#ossasepia Logs for 18 Dec 2019



April 21st, 2020 by Diana Coman
lobbes: diana_coman: so I had a brief exchange with trinque and he suggested a linux-from-scratch exercise as a way to get a better grip on the fundamentals of an OS. [00:15]
ericbot: (trinque) 2019-12-18 trinque: LFS will teach you more about what actually constitutes a working system, because you'll make one with your hands [00:15]
lobbes: It actually sounds like a good idea to me, because part of my reason for wanting to help the tmsr os effort was to pay some of my learning-debt, but as I chew through the ebuild/Portage study I'm coming to terms with the basic fact that the debt I carry is quite large. [00:15]
ericbot: Logged on 2019-12-12 22:35:03 diana_coman: as to remedial/learning-debt-paying, that should get some place at some point too, [00:15]
lobbes: I'm thinking it may be prudent for me to instead just focus on paying that learning-debt down before I try to directly contribute to tmsr os (I fear I may just slow others down atm). I wanted to see what you thought about it anyway. [00:15]
jfw: BingoBoingo, diana_coman: http://welshcomputing.com/jfw_travel-cert.asc [00:59]
jfw: (and to whomever else my travel key may concern.) [01:01]
whaack: diana_coman: EOD Report: I had to edge into the night a bit, but I finished the schedule: writing exercise (2h), saltmines (8h), proofreading writing exercise (35min), preparing tomorrows writing exercise (15min), setting up postgres on VM (3hr), Spanish study (1h). The writing went overtime because I finished at t=90mins without having done a proofread. I have my VM setup. I can connect my local slime to an sbcl process running on [03:20]
whaack: the VM and spin off logbot threads. I have a test bot running in #whaacked. [03:20]
diana_coman: http://ossasepia.com/2020/04/21/ossasepia-logs-for-18-Dec-2019#1013362 - when do you think you'd be done with a lfs exercise? [06:24]
ossabot: Logged on 2019-12-18 00:15:22 lobbes: It actually sounds like a good idea to me, because part of my reason for wanting to help the tmsr os effort was to pay some of my learning-debt, but as I chew through the ebuild/Portage study I'm coming to terms with the basic fact that the debt I carry is quite large. [06:24]
ericbot: Logged on 2019-12-12 22:35:03 diana_coman: as to remedial/learning-debt-paying, that should get some place at some point too, [06:24]
diana_coman: lobbes: to be explicit: it does sound like a good idea and a reasonable way to get you to figure out the basics, yes; at the same time though: 1. you can't now take the next 6 months to figure out what linux is; 2. you already did *some* investigation of the ebuilds system and so you should at the very least wrap that up and publish whatever you figured out (for yourself at a later stage if for nothing else); 3. at this stage it's ... [06:29]
diana_coman: ... anyway unclear if the ebuilds system is going to be of any use so the task there is *exploratory* really, not more and as such, it's something that can & should include documentation/research as needed, without dropping it all; 4. you can *use* lfs as a resource to learn, it doesn't have to be a project by itself. [06:30]
diana_coman: understand also that for almost ~any practical work/project, you'll always have to ride this double-beast of a. making the most of what you *already* know b. learning and improving your knowledge+skill as much as possible as you work. [06:34]
diana_coman: (although I think I said the above some 10 times at least by now, apparently didn't quite say it in the logs so now it's here to link in the future, too.) [06:35]
diana_coman: lobbes: what's your current deadline+deliverable for the ebuilds investigation anyway? [06:36]
diana_coman: http://ossasepia.com/2020/04/21/ossasepia-logs-for-18-Dec-2019#1013366 - your travel key concerns ...you! silly. [06:37]
ossabot: Logged on 2019-12-18 01:01:31 jfw: (and to whomever else my travel key may concern.) [06:37]
diana_coman: http://ossasepia.com/2020/04/21/ossasepia-logs-for-18-Dec-2019#1013367 - did the writing go overtime as in spinning on it or as in had more to write so kept on writing? [06:39]
ossabot: Logged on 2019-12-18 03:20:38 whaack: diana_coman: EOD Report: I had to edge into the night a bit, but I finished the schedule: writing exercise (2h), saltmines (8h), proofreading writing exercise (35min), preparing tomorrows writing exercise (15min), setting up postgres on VM (3hr), Spanish study (1h). The writing went overtime because I finished at t=90mins without having done a proofread. I have my VM setup. I can connect my local slime to an sbcl process running on [06:39]
