Back
[07:17:44] <jthornton> jepler, the forum is in read only mode
[07:18:02] <jthornton> I'll be back Sunday afternoon
[07:47:27] <jepler> jthornton: have a good weekend
[07:47:32] <jepler> I suppose that means I have no excuse not to start now
[07:49:51] <jthornton> I guess so, I'll be checking in on my phone from time to time from the river
[09:10:05] <mozmck> hi jepler: I got a "Message Reported" email from the new forum, so I logged in and moderated the message ;)
[09:10:41] <mozmck> hope I didn't mess you up.
[09:11:25] <jepler> mozmck: OK, that's fine
[09:11:44] <jepler> thanks for telling me though, I was trying to (A) find where the report went and (B) then wondering why something had changed
[09:12:33] <mozmck> haha! I changed the subject and added (moderated :-) ) to the end, and it looks like it moved it to it's own topic.
[09:13:10] <mozmck> Looks like it works anyhow.
[09:13:35] <mozmck> I think reported messages go to all moderators.
[09:16:04] <jepler> yup
[09:16:07] <jepler> that's what I see too
[09:27:52] <jepler> now I just need someone with dreamhost panel access so I can fix the DNS
[09:28:33] <jepler> mozmck: you don't have dreamhost panel access do you?
[09:29:18] <mozmck> No, not for linuxcnc.org I have an account with dreamhost so I'm familiar with the panel though.
[09:29:33] <mozmck> I have a phone number for SWP, want me to try and call him?
[09:29:53] <jepler> I dropped him an e-mail but I'll call later today if I don't hear from him
[09:30:26] <jepler> does the number you have end -5201?
[09:30:42] <mozmck> yep
[10:33:20] <skunkworks> jepler: thanks for all your work.
[11:04:16] <seb_kuzminsky> http://highlab.com/~seb/linuxcnc/wlo/
[11:05:13] <seb_kuzminsky> that's a rough first cut of importing wlo's joomla into jekyll
[11:06:00] <seb_kuzminsky> each page is a file in the git repo
[11:08:01] <seb_kuzminsky> https://github.com/SebKuzminsky/wlo
[11:08:11] <seb_kuzminsky> some caveats:
[11:08:14] <seb_kuzminsky> * i dont know what i'm doing
[11:08:17] <seb_kuzminsky> * i dont know css
[11:08:23] <seb_kuzminsky> * i dont know what a website should look like
[11:08:30] <seb_kuzminsky> * but i'm pretty sure it shouldnt look like that
[11:08:37] <seb_kuzminsky> bbl
[11:46:07] <cradek> aside from the content being as out of date as before, I see nothing wrong with this website. I think it looks less bad than joomla did.
[11:46:20] <cradek> caveat: * i dont know what a website should look like
[11:46:49] <cradek> heh, that picture with the blue gui is humorously out of date
[11:47:07] <cradek> I still really like the logo at the top
[11:48:43] <cradek> maybe the front page should have the "posts" in a column (and probably not all of them)
[11:50:16] <cradek> maybe it would be nice if there was a column for the boilerplate (the existing "welcome" text summary at the top) and then the first few "posts" would be displayed inline in a second column, then a More... link for the older posts
[11:51:21] <cradek> the most important information is probably that "2.6.11 bugfix release" latest post, and it's probably not ideal to have the most important information behind a link
[11:52:25] <cradek> oh how lovely it will be if I can edit the website with a git commit
[11:52:35] <cradek> it makes me want to break out into song
[12:01:15] <seb_kuzminsky> i agree there's too much "news" on the front page
[12:01:25] <seb_kuzminsky> that's just the default jekyll layout
[12:01:51] <seb_kuzminsky> i think i want to add a small but easily findable "latest releases" blurb near the top (like the IRC topic line)
[12:01:59] <jepler> swp says he can make the dns changes this evening
[12:02:12] <seb_kuzminsky> and probably prominent links to the "download and install" page, and to the forum
[12:02:27] <seb_kuzminsky> and almost everything else should either be absorbed into the glo docs, or be deleted
[12:03:55] <seb_kuzminsky> much of the content is in html, because that's how jekyll's "extract-from-joomla" script emitted it, it should probably be rewritten in markdown for simplicity later
[12:04:10] <seb_kuzminsky> for now i'm goign to go outside and work in the garden fora bit, bbl!
