about 450kB then loop. I can use the rest of the sytem but it's slow to repond.

Pity I can't turn off image loading the way you can in IBrowse. Then I could load each in turn. I think the problem is caused by some Flash images on the front page. If I type in URLs to lower-level pages without Flash, they work OK.

What a mixture of responses! No, deleting the cookie collection doesn't help.

Thanks, guys.

...what have they done to my page?

It always was a sad song. Anyway, the local news site has changed their index page since Tuesday and now OWB can't load it any more. This index page (http://www.abc.net.au/news) has been my home page for years, even back in the IBrowse days.

But not any more. OWB just goes into a loop trying to load it. I've noticed that Firefox (on a PC) seems to take a long time loading the page, so something funny has happened.

Can anyone load it with OWB? How does it go with MorphOS' OWB?