Your most wanted feature for the 2019 Hackfest?

[cb88]

No probably not because it isn't that useful.

Something else I just though of you might check around on your motherboard if there is debug header or even solder pads for serial debug output, debug output for kernel debugging will likely remain over serial long long into the future. The only way around this would be if the Bios/UEFI implemented some facility to reconfigure the Ethernet/USB for debugging by simulating a serial port etc...

I don't know why you think a serial port card cost $90.. https://www.ebay.com/itm/PCI-E-Express- ... SwNaRavEyz

That's $5 shipped to the continental USA or most of the world and there are $10 models on amazon if you are in Alaska even. Someone could probably forward it to you in a flat rate box if you are in the non continental USA. For stuff worth under $5-10 you don't pay tax in the EU either AFAIK. Frankly it's hard to imagine where you could be that shipping + tax costs $40 on a $5 device, I've lived in Brazil even that has insane tax rates.

[middings]

We have three options for modern debugging routes:232 debugging cards costing upwards of 100 USD in California and overall a terrible plan...

I don't know what you mean by "(RS-)232 debugging cards". ReactOS and WinDbg use a PC's standard RS-232 connection that is provided on many motherboards or by an adapter card that plugs in to a PC's ISA, PCI, or PCIe connection.

Fry's Electronics stores in California sell PCI-to-RS232 adapter cards for much less that 100USD. PCIe adapter cards are also available. Fry's may even have an ISA adapter card for sale. Suitable adapter cards are also sold on eBay and elsewhere for much less than 100USD.

(E)thernet cable debugging [...] would be the easiest to program in theory because we have network drivers already and could use them to send data to another computer.

Debugging ReactOS over an Ethernet cable would have advantages. Unfortunately, RS-232 hardware and its binary interface in PCs (personal computers) are standardized but those of network interface controller (NIC) hardware in PCs are not. The debugger code in the target computer (the computer being debugged) must know a lot about the target computer's specific Ethernet network interface controller (NIC). As a result, only one or a few NICs might be supported. (The PC I use to test ReactOS cannot use the network drivers included with ReactOS.) If so, ReactOS could not be debugged over the Ethernet connection on many (most?) PCs. For this and other technical reasons, I agree with cb88 that debugging with RS-232 ports will remain with us into the foreseeable future.

[Missingno50]

The nearest frys is a couple of hundred miles away from me AFAIK. On top of that, I get taxed for each package to my house by my landlord, and its not cheap. The tax is there to encourage me to go to a store but still.
Edit:100$ may be an exaggeration but tacking on 40$ is not according to my landlord.

[middings]

Missingno50, in an earlier post elsewhere in the ReactOS forums, you mentioned the Sacramento State University (Sac State) campus. Based on that, I assumed that you could travel to the Fry's Electronics stores in Sacramento and Roseville, California without much difficulty.

[Quim]

BtrFS support in ROS is merging in Hackfest...and it is an important feature to use a newer decent filesystem... but,

... what about vgal work on USB system?

Who are responsible also testing vgal USB code in order to included it officially in 0.4.10 official release?
If not, before the end of this year ReactOS *must* have a full funcional USB 2.0 system or at least include vgal code in official ROS builds....
one important thing is full functional LiveUSB...

[binarymaster]

but,

... what about vgal work on USB system?

Good question, I'm also wondering about the status, and what have been done at the hackfest regarding it.

[hbelusca]

Due to the current quality of the code, it needs a very extensive review.

[karlexceed]

Good question, I'm also wondering about the status

You can always look at the PRs on Github yourself...

https://github.com/reactos/reactos/pull/245
https://github.com/reactos/reactos/pull/301
https://github.com/reactos/reactos/pull/308

[binarymaster]

Due to the current quality of the code, it needs a very extensive review.

Is the code quality really so poor?

You can always look at the PRs on Github yourself...

Thanks Captain Obvious, but I'm wondering what exactly have been done on hackfest re. USB, e.g. real hardware tests, regression tests in buildbots, etc. (at least some work towards the USB boot)

If nothing have been done regarding USB patches on hackfest, just tell me clearly.

[karlexceed]

Thanks Captain Obvious

I try not to assume what people do or do not know. Seeing as the Hackfest ended yesterday and the PRs haven't changed nor do I see any recent commits related to them... I assume very little happened as far as VGal's USB stuff.

[EmuandCo]

Check pulls and of course commit log. And if you are so keen to know before the report is out.... ask where you find the devs aka not in here ^^

[binarymaster]

I try not to assume what people do or do not know. Seeing as the Hackfest ended yesterday and the PRs haven't changed nor do I see any recent commits related to them... I assume very little happened as far as VGal's USB stuff.

Alright. I'm watching closely and tracking any movement towards the USB boot.

[lelekx]

IMO, it should be quite easy to introduce basic debugging over Ethernet:
- Using a dedicated adapter - let's assume that we are choosing cards from the Intel e1000 family - they are solid and, as far as I remember, the driver does not have to put any firmware into it,
- Use only UDP traffic, you will not need to use a rather complicated TCP stack,
- wrap it around the KDB serial interface.
Pros:
- Virtualbox and probably other hypervisors emulate such a card,
- much faster than UART,
- The card is quite popular and available for PCI, PCIe and ExpressCard ports.
Kons:
- The e1000 controller driver is not yet ready (but already WiP),
- Write and test the code for bootloader, debugging at target kernel and host.

[ThFabba]

IMO, it should be quite easy to introduce basic debugging over Ethernet:

That's fantastic, please do it!

I'm not being sarcastic. ROS does not suffer from a lack of ideas, it's just that all devs already have over-full TODO lists. So if you think something is a priority, please go ahead and work on it. It would be much appreciated.

[mrmajik45]

It could help with laptops.