[12:05:37] <seb_kuzminsky> oh, one last thing
[12:06:20] <seb_kuzminsky> if you want to mess with it, you can clone the git repo i linked above, apt-get install jekyll, then "jekyll serve --watch" in the root of the working dir
[12:06:20] <awallin> anyone listening to the tormach webcasts (g+ hangout?) going on? or are they later today..
[12:06:46] <seb_kuzminsky> i'm using jekyll 2.2.0 from jessie, the version in wheezy is much older and i dont know how well it would work
[12:07:35] <jepler> I don't think requiring Jessie to work on the wlo website is a big problem.
[12:13:31] <awallin> is this someone from this channel giving the talk?
https://www.youtube.com/channel/UC3FVtrJ8U91iODq5motHMow
[12:17:00] <awallin> mhaberler?
[12:34:41] <seb_kuzminsky> jepler: i agree
[12:35:22] <seb_kuzminsky> awallin: mhaberler used to be a linuxcnc developer, but he forked the project and no longer works here
[12:36:36] <awallin> seb_kuzminsky: ah ok, seems machinekit and tormach work together now
[12:53:37] <pcw_home> I think PathPilot is a mish-mash of both
[12:54:17] <pcw_home> +Tormachs intrepreter changes
[12:55:04] <pcw_home> (and their GUI of course)
[13:32:38] <skunkworks> actually I think it is based on 2.7 at the moment...
[13:33:01] <skunkworks> other than all of their mods to it
[13:36:39] <seb_kuzminsky> maintaining a local fork of an open-source project can be tricky business, but there are some tricks that make it much easier (that git is good at enabling)
[14:02:32] <jepler> Subject: [ANNOUNCE] v4.1.12-rt13
[14:04:12] <pcw_home> They sure are busy
[15:08:14] <jepler> it's excellent news that they are
[15:19:14] <pcw_home> yeah it looked pretty bleak for a while
[15:59:20] <JT-Mobile> How's it going
[16:01:06] <Tom_itx> out for a saturday PM stroll?
[16:01:32] <JT-Mobile> Down by the Current River
[16:01:56] <JT-Mobile> Won a cabin for the weekend
[16:02:07] <Tom_itx> nice
[16:02:51] <JT-Mobile> At least the sun is out
[16:03:15] <Tom_itx> yeah, nice day here.. cool but nice
[16:03:25] <JT-Mobile> I wonder how jepler is doing on the forum?
[16:03:52] <Tom_itx> dunno, i seldom if ever get on it
[16:10:13] <JT-Mobile> Gotta go
[16:22:56] <jepler> jt-mobile: sorry I missed you. it's all over but the DNS change..
[17:25:39] <jepler> jt-mobile: sorry I missed you. it's all over but the DNS change..
[17:26:53] <JT-Mobile> Cool
[18:10:37] <jepler> whoops, didn't mean to reboot *that* computer
[19:18:51] <jepler> the new forum will be live as soon as DNS propagates. thanks to swp for showing up to do his magic on our dns hosting
[19:20:36] <Tom_itx> so that officially separates the webpage and forum
[19:25:16] <jepler> yes
[19:28:37] <jepler> and as seb_kuzminsky was saying earlier this weekend he's looking at redoing the "main" website with jekyll, a static site generator
[19:29:07] <Tom_itx> nice.. it's long overdue
[19:29:20] <jepler> afk for dinner
[19:43:05] <seb_kuzminsky> i invite feedback on the new website, previewed here:
http://highlab.com/~seb/linuxcnc/wlo/
[19:43:47] <seb_kuzminsky> i especially invite patches and pull requests, source here:
https://github.com/sebkuzminsky/wlo
[19:44:33] <seb_kuzminsky> cradek: i did part of what you suggested, now only the 4 most recent posts show up on the front page, and all posts on the "News" page
[19:46:09] <jepler> you might want to kill the "welcome to jekyll" news item :-)
[19:46:15] <jepler> mmm bowl of chili
[19:46:42] <jepler> but something horrible happened: I have opened what appears to be the last jar of pickled jalapenos I made earlier this year :-(
[19:46:43] <seb_kuzminsky> oh yeah
[19:47:34] <jepler> seb_kuzminsky: I assume the doc links would work once the jekyll website was integrated on wlo
[19:47:40] <seb_kuzminsky> that's kind of cool... i ran 'git rm _posts/2015-11-04-welcome-to-jekyll.markdown', and 2 seconds later the site auto-updated
[19:47:40] <jepler> for now they're 404'd
[19:47:51] <seb_kuzminsky> jepler: yeah, that's the intent
[19:48:30] <seb_kuzminsky> those links point at {{site.url}}/docs/mumble, which will be right once the _site tree is at
http://linuxcnc.org/
[19:48:48] <seb_kuzminsky> there are probably other broken links and missing images and things
[19:49:03] <seb_kuzminsky> and of course the actual content is hilariously out of date
[19:49:31] <seb_kuzminsky> http://highlab.com/~seb/linuxcnc/wlo/about/board-of-directors/
[20:13:18] <jepler> I'm not always sure whether it's good or bad that we essentially lack governance
[20:18:55] <jepler> anybody getting forum.linuxcnc.org yet?
