11:26 <Guest11> good morning folks. I'm using OpenOCD to work on a Nuvoton NUC123, and the chip ID is not recognised. I have added it, recompiled openocd, and can confirm it's working fine. What's the right process to have it included in a future release?

11:27 <Guest11> openocd-0.12.0~rc2/src/flash/nor/numicro.c:{"NUC123SD4AE" , 0x10012315, NUMICRO_BANKS_NUC100(68*1024)},

12:27 <Haohmaru> the right process begins with having patience

12:41 <PaulFertser> Indeed

12:41 <PaulFertser> And reading docs.

12:51 <Haohmaru> nobody reads documents, but at least have patience ;P~

12:52 <braunr> some do

12:52 <PaulFertser> Guest11: you seem to be the impatient fellow

12:53 <Guest11> nah, my laptop went to sleep, and the web client dropped the connection

12:53 <PaulFertser> Guest11: we were just discussing how IRC requires more time

12:53 <PaulFertser> Guest11: there's HACKING in the root of the source tree, you should be able to follow the instructions for contributing there.

12:53 <PaulFertser> Guest11: should be all clear, and feel free to ask if anything is not.

12:56 <Haohmaru> my kitchen window isn't very clear ;P~

12:57 <PaulFertser> I recommend IPA + a newspaper.

12:58 <Haohmaru> i know that trick ;P~

12:58 <PaulFertser> And IPA beer ingested after the process is complete.

13:02 <Haohmaru> nooo

13:03 <Haohmaru> pepsi

13:06 <Guest11> hmmm. gerrit account created, username set, ssh key uploaded, but getting access denied when pushing or even when SCPing the git hook

13:06 <Guest11> do I need to wait for any kind of account activation?

13:07 <Guest11> btw PaulFertser you "seem to be impatient" is quite funny, given that I'm working on this pet project _since one year_ and only in the last week I managed significant progress :D

13:08 <PaulFertser> Guest11: no account activation is needed

13:08 <PaulFertser> Guest11: probably you're missing the Gerrit port? It's not on tcp/22.

13:08 <PaulFertser> So when scp'ing you need to specify it.

13:09 <Guest11> 29418 - I'm copy/pasting from the docs

13:09 <PaulFertser> Guest11: I know how the first impression is often wrong, no wonder :)

13:09 <PaulFertser> Guest11: please use some pastebin to show the commands and output.

13:10 <Guest11> let me check first if there's anything wrong with my ssh keys...

13:11 <PaulFertser> Yes, with ssh -v it's more visible what key is really tried.

13:14 <Guest11> https://pastebin.com/d1v4hG8E

13:17 <PaulFertser> Guest11: probably you're using an outdated RSA key? Have you tried ssh -o 'PubkeyAcceptedKeyTypes +ssh-rsa' ?

13:23 <Guest11> it works with that option, indeed. thanks for the hint, I haven't considered that

13:25 <PaulFertser> It's not really the key itself but the algorithm, and ssh on the Gerrit server can only use this outdated algorithm with RSA keys.

13:25 <PaulFertser> And newer ssh client on your side doesn't like it.

13:28 <Guest11> well, ssh/config to the rescue. and ta-dah https://review.openocd.org/c/openocd/+/7417

13:28 <PaulFertser> Guest11: thank you for the patch! :)

13:29 <Guest11> my pleasure

13:29 <PaulFertser> Using an ed25519 key would be a working alternative too.

13:29 <Guest11> plenty of other services still not supporting them, tho :(

13:31 <Guest11> thank you for your software! the tcp server works great, I've been debugging this thing from the comfort of my living room for _hours_ by using a raspiberry pi3 as interface

13:31 <PaulFertser> Guest11: btw, that extra bit in the device id, have you checked the datasheet if it's probably safe to ignore?

13:31 <Guest11> hm, I haven't but let me check

13:32 <Guest11> https://www.nuvoton.com/resource-files/DS_NUC123_Series_EN_Rev2.04.pdf can't even find a list of those IDs in the datasheet

13:32 <PaulFertser> My idea is that it's probably a minor revision so to be more future-proof it might make sense to mask it out.

13:33 <Guest11> it's not a minor revision, but a different spec chip

13:33 <Guest11> the difference being...the temperature range

13:33 <PaulFertser> Guest11: my pleasure to hear about your experience, as I was the original author of the raspberrypi-native adapter driver.

13:34 <Guest11> with "N" being -40/+85, and "E" being -40/+105

13:34 <Guest11> not even sure why this E model ended up in here, this thing is _never_ going over room temperature :D

13:35 <Guest11> and I don't have access to any other chip from that series to check if that bit is used to distinguish between N and E models, sorry

13:36 <Guest11> oh, and that GPIO/SWD is much more stable than my stlink clone

13:36 <PaulFertser> Guest11: np. I just compared it with the line above.

13:37 <Guest11> btw, if anyone around here works with NUC123, I could use a pre-built bootloader with HID/DFU capabilities :D

13:38 <Guest11> they have samples, I even managed to compile, but they're all too large to fit in the bootloader rom

15:12 <Guest11> after so many years of github/gitlab, gerrit is a....very different way of working

15:23 <PaulFertser> Guest11: I find it intuitive enough once you see it's not deviating from Git at all and just uses Change-Id fields for tracking which commit belongs to what change.

15:25 <Guest11> I still didn't manage to figure out how multiple commits could be part of the same change, understood that they should have the same change-id, but then the push got rejected because there was a duplicate change-id

15:31 <PaulFertser> Guest11: yes, because they can't :)

15:32 <PaulFertser> Guest11: you do not need multiple commits to be the same change.

15:33 <PaulFertser> If you need to somehow group many commits you can specify a "topic" for them.

15:33 <PaulFertser> But it's really optional.

15:34 <Guest11> well, not a problem for today. the changes are there, pending approval

15:35 <PaulFertser> Guest11: if you think about ordering of your commits then there's nothing to do: every commit always has a reference to its parent(s), that's how Git works.