pribib | did someone set up a git server yet | 04:20 |
---|---|---|
houkime | for what exactly? there's one for sch and layout, also there's a git for the website | 04:22 |
houkime | also there's an issue tracking repo and a planning repo | 04:23 |
pribib | https://notabug.org/Houkime/Neo900-Planning | 04:23 |
pribib | this says it is highly needed | 04:23 |
houkime | oh. you see, there's a thing. with a nice interface. | 04:24 |
houkime | the system we use right now is not nice in any way | 04:25 |
houkime | and not really easy to contribute to | 04:25 |
pribib | what about this https://git.zx2c4.com/cgit/ | 04:25 |
Joerg-Neo900 | how about improving it? | 04:25 |
Joerg-Neo900 | pribib: we prefer hosting on our own servers | 04:25 |
pribib | you can run that on your servers | 04:26 |
pribib | with gitolite | 04:26 |
pribib | i think | 04:26 |
Joerg-Neo900 | I think we run gitolite | 04:26 |
Joerg-Neo900 | I suffered quite some rant about this particular fact | 04:27 |
pribib | https://git.parabola.nu/ | 04:27 |
Joerg-Neo900 | ? | 04:28 |
pribib | this uses cgit | 04:28 |
houkime | I probably need to explain a bit what i meant (and actually put details in the planning repo) | 04:28 |
Joerg-Neo900 | sorry I can't follow how that's related | 04:29 |
houkime | sorry about that | 04:29 |
pribib | cgit is a web frontend that can be integrated with gitolite | 04:29 |
Joerg-Neo900 | we need to follow an approach of "what function does the user interface need? how can we implement that?". Not "what softare is available, which looks cool? who's going to install it and then we see if it is ok for our requirements" | 04:32 |
pribib | oh i see | 04:32 |
Joerg-Neo900 | we switched forth and back between iirc 4 different git-esque solutions, since everybody has another pet | 04:33 |
houkime | I actually agree on functional approach. A problem which I currently see with the stuff we use now is the obscurity of ways you can submit a PR | 04:35 |
houkime | Probably one just needs to write a CONTRIBUTING guide but an actual interface for that would be super nice | 04:36 |
Joerg-Neo900 | yeah, it's a tad obscure since we had no many PRs so far. So we managed that on a per-instance basis and found a workable method for every contributer | 04:36 |
Joerg-Neo900 | let me put it this way: whatever you throw at us for your "PR", we probably will find a way to cope with it | 04:37 |
Joerg-Neo900 | actually I think there's no such thing like pull requests in our system | 04:38 |
Joerg-Neo900 | send a patch per email. Or ask for an account so you can git-push | 04:39 |
Joerg-Neo900 | we could use a ML particularly for such stuff | 04:39 |
Joerg-Neo900 | I mean, it works for kernel, no? wgould be fine for us too | 04:40 |
Joerg-Neo900 | should* | 04:40 |
houkime | Linux has more then just a ML. there's a whole system there actually. And it started very long time ago - it would be very strange to think these methods can't be improved on, especially when they have been improved with forking and prs | 04:43 |
houkime | now the core problem we have already discussed is not even this | 04:44 |
houkime | it is the poor compat of kicad and git | 04:44 |
houkime | and bad/unobvious diffsets | 04:44 |
Joerg-Neo900 | err, that's why Neo900 invented eeshow | 04:45 |
Joerg-Neo900 | http://neo900.org/stuff/kicad/proto_v2/2016-11-20/review-howto.html | 04:45 |
Joerg-Neo900 | and the kicad devels (well, several of them at least) were keen to include eeshow concept into kicad genuinely | 04:46 |
Joerg-Neo900 | which to me suggests that it isn't all bad | 04:47 |
houkime | oh, that's the great news about kicad devs | 04:47 |
Joerg-Neo900 | hm? | 04:48 |
houkime | I just remember last time you were upset about them rejecting everything you propose and having other plans | 04:49 |