[20:20:26] <Tom_itx> yes
[20:21:06] <Tom_itx> it's alive!
[20:21:34] <seb_kuzminsky> ooh, i'll update the Forum link on the new website ;-)
[20:21:59] <mozmck> jepler: I just got an email from the new forum - looks like it's up here!
[20:23:10] <jepler> huh .. where did the rss link go?
[20:26:21] <seb_kuzminsky> jepler: i feel like our main problem is lack of developer-time, not lack of process and governance
[20:27:19] <seb_kuzminsky> what process we have is implemented by the test suite & the buildbot, and then constrained by the lack of human eyeballs for review & feedback
[20:31:34] <micges> I think lack of governance is main problem
[20:31:38] <jepler> weird, on my chromebook the rss image is being blocked by ublock origin
[20:35:06] <seb_kuzminsky> micges: please elaborate
[20:38:18] <micges> as I started developing (>2009) we had this problem, in times when we have plenty developer-time this leads to chaotic commits flow
[20:39:06] <micges> one playing with runtests, one in motion, one in interp, noone checked api changes
[20:39:49] <micges> we lost plenty new developers who wanted to make some changes
[20:42:13] <seb_kuzminsky> micges: do i understand you correctly, you are saying: lack of oversight/governance leads to too many commits in different areas, which leads to internal api breakage?
[20:42:25] <micges> they wanted to fix something, there was no info about: project design, rules of approving patches, time plan of releases, future directions
[20:43:31] <micges> seb_kuzminsky: commits in different areas without main idea without general plan for project
[20:44:19] <jepler> at a technical level, git (which we were not yet using in 2009) is helpful for that
[20:44:31] <seb_kuzminsky> i disagree
[20:44:59] <jepler> because a project can be published and receive feedback in a way we couldn't do with cvs
[20:45:18] <seb_kuzminsky> i think micges is pointing out the lack of project guidance, so contributors don't know what contributions would be welcome, or how those contributions should fit into the overall project structure and future plans
[20:45:27] <seb_kuzminsky> micges: is that representing your position accurately?
[20:46:05] <micges> seb_kuzminsky: yes
[20:47:31] <seb_kuzminsky> git is useful because you can implement something and then ask for feedback on the code, but micges is suggesting (as i understand it) that we lack the governance to provide this feedback in a timely/accurate manner
[20:48:15] <seb_kuzminsky> i perceive the problem to be lack of time on the "core developers" part, not lack of bylaws and official processes
[20:49:20] <jepler> reviewing something as big as ja or features is very hard for one person to do
[20:49:29] <seb_kuzminsky> micges: do you have suggestions for how to fix the problem you perceive?