diana_coman: http://ossasepia.com/2020/04/21/ossasepia-logs-for-18-Dec-2019#1013368 - sounds good! [06:39]
ossabot: Logged on 2019-12-18 03:20:38 whaack: the VM and spin off logbot threads. I have a test bot running in #whaacked. [06:39]
dorion_road: http://ossasepia.com/2020/04/21/ossasepia-logs-for-18-Dec-2019#1013360 << you didn't ask me, but here's an idea : as a starting point, why don't you lean of jfw's experience adapting lfs and build and install Gales ? [10:56]
ossabot: Logged on 2019-12-18 00:15:18 lobbes: diana_coman: so I had a brief exchange with trinque and he suggested a linux-from-scratch exercise as a way to get a better grip on the fundamentals of an OS. [10:56]
ericbot: (trinque) 2019-12-18 trinque: LFS will teach you more about what actually constitutes a working system, because you'll make one with your hands [10:56]
dorion_road: spyked, bvt, triqnue and diana_coman all said they have going through the process on their list in the near future. [10:56]
dorion_road: trinque* [10:57]
dorion_road: Doing it yourself and writting a report about it will add leverage to that conversation for everyone as what to adapt from Gales, Cuntoo, else to TMSR OS becomes understood. [10:57]
dorion_road: As for a time estimate, a data point is we guided our clients through the build, install and initial configure in ~4.5 hours (3 ~90 minute sessions). This includes explaining things to them, their beginner command line inefficiency and building on dual core, 4 GB RAM system. [10:58]
diana_coman: dorion_road: would you rather he writes a report on Gales first than on ebuilds? what's exactly the priority/goal/status re ebuilds anyway? I'm getting a bit confused now re this part as I didn't keep it in head to this level of detail. [11:21]
dorion_road: diana_coman ebuilds is most interesting in my view for the automatic dependency resolution and because there's a lot already extant. the cost of terraforming it vs. adapting the Gales build system isn't yet clear. [11:31]
dorion_road: regarding priority, lobbes, how far along are you and how much more time do you estimate on the ebuild report ? when is your deadline ? I know you have other items on your plate, e.g. auctionbot. [11:32]
diana_coman: dorion_road: so what are you after from him as a first report? [11:32]
diana_coman: re ebuilds I mean. [11:32]
diana_coman: dorion_road: for that matter he actually did ask you - did I miss the answer to that? [11:33]
ossabot: Logged on 2019-12-16 20:26:42 lobbes: dorion_road: I also wanted to pick your brain re: my eventual report on the state of the ebuilds. Where I currently stand is I've only ever used Portage; I've never really dug into it. I know it is a beast, so I was hoping that you could help me narrow my focus onto what specifically you'd like me to find out about it [11:33]
dorion_road: diana_coman jesus, I did miss that. [11:35]
diana_coman: good he asked some more then, heh. [11:35]
dorion_road: lobbes focus on the automatic dependency resolution. I don't know how such a mechanism will fit under V / TMSR OS , but that's a big point of leverage that Gales lacks. [11:40]
dorion_road: have you delved into that at all so far ? [11:41]
diana_coman: dorion_road: what's your current vision for this particular part (I guess it would be Package Management from your list)? [11:42]
diana_coman: you are possibly in the best place atm for it anyway since you know Gales' system + you have some idea re Portage as user if not about internals; but from your article there, I don't really get any idea beyond "there's portage and there's gales" and we need something. [11:43]
diana_coman: fwiw I rather doubt it'll be either terraforming portage (I expect it would be a huge task really) OR directly using Gales' system as such; ie what I see missing there is some thought/root of thought into the V-shape of this [11:44]
diana_coman: dorion_road: to ask it from a different angle too: in what way do you think the automatic dependency resolution of ebuilds may carry over directly to tmsr-os given its v-package management? [11:47]
dorion_road: diana_coman the vision is a bit unclear atm and I agree the root of how to integrate with V is the biggest piece missing. I've used both, but my perspective isn't as informed technically and I'm looking forward to what my betters have to say about Gales once they've tried it to help clarify next steps. [11:50]
jfw: See also my thinking aloud and MP's response which was pretty enlightening to me re what V-shaped OS means [11:51]
diana_coman: ha, jfw can't claim the "isn't as informed technically" [11:52]
diana_coman: not to mention that gales is his so... [11:53]
