Nice!

What version is actually part of the latest official OS4 release? 1.8? 1.11?

_________________
- Don't get fooled by my avatar, I'm not like that (anymore, mostly... maybe only sometimes)
> Amiga Classic and OS4 developer for OnyxSoft.

It sounds like the changed DHCP behaviour *might* fix the annoying (but harmless) start-up warning about IP address 0.0.0.0 when another computer boots at about the same time... (Assuming this bug hadn't already been fixed in an earlier version that hasn't reached OS4 either.)

_________________
Author of the PortablE programming language.
It is pitch black. You are likely to be eaten by a grue...

Quote:

What version is actually part of the latest official OS4 release? 1.8? 1.11?

The OS4 version is technically on the same level as the 68k version, has been for a whlie in fact

It only took me three years to get it it into proper shape for release. The installation package had to be updated (including the SDK, the documentation, the installation script, the update script, etc.), and the changes had to be tested on a plain 68000 Amiga, which was the real problem: my A600HD, which I had used as a test platform, ceased to work in 2015, and that delayed everything

@olsen
Great job!!

Last edited by Sir_Lucas on 11-Oct-2016 at 04:04 PM.

_________________
AmiParty Team Member Chelm, Poland

@olsen



What a weird problem to have - there are plenty of people with A600s to test on. I have two myself. Annoying that something so trivial causes delays. I did something unusual the other day, and bought Roadshow, mostly because I was helping someone setting it up, and in doing that ended up installing it myself. The demo kept timing out annoyingly soon, so in a moment of frustration, I just bought it, and newer Cygnus Ed while I was there. At least I can now attest that Roadshow works great on MIST with null-modem and PPP, and A600 with Apollo A630 and various PCMCIA cards. How do one tell whethere one has 1.11 or 1.12 at hand? There doesn't appear to be any notice about that in any binary at least. It would be nice if "version full" on at least bsdsocket.library could give a direct hint about "package version".

Btw, keep an eye on this :)
https://www.vyncke.org/ipv6status/

Oh, and yeah, thanks for the effort.

Last edited by kolla on 14-Oct-2016 at 04:36 AM.

_________________
B5D6A1D019D5D45BCC56F4782AC220D8B3E2A6CC

Quote:

What a weird problem to have - there are plenty of people with A600s to test on.

I did not realize that I was in trouble with the A600HD until it proved unresponsive during testing. This is not my development machine, in fact I must have used it only 3-4 times in the past six years. When I found out what was wrong I didn't have the time to get the hardware looked at and possibly fixed.

Eventually, I asked for volunteers to give the new Roadshow a spin instead. I would have preferred to test it myself, because had something gone wrong with Roadshow for 68000, it would have been really hard to diagnose the problem. Sometimes, for debugging to yield results quickly, you have to be close enough to the hardware to kick it

Quote:

I have two myself.

I suppose that's one of the differences between you and me I only have a handful of 68000 machines, and only one of these even runs Kickstart 2.04 (which would be an A500+ with a Belgian keyboard -- don't ask). The number of A3000 class machines I have in storage is considerably higher here...

Quote:

Annoying that something so trivial causes delays.

Tell me about it I tried to test this in 68000 emulation via WinUAE but was ultimatively unsuccessful.

Funny how such seemingly small obstacles can derail things. What went wrong had further repercussions along the way. As they say about accidents, there's usually not just one thing which goes wrong, but a combination of 3 or more factors which come together and result in desaster.

Quote:

I did something unusual the other day, and bought Roadshow, mostly because I was helping someone setting it up, and in doing that ended up installing it myself. The demo kept timing out annoyingly soon, so in a moment of frustration, I just bought it, and newer Cygnus Ed while I was there.

Thank you

Quote:

At least I can now attest that Roadshow works great on MIST with null-modem and PPP, and A600 with Apollo A630 and various PCMCIA cards. How do one tell whethere one has 1.11 or 1.12 at hand?

There is a file called "version" in the "Roadshow.lha" archive which for 1.12 contains the text "$VER: Roadshow 1.12 (5.10.2016)". This covers the whole package. Each component, of course, has its own version number.

Quote:

There doesn't appear to be any notice about that in any binary at least. It would be nice if "version full" on at least bsdsocket.library could give a direct hint about "package version". Btw, keep an eye on this :) https://www.vyncke.org/ipv6status/ Oh, and yeah, thanks for the effort.

You're welcome

I enjoy using RoadShow. Thanks for the update!

I think it's a bit faster as well than the previous version.

_________________
AmiParty Team Member Chelm, Poland

Quote:

I think it's a bit faster as well than the previous version.

It is? Well, that's welcome, of course. The main goal was to improve reliability, which in particular concerns the TCP/IP stack's own internal timekeeping.

Prior to this release, the timekeeping was intimately tied to the system clock time. If the clock time jumped back or forth, such as during a dailight saving time change, it could affect events related to how data throughput is measured, for example. This could have had repercussions on how TCP transmissions are "tuned", affecting throughput.

The new code is a bit more accurate than the old one. But you never know what other effects there might be.

So far I'm happy that it does the job, because it is rather more complex than the old solution I employed.

From my observations it is around 2 - 3% faster than the previous version.
Once again great job!!

_________________
AmiParty Team Member Chelm, Poland

As told tech... the software is on the exact same level as the 68k ver., and the improved --- or changed --- source code should not give better benchmarks --- perhaps a coincidence or a approach by God --- ?

High!