@jabirulo

Whenever you make your next test run. If you can give the debug output and maybe even the a wireshark capturing would be nice.

@trgswe

Good to know that i don't need to swap the data. That was a little question in back mind, which got answered now.
Will see that i rewrite the network read/write stuff.

@flash

The sources can be downloaded here https://drive.google.com/file/d/1VgN9h ... yRKq-bNf/view?usp=sharing

Something like that might it be, but currently i have no idea where to look. But i have the fleeing it might be wrong format string for some output and/or 64 ints alignment stuff.

@jabirulo

I had access to an Win10 Server and could test my implementation and ran in the same bug as you discovered in my implementation. Here is a fixed version, please if you like test it again:

https://drive.google.com/file/d/1YYpF6 ... xFRS3YCf/view?usp=sharing

@all

Even that the strange AllocBehavior disappars more and more. i got output on the shell, which isn't programmed. Anyone an idea what could cause that. See the image and the zeros:

Quote:

(Well don't how to add an image, so typed out) It seems to be exactly 64 zeros.

@trgswe

Well smbv1 doesn’t bother me. There is already a client available for it, but it doesn’t support smbv2. Thus I just targeting to implement a smbv2 client with NTLMv2 authentication. So that a samba share can be accessed from the amiga without playing around at the server.

@tonyw

Of course it is a Wombat. You might not get mixed up by a Quokka. Even cute, bur haven't met any in real life yet, so don't know if they get my excited as the wombats did, as i first encounter one in Crains zoo, along time ago.

@jabirulo

Interesting error, looks like a missed translation of an endian encoding. See if that version https://drive.google.com/file/d/16OLQg ... hVlSG9IF/view?usp=sharing
gives more millage.
BTW what SMB Server do you use? Was it Win10?

It would be really interesting to see if anybody else has the same behavior.

@all

Thanks for all the input and hints. I'm have probably found thousand of bugs. But no found bug completely eliminates the "AllocVec" crash. It still comes up.

But if it doesn't crash, i can authenticate against all my SMB server, for which i have access to.

If anyone want to try, please download the binary here https://drive.google.com/file/d/1dZaFw ... 4K_iJBK5/view?usp=sharing , but be aware it is still buggy!! Use at your own risk.

Any feedback is welcome, so i can improve the implementation.

The binary now requires that you supplied all related information on the command line like this:

smbv2.debug SERVER=<ip/dns> SHARE=<sharename> USER=<username> PASSWORD=<clear text password>

@jabirulo

Thanks for the wireshark capturing. I will used it to improve the communication with other servers,

@LiveForIt

Thanks for idea with the macros.

Quote:

You are completely right. That's is way i want to really understand whats going wrong. I think one way would be to really debug the stuff on premises. But the spotless debugger from os4depot doesn't work for me. And the latest release from the Spotless debugger thread here on amigans, isn't accessible anymore. I Hope alfkil will soon release a version general available. And the GDB debugger, there I miss a little graphical support, plain text based interface is way out for me.

@broadblues

Quote:

I have 2GB Ram and i test it just after a fresh restart, so i don't think that lack of memory or fragmentation paly a
role. Furthermore on my machine the allocations made up till the crash. Each is less than 1024 bytes (and only around
6 allocations) and one or two bigger chunks around 65KB.

And you are right that the return result should be checked. If did checked it, i just forgot about it. If you see a
something in the code, please let me know.

But the crash seems to be in AllocVecTags, thus it just returns nothing, which can be checked.And that drives my crazy.
I think i'm accidentally poking in memory location, which i shouldn't. But my C skills aren't that good, that i
immediately see such places.

@jabirulo

Yes. You only need to select the network card, where the traffic runs through. And don't forget to actually start the capturing in wireshark.

@jabirulo

Thanks for trying out.

But investigation your output it seems that my program cannot handle the initial response from your server. I test against a mac computer. I will see if i can upgrade the implementation, to handle the response. Than we can see if it progress further with your setup.

Might it be possible, that you can capture the network traffic with wireshark on your Win10 machine, and give it to me?