jfw: note that 1) dependency resolution in the abstract isn't rocket science, toposort as seen in V is maybe 15 lines of code; 2) I specifically didn't include it in Gales because didn't want to over-develop something that had to be thrown out anyway once V-OS was clarified [11:53]
jfw: from the above link, the ideal as I understand it is: there is no "package X requires separate package Y, some portage is needed to glue them together from different sources"; only "patch X builds on Y vtree, and system needs to support whatever X you wish to press". Static linking seems quite essential for that actually [11:56]
dorion_road: diana_coman right about jfw , and as far as ownership, I think he would be a good fit. we haven't talked about what he wants to do yet, in part because wallet dev, travel, etc. I also think he'd be a good fit for compiler and C library if ave1 doesn't want it. [11:57]
diana_coman: dorion_road: ah, I wasn't trying to push jfw into doing it all, lol; just meant that you *can* pick his brain to help you get a better vision there esp as he's the authority on Gales, what. [11:59]
whaack: diana_coman: Writing went over time for a reason closer to the ~ 'had more to write so kept on writing.' I finished getting down what I had to say at ~ t=90 minutes, and then still had to proofread. [11:59]
dorion_road: diana_coman I didn't take it that you were pushing, and pick his brain I will :) [12:00]
diana_coman: dorion_road: and note that lobbes is basically saying precisely that he's not all that competent technically either so make up your mind there re what does installing Gales do anyway - you can't simultaneously claim that it didn't do much for you re getting a good technical grasp and at the same time suggest it to lobbes for..getting a good technical grasp, you know? [12:01]
diana_coman: whaack: well, that's the good type of going over time then; hopefully it will get even longer & you'll simply split it in 2 installments/articles if needed. [12:02]
diana_coman: whaack: btw, comment in your mod q. [12:03]
whaack: diana_coman: Ah thank you. I forgot to mention it to Jacob / comment on his - I had read that piece by jfw, and that's what inspired me to write my mine. [12:08]
diana_coman: whaack: tsk; do link refs already, don't do the silly "forgot to link" stuff. [12:10]
diana_coman: esp if it actually inspired you to read yours, just how exactly you ..."forgot"?? [12:10]
dorion_road: diana_coman that's true. for example, it helped me clarify what static linking, cross compiling, etc are. my comment about technical grasp was more about steps forward in V-ifying the package management, rather than strengthening me from where I was. [12:10]
diana_coman: dorion_road: if only you'd wrote about it at that time :P [12:15]
whaack: diana_coman: I had read his article a few days prior and thought that a 9/11 piece would be a good topic for an open ended article. I guess it would be better to say while writing the article i didn't take the action to link / comment at the point when I remembered jfw's. [12:16]
diana_coman: dorion_road: anyways, to coordinate a bit here so poor lobbes doesn't end up entirely confused: from tmsr-os perspective, would a Gales-installation report be more useful atm than a beginner's take on ebuilds investigation? [12:16]
diana_coman: whaack: so take that action each and every time! and yes, better said like that. [12:16]
dorion_road: diana_coman lobbes I think a Gales installation report would be better atm. espcially if there still remains a lot of work cut out for lobbes on the ebuilds. ebuilds can be approached as necessary after. [12:26]
diana_coman: dorion_road: cool then. [12:26]
diana_coman: lobbes: do a write-up of whatever you have already on ebuilds, just so it doesn't go to waste; take Gales and install it + report on it and remember that it's basically an accelerated lfs + you might still get direct help from jfw and/or dorion_road if you get stuck (and ask, ofc). [12:28]
diana_coman: jfw, dorion_road while I haven't really spent time specifically pondering the v-shape of a package management system, I think it essentially stems from that fundamental "each V-tree" is the context of its all vpatches; as such, for one thing there isn't an "automatic dependency resolution" external to the v-tree because what would it do anyway, mess up trees or what exactly? and the problem otherwise switches more to a proper grind ... [12:33]
diana_coman: ... of the tree ie a reader/signer's thing rather than a system; ie possibly even different grinds of "my OS" since anyway, tmsr-os or not, it's still up to each user what exactly they run on their machine. [12:33]