Joerg-Neo900 | when you're interested in contributing to Neo900 on a technical (EE) level, you should make familiar with https://neo900.org/stuff/eeshow/#motivation | 04:49 |
Joerg-Neo900 | no, you got that wrong | 04:49 |
houkime | ok | 04:50 |
houkime | we still have nothing to actually review diffs for layout though | 04:51 |
houkime | and maybe for sch also. Eeshow doesn't really seem to have the ability to highlight diffs, or has it? It is aware of commits, but diffs? | 04:53 |
houkime | i mean, you can't really afford to recheck EVERYTHING on every patch, right? | 04:55 |
houkime | the diff should be located even if it is big formally in file | 04:55 |
Joerg-Neo900 | huh? | 05:00 |
houkime | pribib: I might have been wrong about need for large software changes, but I really think the contribution workflow should be streamlined or at least have some detailed instructions. | 05:01 |
pribib | i dont know what i have to offer to help with that | 05:04 |
houkime | My first thought about that was to go self-hostet gogs or gitea to support easier fork+PR, but probably there're better ways | 05:05 |
houkime | and there's still a fact that the way one reviews prs/patches for kicad stuff should be streamlined | 05:06 |
houkime | even more | 05:07 |
Joerg-Neo900 | did you ever have a complete read of https://neo900.org/stuff/eeshow/ ? Does this look like a decent diff for a schematics https://i.imgur.com/z2URU6s.png , to answer the question "what been done in commit 57eebdc by Werner Almesberger (https://i.imgur.com/zGaEiyF.png)? | 05:07 |
houkime | oooh... Yep, I'm blind. | 05:09 |
houkime | sorry Joerg | 05:09 |
Joerg-Neo900 | asking for streamlinig is pointless without constructive criticism and suggestions where to improve what in which particular way | 05:09 |
houkime | so it is down to nice contribution instructions then. ok. | 05:12 |
houkime | btw, I haven't found where the website git repo is. | 05:13 |
houkime | atk says there is one but i've never seen it | 05:14 |
Joerg-Neo900 | yeah, resources lacks a link to the git repo | 05:14 |
Joerg-Neo900 | https://neo900.org/git/www/ | 05:15 |
houkime | thx | 05:15 |
Joerg-Neo900 | yw | 05:15 |
Joerg-Neo900 | >>so it is down to nice contribution instructions then.<< yes, never can get verbose and encouraging and visible enough | 05:19 |
Joerg-Neo900 | however note that it's not exatly impossible to find http://neo900.org/stuff/kicad/proto_v2/2016-11-20/review-howto.html - and we got zilch response on that one | 05:20 |
houkime | obscurity is bad. Actually even long multistep contribution procedures i think are bad that's why I really like fork+pr workflow | 05:24 |
houkime | makes it simple | 05:25 |
Joerg-Neo900 | err | 05:25 |
Joerg-Neo900 | go ahead, nobody stops you | 05:25 |
houkime | probably time for me to actually write guides and do small website changes | 05:26 |
houkime | so that even a person as blind as me can find everything | 05:27 |
enyc | coo | 05:27 |
houkime | pribib: thanks anyway for initiating discussion | 05:32 |
pribib | no problem | 05:33 |
houkime | pribib: feel free to post issues, suggestions and whatever. | 05:33 |
houkime | pribib:I will update planning repo tomorrow, because it's a sleep time | 05:33 |
houkime | pribib: aaand probably will try to stream stuff | 05:34 |
houkime | ok, updated neo900-planning a bit | 22:04 |
houkime | now, to publishing a nice layout planning in the tech issues repo | 22:06 |
houkime | and finally to work. There was a long hiatus that needs to be ended | 22:06 |
houkime | now I think that finding people is equally important because alone I obviously can't do it fast enough due to IRL | 22:09 |
houkime | It would be very bad if the thing becomes obsolete before the birth | 22:11 |
houkime | ok, some planning in place | 22:58 |
sicelo | neo900.org cert expired | 23:26 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!