Hi all,

anyone interested to help me in figuring out what i'm doing wrong in my code?
I using AllocVecTags to get memory from the system. Normally it works, except suddenly a call to it fails, with the
following output via the serial port. The output is mainly from my program except for the kernel at the end, which
i cannot make any sens of.



I had this once, and figure out that before AllocVecTags fails, i previously allocated memory with less size than expected
and wrote over the allocated space, which seems to disturb AllocVecTags during the following call. I can not see that i'm
still doing it. So what else can make AllocVecTags fails?

The source is avaible here.

If anyoue want to know what the programm is inteded to do at some point: Currently it aim is it to implement the
authentication procedure for a smbv2 client.

@alfkil

Where can the latest version of spotless be downloaded? The os4depot version seems to be from 2020 and on GitHub aren’t any releases for download.

@all

just for the curiosity. Has someone a RX580 gfx card successfully running, or are they not supported?

@trgswe

I would probably be aware, if i would targeting 68k. But the __libcall__ attribute was introduced with Amiga OS4, os i thin only PPC targets are of concern of this.

@walkero

Thas nice to hear. Because i just picked one of our docker image from github, to build my envirnoment on. I don't know if it was the best choice, but i made Pull Request to include the latest SDK.

@all

I really recommend anyone having a github account and using Visual Studio Code or thinking of to use it in the future for cross compiling for AOS4 ti vote on the issue.

Hi everyone,

Im using VS code to cross compile for amigaOS 4, like George Sokianos described it in this Video on Youtube: docker based AmigaOS cross-platform development environment

This really works nice, except for that VS Code comes up with "too few arguments.." on each call to a Amiga Library function.

I wanted to go after it and found this posted issue: Intellisense support for gcc __attribute__((libcall)) not found

It seems that if enough interest is presented, one of the folks might add support for APICALL.

So anyone having interest in this feature of VS Code should maybe express it on Github.

@mrupp

Well the killing of the Sonos device only happened with my first run with SonosController. Now that i wanted to verify it before posting, it didn't happen again. And even Auto discovering works really fine.
At work i only use Sonos(Controller) to mute and change radio station, and that works without any troubles, even with v1.2 version. But most of the day the app is just idling.

Oh, i have a Micro AmigaOne with OS4.1 with latest update. No enhancer stuff, because i (A-Eon) don't get my account working. Having 512Mb of RAM.

@walkero

Please let us now how it works with your intended setup.
Because i have a IDE44 to SATA converted on the main IDE
interface, which gives none deterministic and random read
errors from while to while. I'm thinking about to replace
it with a Transcend PSD330, which is a SSD with PSATA (IDE
44) interface. But i think your solution could be nicer.

@mrupp

I have access to SONOS speakers at work, so i tested your v1.1 around the time you released that version.

With my first test, SonosController had trouble to discover all speakers. So i added them one by one via there IP address. That somehow resulted in my colleagues haven't no access to the speakers via the Sonos app. Thus they wanted the control back, and we had a hard time to tell the original Sonos app, again about we speakers. We had to reset them etc, until access was possible. So i didn't use SonosController any further, until now.

I started v1.1 again and, whatever chnanged, now the speakers are automatically discovered. I can control them and my colleagues haven't yet a clue that i do it on my amiga. Maybe the cause was a issue on the company network? So works nice and great.

Next i started v1.2 and works as nice as v1.1, and even the groups are correctly discovered.

I did following tests:

a) On the original Sonos app, i split a group. SonosController did automatically update itself to reflect the change. Nice.

b) Rejoined the group with SonsoController. Was performed and the Sonos app did update itself to reflect the change.

If you want to have something specific tested, let me know. For now i stopped any testing, in the case something goes wrong, so that my colleagues don't get upset on me

@amigakit

Thanks for the second email. I could log on the webpage and change the initial given password.

But if i use the AmiStore application on the amiga, i can not checkout. The app ask for my credential and seems to log in, at least no error comes. But the browser isn't opened (is configured as described). And looking from screenshots, i don't get a checkout button after login. What is wrong with my account?