Roadshow 1.12 released

[wXR]

Originally Posted by Olaf Barthel

Well... I suppose it would require me to abandon all responsibility for Roadshow, with no further development or support work being done. Right now I cannot see this happening for a very long time.

Why would it require you to abandon all responsibility for it?

[utri007]

Originally Posted by kolla

utri007: are you sure you are not just running NAT behind NAT, and not at all a bridge? Do other, non-amiga, wireless devices show up?

It would help if you could tell what IP-adresses, netmasks and gateways you have for your various devices, I suspect your wireless network and wired network are not on same LAN

Wired and wireless network are on same LAN. Behind of that Zyxel / bridge is only one device, my OS4 machine. It gets it IP from router, so this is no NAT over NAT situation. Only on DHCP on network. 68k Amigas are connected with wlan with their own adaters. Orinoco Gold and Silver PCMCIA cards.

[Olaf Barthel]

Originally Posted by utri007

First I tried to do a bridge with D-Link acces point. It worked, but it shoot down every other device using WLAN. Zyxel has worked nicely. But it can't be reason for this, it is not always on. In a matter of fact is uses same power cord than my Commodore 64, Amstrad CPC, Sinclair ZX Spectrum, Atari etc. get their power.

68k Amigas are all in same power cord.

The strange network traffic limitations should have little to do with the wiring, unless the source of power is constrained, too, delivering reduced electric current to the attached devices.

Mind you, the only time I saw something like that in real life was when I discovered that the expansion card I had added to my home PC required a more powerful power supply unit. With the old power supply unit installed the system became unstable and would hang/crash. Sometimes strange system behaviour is a side-effect of hardware issues.

But we are getting off-track, I suppose.

Could you explain how your network layout actually looks like, and which ZyXEL access point model you are using? Why are you using the ZyXEL access point in the first place?

[Olaf Barthel]

Originally Posted by wXR

Why would it require you to abandon all responsibility for it?

I feel responsible for "Roadshow", warts and everything. It has taken some 13 years to get it released for 68k machines, which represents a considerable investment of time and effort. My commitment to Roadshow is such that I will do my best to keep it in good shape, update it as needed, and provide support for it. When I am no longer able to do that, or Roadshow has outlived its usefulness, then that would be the point to walk away from it and release it as an open source project, for somebody else to care for it.

That point has arguably been reached for the "ppp-serial.device" and "ppp-ethernet.device" drivers already which no longer are as important today as they used to be about a decade ago. I am seriously pondering releasing these as an open source project.

If anything, they could serve as useful examples of how one might create an Amiga device driver, and in particular how to create a SANA-II network device driver.

Also, there are not that many open source PPP implementations around, and one more might be helpful for somebody. My implementation, with the exception of the MS-CHAP support, was written from scratch without building upon existing solutions such as the PPP daemon created by Paul Mackerras (which has NeXTstep support, which tells you how long it has been used and developed).

Writing PPP client code is tough, in spite of the solid specifications, I can tell you...

[Schoenfeld]

Originally Posted by Olaf Barthel

That point has arguably been reached for the "ppp-serial.device" and "ppp-ethernet.device" drivers already which no longer are as important today as they used to be about a decade ago.

Not exactly sure what these two do, but if ppp-serial.device handles PPP over a modem, then this may get interesting with USB-Surfsticks again. Essentially, they behave like a modem, and they even support the Hayes AT-command set. Ever thought about supporting these? If so, I have quite some experience with them from my "Nequester" project.

Is ppp-ethernet.device doing PPPoE? Then yes, this is almost obsolete these days. It's hard to buy "modem only" units, and there is no use in giving an Amiga a dedicated xDSL line.

For the X-Surf (10 MBit version), we created a PPPoE device some 15 years ago: It "speaks" Hayes-modem and wraps packets according to RFC 2516. It is largely useless today, but requires the stack to support "PPP over serial".

Let me know if you would like to look into "Surf stick on an Amiga" - I'm willing to donate a RapidRoad for testing. Got to talk to Chris Hodges to see if there's already support in Poseidon. And while we're at it, we should also check if USB tethering with an Android- or iPhone can be added in a similar fashion.

Jens

[roomeo]

Originally Posted by Schoenfeld

And while we're at it, we should also check if USB tethering with an Android- or iPhone can be added in a similar fashion.

Jens

Awesome

[Olaf Barthel]

Originally Posted by Schoenfeld

Not exactly sure what these two do, but if ppp-serial.device handles PPP over a modem, then this may get interesting with USB-Surfsticks again. Essentially, they behave like a modem, and they even support the Hayes AT-command set. Ever thought about supporting these? If so, I have quite some experience with them from my "Nequester" project.

