Notices by Alexander Bochmann (galaxis@mastodon.infra.de)

Ah, so, maybe it's SNMP?

On the topic of monitoring... Useless audit message of the day:

[2046] 18:40:17 2023-12-03 level: 6
"Authentication failure notification."

Yes?

Who?

From where?

Which protocol?

:flan_eyeroll:

From IRC: Apparently $someone intercepted XMPP traffic to jabber.ru (hosted by Hetzner and Linode in Germany) with a middlebox, and this was detected only because the Let's Encrypt certificates used for the interception ran out: https://notes.valdikss.org.ru/jabber.ru-mitm/

This kind of attack is not usually possible without cooperation from the hosting provider, quote: "We believe this is lawful interception Hetzner and Linode were forced to setup."

#infosec

I guess there is some kind of botched LLM processing involved when an article on some previously unknown tech news site starts with:

"rewrite this content and keep HTML tags

Fix grammar in this content, make it friendly formal tone and keep HTML tags"

@darrenmoffat @kalleboo Yeah, might be in the wrong age for that one, according to the IE5 for UNIX release notes: "Internet Explorer is not supported on Solaris 2.5.1 and earlier."

I've seen mentioned that it's kind of possible to hack up OpenStep for Solaris to run on 2.6, but it's brittle even on the version it was released for...

@darrenmoffat @kalleboo (...though I see now that there was also a IE 4.01 release that was compatible with Solaris 2.5 and up. Might try that one at some point.)

Hrm, I completely forgot about LuBu OpenMagic, which was a redistribution of the OpenStep 1.1 files with some changes and alleged optimizations, and I seem to remember it should work on newer Solaris versions.

Someone uploaded that to github at some point (https://github.com/itomato/lubu), and archive.org also has the original .tar file.

There was also a post on one of the Sun blogs with a little bit of history of OpenStep for Solaris, which is on the Archive too, at https://web.archive.org/web/20060215231443/http://blogs.sun.com/roller/page/richb?entry=openstep_history

Some of the archived comments talk about the Lighthouse apps...

For a long time I thought that the Lighthouse Office Suite would be unobtainium forever, but it seems someone uploaded them to nextcomputers.org in 2016, and I missed it? (www.nextcomputers.org/NeXTfiles/Software/NEXTSTEP/Apps/Lighthouse_Design/)

Not sure if any of those run on non-m68k hardware though (the FrameMaker in that directory seems to be NeXT/Intel/HPPA according to the package name - not quad-fat for OPENSTEP/Sparc).

Back to OpenStep for Solaris, it seems there was a beta of an WorkShop OpenStep development environment at some point, but I've certainly never seen a download of that one anywhere - https://web.archive.org/web/20050427180415/http://docs.sun.com/app/docs/doc/802-2110/6i63kq4s7?a=view

Hrm, so according to fbinfo I have a 1MB TGX in this system, but Xsun claims it can only do depth 8 nevertheless...

So I tried switching to GrayScale mode, and while Xsun then starts in grayscale, the Openstep desktop will still be in colour, and there's flickering when touching window titles or the background?

Apparently one of the SBus graphics cards in my stash is a Vitec RasterFlex32, which supposedly is one of the better frame buffers of that time?

The other is an older 1 MB TGX.

The RasterFlex has a custom framebuffer logo.

#retrocomputing #SunOS

According to the dmesg, I have:

VITec,RasterFLEX-320 at sbus0: SBus slot 0 0x0 SBus level 5 sparc ipl 9
VITec,RasterFLEX-320 is /iommu@f,e0000000/sbus@f,e0001000/VITec,RasterFLEX-32@0,0

Sourced RasterFlex drivers from fido.aalin.co.uk/vault/hw/system/sun/3rdparty/rasterflex/

My Xsun configuration for Openstep is now broken though...

Maybe I should take the opportunity to restart from scratch and set the system up with Solaris 2.6 instead?

Huh. I was lazy and put the card into the one SBus slot that didn't have a cover anyway (Slot 0) - which seems to run only at 20MHz in the SS10, whereas all the others are 40MHz...

...and the system has 1 * 32MB memory, and 3 * 16MB, but only reports a total 64MB?

Weeeellll, according to the old hardware FAQ, 32MB SIMMs supposedly don't even work on an SS10 (SS20 only). Great. Now the question is - why didn't I run into problems up to now?

Spent the evening with:

Rearranging RAM modules because the SS10 would only blink the capslock LED after powering on ... until I remembered I had done an setenv diag-switch? true to do a full RAM test, and then it takes bloody ages for the screen to come up, and I was too impatient...

Fighting the RasterFlex card after reseating it into a different SBus slot - until I read the FAQ, which said that the drivers need to be uninstalled and reinstalled after a reboot -- -r so the internals get rearranged appropriately.

Now back to where I was this afternoon (except the RasterFlex does not seem all to spiffy in TrueColor mode after all, and also the screen blanks occasionally for some reason)...

So since someone mentioned this abomination further up in the thread: Here's a screenshot of Microsoft Internet Explorer 4 and the accompanying Outlook Express running on Solaris 2.5.1, with the desktop provided by Solaris Openstep instead of OpenWindows.

#retrocomputing #solaris

Peak Solaris: SparcStation 10 running Solaris 5.5.1, with Solaris OPENSTEP as window manager, and Wabi (Windows 3.1 environment) plus Macintosh Application Environment as applications.

Unfortunately the colours in Wabi and MAE are off, as I only have a cgsix...

#retrocomputing #solaris

Uuugh. Took me until today to understand that pulling a window down on the Android 12 app switcher locks it so it can't be closed - and pulling down again unlocks...

For some reason this has only occasionally happened by accident up to now, and while I usually managed to do the unlock, I thought it was by tapping the small lock icon (in reality I probably happened to do enough of a swipe to trigger the change)...