jfw: whaack: glad to be of inspiration and looks like you're off to a strong start on the daily writing! [12:35]
whaack: jfw: ty! it is quite enjoyable [12:35]
jfw: lobbes: feel quite free to ask for help re Gales as needed. [12:35]
whaack: diana_coman: ack, I won't hesitate to link going forward [12:35]
jfw: the LFS handbook might make a good companion reference too, as I recall it went into greater depth on some explanations, though many of the steps involving glibc / dynamic linking aren't needed. [12:37]
diana_coman: ^^^^^^ [12:37]
jfw: (needed by Gales at any rate.) [12:37]
diana_coman: jfw: what's your plan re chan and blog over the holidays? [12:38]
jfw: I plan to follow the log and chime in as needed though it might be sporadic; on the blog, will keep an eye out for comments but not planning to write new articles [12:40]
diana_coman: prolly better off calling it "I'll be offline but I'll try to at least read the logs"; makes for closer-to-likely-reality expectations all around. [12:43]
diana_coman: kind of curious if you'll find it harder to write after the break :D [12:44]
jfw: hmm... your closer-to-likely expectations have been on target so far so I won't dispute. [12:44]
diana_coman: well, I can always bait you with a lot of discussion of your own code and then you'll get online. [12:46]
jfw: likewise curious on the writing, maybe I'll at least keep a journal to stay in something of the habit [12:46]
jfw: could be, heh. [12:47]
jfw: I'd like some help as to adequately approaching the recall and reload / setting down carefully and time-consumingly process [12:50]
ossabot: Logged on 2019-11-17 16:36:29 diana_coman: whaack: that is very possible; the other part to this is to start getting better at helping your recall and reload really; let me fish something out for you. [12:50]
ossabot: Logged on 2019-12-17 04:11:25 diana_coman: To be explicit: there are just 3 options there, namely a. you drop the work quickly and painlessly now to pick it up painfully in January b. you set down the work carefully and time-consumingly now to pick it up easily in January c. you don't do either and therefore take the work (or parts of it) with you on holidays too, whether you sit down to do any of it or just carry it in your mind anyway. [12:50]
jfw: so far I've listed the tasks in flight and planned, and there's the earlier planning articles too [12:51]
diana_coman: is looking for that... question mark. [12:51]
jfw: Can you clarify the steps that you find helpful there? [12:52]
diana_coman: yay, question! jfw, this is a bit of an iterated/person-specific process really but essentially what I found I'm aiming for is to jot down the main bits of context really [12:54]
diana_coman: tasks in flight and planned as you say sound good but also ideas/why [12:54]
diana_coman: because - or at least the way I work - there's a lot of thinking before/in those "tasks" and it's mainly *that* part that is painful to lose/have to pick up from start again (or even remember that uhm, I was thinking of that/this/the other/pondering) [12:55]
diana_coman: so I end up listing all sorts, including constraints (why I wanted to do this and not that) and wtfs [12:57]
jfw: . [12:57]
jfw: whoops. Makes sense [12:57]
diana_coman: or if it helps, maybe look at it from the other end: what would you want to have on one page to read to get you quickly back up to speed when you come back to it? [12:58]
diana_coman: (then again, do note that my eulora work has consisted up to date mainly of figuring out and trimming a ~500k+ clusterfuck, so possibly on the heavy side in terms of tangles.) [13:00]
jfw: aha, seems I can manage mostly by reference to my own things from here [13:02]
diana_coman: cool. [13:02]
jfw: well except those trb patches, heh, but they're still not too invasive. [13:02]
diana_coman: happy you. [13:03]
whaack: diana_coman: For my schedule today (2hr 15 min on writing, 1h Spanish study, 8hr 45 min on logbot / irc exploration project) I prefer to go ~2 hours short for logbot / irc so I can get to bed earlier. Last night I had to pick between going against my word or staying up late, and I chose the later. I would not like my schedule to shift into the night. [13:16]
diana_coman: whaack: makes sense. [13:17]
whaack: CR is not a place for night owls, it is much nicer to wake up at ~5am here lest you cook in your bed in the morning [13:19]
dorion_road: http://ossasepia.com/2020/04/21/ossasepia-logs-for-18-Dec-2019#1013429 << the matter needs to be considered further and i read your reasoning to be strong. in each deployment the user subjectively selects the vpatches and sigs for a particular machine and his plans for it. [15:16]