My, looks like my old code so obsolete that it's useful again

I'd be happy to revive it if need be. This is the first time I read about USB surfsticks.

Is ppp-ethernet.device doing PPPoE?

Yes, it's for exclusive use with PPPoE. It's also one of the very few PPPoE implementations which do not convert asynchronous PPP frames (intended for serial links) into Ethernet frames: it natively "talks" in PPPoE frames.

Then yes, this is almost obsolete these days. It's hard to buy "modem only" units, and there is no use in giving an Amiga a dedicated xDSL line.

Yes, there is no justification for hooking up your Amiga directly to the Internet today.

But the world used to be a different place some 15 years ago, to put it mildly. An Amiga directly connected via PPPoE to the Internet would probably survive it, but likely have to use a firewall (as built into Roadshow) for protection against the noise of incessant port scans that would hit it like hailstones

For the X-Surf (10 MBit version), we created a PPPoE device some 15 years ago: It "speaks" Hayes-modem and wraps packets according to RFC 2516. It is largely useless today, but requires the stack to support "PPP over serial".

Let me know if you would like to look into "Surf stick on an Amiga" - I'm willing to donate a RapidRoad for testing. Got to talk to Chris Hodges to see if there's already support in Poseidon. And while we're at it, we should also check if USB tethering with an Android- or iPhone can be added in a similar fashion.

Jens

Blimey. And I thought the PPP work, which consumed so much of time, was in vain...

[cmsj]

Originally Posted by Schoenfeld

NAnd while we're at it, we should also check if USB tethering with an Android- or iPhone can be added in a similar fashion.

Jens

I can't tell from the context if that is relevant to the PPP part of the discussion, but iPhones at least, present an ethernet device for tethering, so PPP would not be relevant.
Edit: there's also some kind of pairing that has to happen, before the ethernet device will work, and at least on Linux, there is a whole separate driver for iPhone ethernet.

In short, I'm pretty sure that isn't going to work already with Poseidon

[S0ulA55a551n]

Originally Posted by JJKom

This might be just shot-in-the-dark but I have noticed similar problem in pure win network and new router.
For some reason that router is limiting access to only those IPs that are in DHCP range, in this case those Amiga IPs are outside it.
No matter if connection was wireless or wired, that router did not "see" or allow traffic with IPs outside its DHCP range.

Fix was to use only IPs that are inside, in this case set router to give always same IP to those Amigas and use DHCP with them.

This makes most sense. I use dhcp but set the router to always assign certain IPs to machines. Why do you have amigos outside dhcp range anyway


Sent from my iPhone using Tapatalk

[utri007]

Originally Posted by S0ulA55a551n

This makes most sense. I use dhcp but set the router to always assign certain IPs to machines. Why do you have amigos outside dhcp range anyway


Sent from my iPhone using Tapatalk

OK, sounds promising. I need to test this. Using DHCP takes time and it is unreliable. Just checkd, I can make a reaservations to Huawei's DHCP server.

[wXR]

[QUOTE=Olaf Barthel;1116208]I feel responsible for "Roadshow", warts and everything. It has taken some 13 years to get it released for 68k machines, which represents a considerable investment of time and effort. My commitment to Roadshow is such that I will do my best to keep it in good shape, update it as needed, and provide support for it. When I am no longer able to do that, or Roadshow has outlived its usefulness, then that would be the point to walk away from it and release it as an open source project, for somebody else to care for it.

No no, I mean, why, if the source were made available, would you no longer be responsible for it? I don't get the connection. The point I am trying to make is that the lack of open source software harms the retrocomputing hobby; it does not help it. That is why I asked what would be required for you to make the source available. I am not asking "What would it take for you to drop it and stop working on it?", I am asking what would be required for you to release it under a permissive license so that many people have the option of inspecting it and working on it, if they should so choose. Before any outdated opposition is raised, please note that in the age of Patreon and Kickstarter, making the source of your TCP/IP stack available does not mean that you cannot make some money by continuing to work on it.

[vulture]

@wXR

not Olaf, but if it's closed source and Olaf still supports it, I don't understand how this is hurting the hobby. Now, if Olaf had abandoned it and kept it closed, you'd be correct of course. If Olaf opens it, I suppose that freeware binaries can be built and nobody would have to buy it and that would be a shame, since Olaf has put way too much work into it, developing it and supporting it.

[Predseda]

If I buy Roadhow, will it display something like "This copy is registered to..."? I would be proud of it.

[wXR]

@vulture

It indeed hurts the hobby. Decisions like this give no one but Olaf the opportunity to inspect the code, or to improve the code. It locks people into his decisions, and also sets a precedent that proprietary software is somehow a good idea on the Amiga in 2016. Yes, one can and should certainly hold his own prerogative, but my point is simply that this model no longer benefits even the developer.

