I ran AmiUpdate today, which was the first is probably a year.

It started, checked for update OK, and scanning reveiled AmiSSL 5.8, IBrowse 2.5 and MUI libraries plus something more.

But most failed to update! Only two out of eight items updated properly. Any known issues?

Here's part of the log:
...
21:23:20 Creating working directory...
21:23:20 Created
21:23:20 Starting update of 'AmiSSL-5.8-OS4.lha'
21:23:20 Error returned from downloading file 'AmiSSL-5.8-OS4.lha': Transfer failed
21:23:20 Update finished
21:23:20 Checking for other entries in the same archive...
21:23:20 None found
21:23:20 Starting update of 'emotionupdate1.1-1.9.lha'
21:23:20 Trying to connect to: www.entwickler-x.de on port: 80
21:23:20 Connected
21:23:20 Redirection required, following link....
21:23:20 Error returned from downloading file 'emotionupdate1.1-1.9.lha': Transfer failed
21:23:20 Update finished
21:23:20 Checking for other entries in the same archive...
21:23:20 None found
21:23:20 Starting update of 'IBrowse2.5.8-PPC.lha'
21:23:20 Error returned from downloading file 'IBrowse2.5.8-PPC.lha': Transfer failed
21:23:20 Update finished
21:23:20 Checking for other entries in the same archive...
21:23:20 None found
21:23:20 Starting update of 'MUI-5.0-20210831-os4.lha'
21:23:20 Trying to connect to: github.com on port: 80
21:23:20 Connected
21:23:20 Redirection required, following link....
21:23:20 Error returned from downloading file 'MUI-5.0-20210831-os4.lha': Transfer failed
21:23:20 Update finished
21:23:20 Checking for other entries in the same archive...
21:23:20 None found
21:23:20 Starting update of 'MUI-5.0-20210831-os4-contrib.lha'
21:23:20 Trying to connect to: github.com on port: 80
21:23:20 Connected
21:23:20 Redirection required, following link....
21:23:20 Error returned from downloading file 'MUI-5.0-20210831-os4-contrib.lha': Transfer failed
21:23:20 Update finished
21:23:20 Checking for other entries in the same archive...
21:23:20 None found
21:23:20 Deleting working directory...
21:23:20 No succesfull updates performed

Thanks for a very good explanation!

So even though the file sizes are not the same for both "monitors", they are the same. Do they load anything, or is someone just using their tooltypes?

Anyway, I'll remove the 6750 monitor and see how it goes.
Have an RX560 on the way, so hope I can get that working too

And maybe my memory is a bit hazy about P96Mode but at least it allowed you to try out stuff on-the-fly.

@tonyw

Yes, it was a "new" monitor, with a higher resolution than my current modes.

I think I chose to have "hardcoded" modes rather than DDC to get fewer choises in screen mode requesters so I don't have to scroll and mess around with useless modes when using PPaint and certain pic viewers and stuff.

I did switch on DDC to test this monitor though because the modes I made manually didn't work for some reason. Wrong frequency probably...

Turns out the 6750 can't display 2560x1440 on HDMI, and my DP->DVI adapter apparently doesn't work backwards from DVI output to DP input. So couldn't use that monitor.

Still, what's "Board 0"? Is that supposed to be the first option in the cycle gadget on the monitors page or what? How hard could it be to put the actual monitor name there, when it has it in the list..? unintuitive... Also very unclear which of the Radeon "monitors" it actually picked (since none of then is called "Board 0" ), and if both were compatible with my radeon why it didn't show a selection for Board 0 and Board 1..? That GUI needs some further thought I believe.

@Raziel

Yeah, I read about your incident! But that was the jumper, wasn't it?

But the problem with not using the menu with the HD card isn't that I don't get a picture in WB, it's that it can't find the harddisk to boot from for some reason. Can't remeber the exact error, but it just errored out saying it couldn't find the boot parition or something. Why?

So instead of loading the menu, it should just boot immediately from the configured disk, right? That's what it does with the RX card? So wouldn't the same command work with a HD card in the machine?

So, I was testing a new monitor and needed a new resolution for it.

In ScreenMode prefs I have 3 (x3) resolution for "Board 0:". Under Monitors I have two monitors: "Radeon" and "Radeon HD 6570".

I have a HD6570 in my X1000.