[20:50:13] <seb_kuzminsky> i've spent countless nights trying to get things like ja and ubc into my head
[20:51:14] <ssi> for an outside perspective, it's very hard to break into contributing
[20:51:26] <micges> seb_kuzminsky: not at the moment as it's 3 am here
[20:51:29] <ssi> I've wanted to contribute a lot more than I have, but it's a tough ecosystem to get immersed in
[20:51:57] <ssi> and I think micges's points speak to that to some degree
[20:52:08] <seb_kuzminsky> micges: no rush, input is welcome any time
[20:52:43] <seb_kuzminsky> micges: i agree that we seem to have a problem, though i dont agree with you on exactly what the problem is, and i'm not at all sure how to fix it
[20:52:48] <micges> while working with ja I hit few of problems talked here
[20:53:22] <seb_kuzminsky> ssi: i'd love to hear your input on what the difficulty in getting in to our ecosystem is. is it technical? political? social?
[20:53:45] <ssi> heh a little of all of the above honestly
[20:53:51] <ssi> it's a HUGE technical challenge of course
[20:53:55] <ssi> because it's a very complex project
[20:54:24] <ssi> I'm in the middle of developing a smartserial remote at the moment, and even just that was a herculean task to get spun up on how sserial works and how it interats with hal via hm2
[20:54:30] <seb_kuzminsky> micges: heh, my view of the difficulty of ja is that it did not come with a clearly explained design, so as a part-time slacker reviewer it was/is diffcult to understand what all those commits are trying to do
[20:54:59] <seb_kuzminsky> ssi: so, lack of technical documentation, in that particular situation?
[20:55:06] <ssi> I personally feel like git is a good choice, but tools like github really improve it a lot with respect to contributions and reviewing
[20:55:16] <ssi> but github isn't favored in this group
[20:55:21] <ssi> there's plenty of technical documentation
[20:55:28] <ssi> but it's mostly reference docs
[20:55:34] <ssi> it's very hard to get into it "from scratch"
[20:55:55] <ssi> it's hard to know what docs apply to what you're trying to accomplish
[20:55:56] <micges> seb_kuzminsky: right
[20:55:59] <seb_kuzminsky> github PRs are very welcome:
http://linuxcnc.org/docs/2.7/html/code/contributing-to-linuxcnc.html#_linuxcnc_on_github
[20:56:27] <ssi> ah ok good
[20:56:47] <seb_kuzminsky> ssi: i totally agree our docs (specifically our technical docs) could use a lot of work
[20:56:47] <ssi> when the git move first happened I remember discussion about it, and the outcome at the time seemed like eschewing public repos was the plan
[20:57:15] <ssi> there are a lot of good techincal resources (in terms of people) in here, the forums, and the list
[20:57:36] <ssi> but a lot of times they are unable or unwilling to dumb down to the level of someone trying to get a foothold into the code
[20:57:38] <seb_kuzminsky> ssi: yeah, having people to talk to is good, but it's also good to have docs to read on your own time
[20:58:12] <seb_kuzminsky> "dumbing down" is really hard - the devil's in the details
[20:58:15] <ssi> yes I know
[20:58:26] <ssi> and it's hard to know how far to dumb down
[20:58:34] <ssi> do you need to start with "this is what we call an integer"?
[20:58:35] <seb_kuzminsky> yeah :-/
[20:58:35] <ssi> :)
[20:59:03] <ssi> since it's fresh in my mind, I can speak a bit on the topic of the sserial stuff
[20:59:07] <ssi> it's not really linuxcnc though
[20:59:33] <ssi> but mesa has docs for sserial, but they're written from the perspective of someone writing code for the proc in the hm2 firmware
[20:59:40] <seb_kuzminsky> we have john thornton working on the docs, and the rest of us patching and expanding here and there, but we could always use more people working on that (same as everything else)
[20:59:48] <ssi> as someone trying to write a remote, it's been challenging trying to figure out how the docs apply to my situation
[21:00:02] <seb_kuzminsky> i bet
[21:00:05] <ssi> we did most of it by reverse engineering what the 7i76 did by protocol dumping the line
[21:00:12] <ssi> and now that I understand it, the docs make sense
[21:00:20] <ssi> but it was just a mess of out of context technical data initially
[21:00:27] <seb_kuzminsky> did you write down the new things you learned, that makes the existing docs make sense?