If Olaf is putting tons of work into Roadshow, developing it, supporting it, etc, then he can still do that when the source permissively licensed. If it's about having some trickle of money, then Olaf could for example setup a Patreon.com account. We'll all chip in some small amount on a monthly basis, and viola: Olaf is now a full-time free software programmer on the Amiga.

This is not the 1980s; we can certainly do better than some ancient vendor-style relationship with a developer.

[wXR]

Just one more note and I will finish, because I actually don't want to completely hijack this thread with my own ideology. I just want to make it clear that I am not attacking Olaf, or Roadshow, or any company or project in the space. Rather I am strongly questioning the persistence of certain ideas that seem to hold back a substantial degree of community participation and growth. That is all. Wherever it is eventually discussed/debated, I hope that will be the understood context, rather than the appearance of personal attacks.

Perhaps the rest can be continued here:

http://eab.abime.net/showthread.php?t=77722

My apologies for the distraction, Olaf.

[idrougge]

I know a way for all parties involved to be happy: Olaf makes his TCP/IP stack and wXR makes his own.

[utri007]

Originally Posted by utri007

OK, sounds promising. I need to test this. Using DHCP takes time and it is unreliable. Just checkd, I can make a reaservations to Huawei's DHCP server.

Tested, I couldn't get DHCP work with Raodshow, but it worked with static IP. Tested with IP inside of DHCP range it worked, but still not not visible to other devices. Couldn't ping Amiga, even when it was 100% sure online and internet.

Turned my Amiga OS4 machine on, changed static IP to DHCP, worked just fine and it was visible to other devices.

Tried to add reservation to DHCP server with MAC address, but got syntax error.

Decided to be happy with current situation.

[JJKom]

Originally Posted by utri007

Tried to add reservation to DHCP server with MAC address, but got syntax error.

Decided to be happy with current situation.

Which Huawei 4G model do you have?

[Olaf Barthel]

Originally Posted by wXR

No no, I mean, why, if the source were made available, would you no longer be responsible for it? I don't get the connection.

It's the other way round: if I could no longer take care of the TCP/IP stack then it would be prudent for me to step down and let somebody else take care of it, by making it an open source project.

The point I am trying to make is that the lack of open source software harms the retrocomputing hobby; it does not help it. That is why I asked what would be required for you to make the source available. I am not asking "What would it take for you to drop it and stop working on it?", I am asking what would be required for you to release it under a permissive license so that many people have the option of inspecting it and working on it, if they should so choose.

What it would take for me is to be unable to support and develop Roadshow as I would want to, according to my own strange standards for Amiga software development.

Before any outdated opposition is raised, please note that in the age of Patreon and Kickstarter, making the source of your TCP/IP stack available does not mean that you cannot make some money by continuing to work on it.

Incidentally, the complete source code to AmiTCP 2.2 is still available on Aminet (this was part of the AmiTCP authors' thesis, and it is licensed under the GPL). Look for "AmiTCP-src-22.lha" and the associated archives (search for the text "stack as a shared library").

This is the blueprint for creating an AmiTCP-API compatible TCP/IP stack for the Amiga, using a BSD Unix-derived kernel/userland and shell command code.

As far as I know, Holger Kruse followed this blueprint, and I did the same.

Somebody else could follow it, too. The reference documentation for the 4.4BSD TCP/IP stack is still in print ("TCP/IP illustrated, Vol. 2: The implementation", by Gary Wright & Richard Stevens).

[Olaf Barthel]

Originally Posted by wXR

@vulture

It indeed hurts the hobby. Decisions like this give no one but Olaf the opportunity to inspect the code, or to improve the code.

Small note: the Roadshow code has been in the AmigaOS4 source code repository since the project's inception, where it did benefit from review by other developers. The TCP: device handler which is part of Roadshow, for example, was reviewed and significantly rewritten by Jörg Strohmayer.

Roadshow also did benefit enormously from beta testing within the AmigaOS4 tester community. This is one of the reasons why there have been so few updates to Roadshow 68k so far: the big problems were resolved before I finally managed to get the 68k version released.

It locks people into his decisions, and also sets a precedent that proprietary software is somehow a good idea on the Amiga in 2016. Yes, one can and should certainly hold his own prerogative, but my point is simply that this model no longer benefits even the developer.

Hey, whoever wants to make money from Amiga software development is likely going to be disappointed in the long run...

Let me put it this way: the money that comes in through sales will never pay for the amount of work I put into it.

I am a very curious person (probably in more than one meaning of the word), and development of Roadshow was mostly driven by my curiousity.

The feedback from Roadshow users and of course the money that comes from sales of the product "only" validate my commitment to Roadshow, but that's quite something in its own right