The 3 resolutions I have available under Screen modes are the same as those in the Radeon monitor (which I don't think I use). The "Radeon HD 6570" monitor have several resolutions ticked than what I have available.

- What's "Board 0"? Where does it get that from?
- What's the difference between my "Radeon" and "Radeon HD 6570" monitors?
- What is a DEVS:Monitors/"monitor" anyway? The driver is a Kickstart/*.chip file.

Why can't I add a resolution on-the-fly to test? Why a reboot?!?

This prefs just seems to very unintuitive, and I've been an active Amiga user since '92. Picasso96Mode was clear and straight-forward. Technical, yes, but not fuzzy and unclear. I've also use CyberGfx and other tweak tools that could do realtime change to the frequencies etc.

Hi

I just ordered myself an RX550 for my X1000. For this I need that CFE hack to get it to boot where it doesn't run the menu and the jumper to push output to serial.
Ref: https://www.amigans.net/modules/newbb/viewtopic.php?topic_id=8920

One question: Why can't I always use the non-menu boot option also for my old HD6750? That way only the jump would have to be changed if I need to swap cards.

Is that menu *required* for CFE to boot when the serial jumper is not set?
Ref:

I ask because I tried it, to be repared, but it doesn't seem to work (with my existing Radoen HD and the jumper not set). Says it can't find the boot disk or something, even if I use the same line as the menu does.

Thoughts on this?

Hi

I'm using RDesktop to connect to my Raspberry Pi from my Amiga, but I can't get the clipboard to work.

I searched the net and there are issues, but I don't see the same errors they talk about: Failing to connect to chansrv after net dropout f.ex. I just simply doesn't work.

It works fine from windows though. And it works fine between my Amiga and windows too.

However, if I remote to windows from Amiga, and then to the Pi from windows and then try to copy stuff from linux to the Amiga (two RDP session away) strange things happen. Directly from linux to Amiga doesn't work, but if I middle-land the text in Notepad on windows first, then past the text on the Amiga, it sort of freezes!

It seems RDesktop on Amiga locks the clipboard, but fails to get data from windows and so all device access on the Amiga freezes, like disk, clipboard etc. But Amiga windows and RDP still work (until I try to copy something more on windows, then RDP freezes too).

Any ideas?

Could RDP on linux be sending data on clipboard the Amiga doesn't understand ?

@Raziel

So it seems you got it running fine with your X1000+RX560.

Why do we need to do what you do with CFE? Does it wait for a supported gfx card to be ready to be able to proceed with the boot or what? And so we need to disable gfx output completely to make it ignore that step?

I don't seem to have any X1000 or CFE docs on my X1000 for some reason... have stuff about U-Boot on A1XE, but that's not helping I guess ;)

So basically what we need is to turn off gfx with some command sequence. Do we need the jumper, or is that just for recovery, to be able to restore settings?
Can't these settings be changed from OS4 with some NV commands? I thought so.

@trixie

Thanks!
I'll try it.

Turns out OS4 does return 1.6GB as largest available, but SampleManager doesn't believe it. It has some internal limit of 100MB. Patched exec->AvailMem to force it to return more than 100MB, only to find out that it returns 1,6GB and SampleManager says 100MB no matter what... ;)

I came across an annoying problem.

I need to edit an audio sample that is 170MB large. I use SampleManager because it is the best (on a machine where AudioMaster doesn't work). But the program says it cannot allocate more than 100MB (of presumably continuous memory), but I have a freshly started system with 1.5GB memory free, which means there should be at least 1GB continous available.

Is this an annoying limitation of OS4 memory management, to simply forbid allocations above 100MB, or at least report no more than 100MB when asked about the largest available chunk?

X1000. OS4.1 FEU2
I have swapping off, since I have 2GB physical RAM.

Do I have to hack this program to remove this check or what? Patch/hack exec perhaps...

@Raziel

Good to hear you say that and sorry for rambling, but I read/skimmed through the comments about allowing VSYNCMIN to be set below 50Hz with regard to CRT monitors.

You wrote: ("by Raziel » Thu Jan 13, 2022 9:06 pm")
....
wrt the changes

First of all, thank you
Second I would like to know if there is any chance I could test it?
Maybe with a time-out after a week or so or a timer that counts the library open state and stops working after 50?
....

Ah, maybe that was regarding to a beta-version timeout, not a requester warning for the VSYNC setting?