[21:00:28] <ssi> and linuxcnc's codebase feels similar to me
[21:00:40] <ssi> no, not really, although I did report some doc bugs back to pcw
[21:00:43] <seb_kuzminsky> that would be a very valuable contribution, i bet you know more than just about any of us at this point
[21:01:06] <ssi> I'll roll that around in my head a bit and see if i can come up with something
[21:01:45] <ssi> there have definitely been some surprises that i only discovered empirically
[21:02:26] <seb_kuzminsky> we're all like that
[21:02:54] <seb_kuzminsky> some of us have been at this longer than others, that's all
[21:03:00] <ssi> :)
[21:03:23] <seb_kuzminsky> i started in 2007, and i totally fucked up the first version of the hostmot2 stepgen because i totally misunderstood what the commanded position meant
[21:03:43] <ssi> the last time I was actually in the linuxcnc codebase was writing hal drivers for beaglebone black hardware under mhaberler's guidance before the machinekit stuff forked
[21:04:43] <ssi> but I'll tell you this
[21:05:06] <ssi> I'm a professional software engineer by trade, I've been coding for money for 20 years, and coding for almost 30
[21:05:22] <ssi> and there have been plenty of times trying to work with this stuff that I've thought "I don't think I'm good enough to do this"
[21:05:34] <ssi> I'm sure lots of potential contributors have that problem
[21:06:00] <seb_kuzminsky> heh, same here... i've been coding for money since my undergrad days in the mid 90's
[21:06:21] <seb_kuzminsky> linuxcnc is one of the most complex code bases i've ever worked with
[21:06:27] <ssi> yes
[21:06:38] <ssi> and I wrote much of the engine that runs cnn.com :P
[21:07:35] <seb_kuzminsky> there are ways i can imagine to make it less complicated, but they're huge efforts, and not guaranteed to succeed
[21:07:52] <ssi> part of the problem is diving back into pure C and embedded code is challenging after working in higher level systems predominantly
[21:08:15] <Tom_itx> is it a bucket of patches that has been handed down from the dark ages?
[21:08:41] <ssi> Tom_itx: what, cnn? no, fortunately we get to throw it all away and start over every couple years :)
[21:08:45] <seb_kuzminsky> i started the liblinuxcnc-ui project with jepler in the hope of simplifying the UI/Task interface, but that's been stalled for over a year now
[21:08:45] <ssi> the CONTENT however is
[21:08:47] <Tom_itx> no, lcnc
[21:09:14] <seb_kuzminsky> Tom_itx: the age of the project is definitely implicated in its complexity
[21:09:25] <ssi> seb_kuzminsky: this sserial work is definitely the most satisfying and exciting work I've done with linuxcnc's codebase
[21:09:41] <seb_kuzminsky> awesome :-)
[21:09:44] <Tom_itx> i know i've been non-actively following it for a good many years
[21:09:46] <seb_kuzminsky> i'm looking forward to seeing it :-)
[21:10:13] <ssi> it's someone else's project, but I'm happy to be a part of it
[21:10:26] <ssi> it's a servo drive somewhat like the 8i20, sserial command and feedback
[21:10:31] <ssi> (among other options)
[21:10:50] <ssi> but it ought to be a good position mode drive for linuxcnc, which I hope will drastically simplify cabling complexity
[21:11:02] <seb_kuzminsky> i'm just some guy, but my opinion is this: linuxcnc is just the bunch of us, working together to scratch our own itches and occasionally make the shared project better/simpler/cleaner, and the project progresses to the extent that we work to make it better, individually and together
[21:11:15] <ssi> seb_kuzminsky: I love that attitude
[21:11:21] <ssi> I don't always feel that attitude coming from the group :)
[21:11:29] <Tom_itx> yeah
[21:11:37] <jepler> yeah we fail at it sometimes
[21:15:15] <ssi> the contributing to doc is really good btw
[21:15:21] <ssi> espceially the notes on effectively using git
[21:15:25] <ssi> that's something I don't know well enough
[21:15:34] <ssi> stuff like using rebase to fix up the commit history
[21:15:49] <ssi> I very much fall into the xkcd camp with git
[21:16:04] <ssi> http://www.xkcd.com/1597/
[21:16:24] <ssi> I like to think I'm smart enough to use git effectively, but it often humbles me :)