ossabot: Logged on 2019-12-18 12:33:03 diana_coman: jfw, dorion_road while I haven't really spent time specifically pondering the v-shape of a package management system, I think it essentially stems from that fundamental "each V-tree" is the context of its all vpatches; as such, for one thing there isn't an "automatic dependency resolution" external to the v-tree because what would it do anyway, mess up trees or what exactly? and the problem otherwise switches more to a proper grind ... [15:16]
diana_coman: it's just my initial thought on it really, so I expect there will be more discussion on it esp as things get moving otherwise, certainly; the main thing though is that in this/similar, I fail to see much use/space for ebuilds/portage/very-similar. [15:21]
diana_coman: whaack: re ~5am in CR, that's pretty much the hour I was up each and every day of those 2 weeks I was there (+ even in the streets it was easy to notice that everyone had the schedule essentially earlier than the usual in Europe); the light was enough for the wake-up-at-5. [15:26]
diana_coman: that being said, it was rather... exhausting overall. [15:27]
whaack: diana_coman: I imagine it was exhausting from the heat + the walking/traveling you did, not from waking up at 5am since ostensibly you also went to bed earlier [15:33]
whaack: diana_coman: Going to bed before 10pm consistently was a major shift for me when I got here, but it is the only schedule that makes sense. Given the heat the two best times to be outside are sunrise and sunset. A walk at night in CR is awful, especially in non urban areas where there is no light and instead just a bunch of bugs out to eat you. [15:33]
diana_coman: whaack: the "went to bed earlier" didn't quite...work out, no, lol. [15:36]
diana_coman: but there was also some jet lag and all the rest, indeed. [15:38]
whaack: diana_coman: aha well at least there was good coffee here to aid you. (iirc you mentioned in your smilely alligator article that you drank some even though you do not drink coffee normally) [15:39]
diana_coman: yes; and the rum to not-help. [15:40]
diana_coman: heh, not complaining though, no. [15:41]
whaack: lol [16:18]
whaack: well i have found a quite weird bug in my mpwp. I have 5 superscripts linking to footnotes in the body of my article, as I should, but only 4 footnotes appear down below. [16:19]
whaack: figured out problem. It was caused by a double quote, ", in an anchor tag being accidentally written as a single quote, ' [16:23]
jfw: diana_coman: in searching for test data for the full keccak sponge I've found "known-answer test result" zip files at https://keccak.team/archives.html ; despite the "warning, obsolete!11" these appear to come from various iterations of the NIST tweaking and I've not been able to reproduce at least by a naive interpretation of the test format. [16:57]
jfw: Do you have such test data for the original keccak? Otherwise, would you prefer I re-genesis now with fixed self-gen test, or postpone and try to figure this out? [17:00]
jfw: I see one sponge test at http://ossasepia.com/2018/01/25/eucrypt-chapter-7-keccak-sponge/ but seemingly for a 5-bit input whereas my prog demands bytes. [17:06]
asciilifeform: diana_coman: i wrote a comment re your vtronics piece (sadly much belated, was in uruguay) , is in your spam queue atm [17:06]
whaack: diana_coman: My next step is to spin up a new VM and test run my instructions by reinstalling everything. There were a few "do, undo" steps I took and I want to make sure that (1) I have a pure environment without files I forgot to uninstall (2) I/anyone else can quickly rebuild using the instructions [18:12]
lobbes: http://logs.ericbenevides.com/log/ossasepia/2019-12-18#1013441 << sounds good to me, I will aim to have write-up of what I have already on ebuilds done by the Sunday. After that I will do a Gales install and report. [20:52]
ericbot: Logged on 2019-12-18 16:23:58 diana_coman: lobbes: do a write-up of whatever you have already on ebuilds, just so it doesn't go to waste; take Gales and install it + report on it and remember that it's basically an accelerated lfs + you might still get direct help from jfw and/or dorion_road if you get stuck (and ask, ofc). [20:52]
lobbes: ty diana_coman, jfw, and dorion_road for the feedback, and I'll definitely reach out for help during install. [20:52]
trinque: diana_coman: since jfw and dorion_road appear to be in your house, where would you like the OS thread carried? [22:54]
trinque: here? #trilema ? my chan? [22:54]

Comments feed: RSS 2.0

Leave a Reply