@trixie

Aah!! I read from Hz to hz... bah, sorry ;)

@Raziel

Thanks for the explanation.
I seem to have lost my password at the Hyperion board so I hope your read here:

I noticed you suggested some weekly warning timer (supposedly to popup on bootup or whatever) if VSYNC is set to low. Please DO NOT let some horror like that plague the rest of the Amiga life with 4k+ screens, that's insane! Either it works or your monitor blows up. Don't need a reminder every week or any such at all.

Warning before allowing to Test, Use or Save screenmode is all we need. No other silliness on bootup at 1024x768 (that won't likely display on CRT anyway). Even if someone with CRT sets VSYNCMIN to 25 doesn't mean it will ever choose such a frequency. It's just a limit of what you can set.

@tonyw

> Anyway, we've changed the hard-coded limit from 50 Hz down to 25 (and corrected the spelling of "Hz" from "hz").

That's wrong! Herts is abbreviated "Hz" and nothing else. "hz" is wrong. Change it back.

Like "mb" (millibit, whats that?). MB=megabyte, Mb=megabit, kB=kilobyte, kb=kilobit, KB=Kelvinbyte (what's that??).

As the title says. I'm planning to upgrade my monitor for my X1000, but I'm not sure if there are any gfx-cards that has support for 4k displays.

I also want video acceleration with the new DvPlayer and Enhancer pack, but for this I need a new gfx-card. But which one? Anyone of the SI cards or the RX cards that supports 4k *video* decoding?

And this...:
News for RadeonHD.driver from 2013 says this:
"The new Version 1.0 RC2 (Release candidate) adds a few new features and bug fixes which improves card initialisation and solves intermittent boot problems which affected some Sapphire RadeonHD 6570 cards."

I have one such card and have started to experience intermitten boot problems a few years ago or so. Specifically after cold boot. Has the bug returned?

Same CDFileSystem 53.8 (20.10.2019).
I don't it gets that far. The CD-burner programs can't read drive status of a DVD. As if it isn't readable.

Could be hardware, but CDs work..?
I'll try some more disks and perhaps some cleaning before trying to dig out another drive.

For some reason I cannot read or write any DVDs on my X1000.

I'm pretty sure I burnt some DVDs in the past on this machine, but now it just won't recognize any DVD disk. Not plain DVD, not DVD-R, not DVD+-RW. CDs work. What happened?

The drive is the original that came with the X1000: "Super Writemaster SpeedPlus+". Vendor: TSSTcorp, Product: CDDVDW SH-222BB, as recognized by FryingPan.

Did something happen in Update2 that made it stop reading DVDs?

It seems both FryingPan and MakeCD fails to read media status when there is a DVD inserted. Workbench shows no icon or content when inserting a DVD.

I'll do some more tests, like using Popstring, and creating the Popasl separately by itself. Both in the main thread and then in the GUI thread and see what happens.

Reproducability is key to find such a bug. A simple crash log is about as helpful as a simple DAR register...

Just wanted to see if this is somehow a known problem or at least heard of before. The Popasl obviously works in normal cases or many MUI programs would fail. The only thing that makes this program stand out is that the MUI GUI isn't created and run on the main process of the program.

I have a project that runs a MUI GUI on a separate process that I start with NP_Child,true. The rest is pretty much default.

Everything works fine. I have a message passing system between tasks so no cross-thread access of MUI etc. But I cannot include PopaslObject in the GUI because then the GUI creation crashes with wierd and random results. DAR is typically 0x39553933 ("9U93").

If I remove PopaslObject from the MUI GUI creation tree all else works fine.

Does PopaslObject have a problem with being run on a process other than main?

OS4 latest final update.

@NinjaCyborg

This is make (from gcc), not AmigaDOS. There is not failat in make.

@LiveForIt

Thanks, I found a similar solution, using touch first to create a dummy and then deleting with #?, before realising that you can prepend a command with - to ignore its return code.

target:
-delete MyFile
continue with important stuff

That works!

I try to use the normal delete command in a make script, but no matter what I do, make stops if delete can't find the file to delete (already deleted):

delete MyFile
MyFile Not Deleted: object not found
make: *** [target] Error 5

I have tried to add QUIET and/or FORCE, but to no avail. Delete always seems to return WARN (5) when it can't find the file, so the makefile always stops. And QUIET does NOT make it quiet. It still prints
"MyFile Not Deleted: object not found"

Any comments on this?