rinigus | do we have some macro defined on SFOS OBS (or even better on all SFOS SDKs) that can be used to distinguish in SPEC that the build is for SFOS? so I could use the same SPEC for SFOS and suse/fedora? | 07:49 |
---|---|---|
T42 | <adampigg> What u trying to do rinigus? | 08:33 |
rinigus | @adampigg: learning how to use opensuse obs. so, as a first step, packaging maps apps for distros via obs | 08:34 |
T42 | <adampigg> K | 08:35 |
rinigus | not sure it is all correct, but current test at https://build.opensuse.org/package/show/home:rinigus:maps/libpostal | 08:35 |
rinigus | notice differences in _service | 08:35 |
T42 | <adampigg> How easy is it to add the sfos targets as additional distros? | 08:36 |
rinigus | @adampigg: to suse obs? they don't want anything closed source | 08:39 |
rinigus | https://en.opensuse.org/openSUSE:Build_Service_application_blacklist | 08:40 |
T42 | <adampigg> I wondered if you had rolled your own instance | 08:40 |
rinigus | @adampigg: no, I didn't :) | 08:40 |
rinigus | being lazy over here. we will be told how bad it is and then can take it from there | 08:41 |
T42 | <adampigg> Yep..friday soon! | 08:41 |
T42 | <adampigg> Have a cheery holiday snap | 08:42 |
rinigus | as for using suse obs, it could be good to push apps for other distros. and probably for nemomobile | 08:42 |
T42 | <adampigg> (Photo, 2560x1440) https://irc.thaodan.de/.imgstore/M3PjNJ0kcl.png | 08:42 |
rinigus | @adampigg: and all is great again! | 08:42 |
T42 | <neochapay> yea...i searching hosting for nemo obs | 08:43 |
rinigus | @neochapay: let's see, suse obs could be the one for nemo. I will test it on maps stack, report back, and then will look on nemo packages | 08:46 |
T42 | <edp_17> @adampigg: is that the new boat for sfos or how you feel about the current? 😄 | 08:49 |
rinigus | :) | 08:51 |
r0kk3rz | suse obs wouldnt work | 10:15 |
r0kk3rz | or at least wouldnt use sb2, which might cause problems | 10:15 |
rinigus | r0kk3rz: that is for sfos or nemo? | 10:51 |
rinigus | they do have arm 32/64 bits available | 10:51 |
rinigus | @adampigg: https://forum.sailfishos.org/t/obs-shut-down-and-next-steps/1814 | 10:59 |
atlochowski | rinigus: adampigg: https://forum.sailfishos.org/t/obs-shut-down-and-next-steps/1814 | 11:00 |
rinigus | r0kk3rz: very well said on obs thread, thanks! | 11:52 |
T42 | ç™½æ˜Šæƒ %lastname% was added by: ç™½æ˜Šæƒ %lastname% | 11:56 |
r0kk3rz | rinigus: yeah no need for long tirades | 12:10 |
T42 | <adampigg> Rinigis, r0kk3rz up to date... Not terribly informitive as to what solutions are proposed | 12:10 |
r0kk3rz | yeah i dont think there is any | 12:11 |
r0kk3rz | they'll figure something for ports maybe, but not anything else | 12:11 |
T42 | <adampigg> He says obs is no longer needed for porting with new hadk...but where are we to host ota updates etc | 12:12 |
T42 | <adampigg> The 'legal risk' argument sounds a bit wishy washy, like thet havnt actually identified anything conclusive | 12:22 |
T42 | <edp_17> From how I read in the latest and previous announcements, it is clear to me that Jolla feels big enough now to spit on the community. It doesn't matter what we say or do. We can raise our voice, it doesn't matter anymore because the decision has been made. It's clear. Leave it or obey it but I don't think this can (will) be changed. | 12:32 |
T42 | <edp_17> I was happy when found SFOS and wanted to port it to all of my devices but now, ... I am sad. | 12:32 |
T42 | <adampigg> Im not ready to rage quit just yet, like rinigus i want to see solutions first, ive invested many years in porting, but seems jolla management arnt listening to their own devs, or properly discussing with community | 12:39 |
rinigus | @adampigg: not informative in terms of solutions, indeed. and, as r0kk3rz suggested, probably no good solution expected | 12:40 |
rinigus | don't think we have to rage quit, but instead work on exit strategy | 12:40 |
rinigus | r0kk3rz: with suse obs, do you think it is a problem with compiling nemo on top of some suse arm distro? due to absence of sb2? | 12:41 |
mal | has OBS been recently use much for other things than adaptation builds/hosting? apps are in store or openrepos | 12:57 |
vknecht | rinigus: there's this nemo-on-postmarketOS, so I'd guess it's doable on Suse too... http://nemomobile.net/pages/Packaging_Glacier_for_postmarketOS/ | 13:03 |
T42 | <adampigg> I think.we.need the sentiments of this channel reflected on the thread | 13:05 |
T42 | <adampigg> Mal, i dont think apps are the issue, there is the sdk for that with hosting on store / openrepos. .... | 13:06 |
T42 | <adampigg> Ports are the issue, and tbe ease obs gives us with updating packages, auto building and hosting | 13:07 |
mal | btw, I already have the self-hosted repos in test use on a device | 13:08 |
mal | still on my own server | 13:08 |
T42 | <NotKit> rinigus: I experimented with suse OBS a bit on the weekends and at least it can build libhybris just fine for arm/arm64 with small changes, so I think missing sb2 is not an issue | 13:13 |
rinigus | mal: maps apps (pure maps and osm scout server) are on a project with 17 packages. once in a while I update one or another dependency. those are also recompiled on sfos update | 13:14 |
rinigus | flatpak is less - 7 | 13:15 |
rinigus | so, larger apps do require some automatic building | 13:15 |
mal | ah, that is indeed a bit annoying locally unless scripted | 13:16 |
rinigus | @NotKit: interesting. I will have to figure out how to feed it best to their obs (_service) and then would look into nemo. do you have some repo with hybris over there? | 13:17 |
mal | but with a bit of script magic that is not so difficult to do, like making a script with a simple dependency tree and if you trigger it with some package then it will build that and all packages which depend on it | 13:18 |
mal | I do understand making such script is annoying compared to OBS | 13:19 |
rinigus | mal: add to that ability to support multiple sfos releases - as I did before but probably not anymore after gcc update - and there was my use of obs | 13:19 |
mal | yeah | 13:19 |
T42 | <NotKit> @rinigus [@NotKit: interesting. I will have to figure ou …], https://build.opensuse.org/package/show/home:NotKit:hybris/libhybris | 13:20 |
rinigus | @adampigg: I will probably write something tonight in the thread, will have to think it through. please feel free to voice your opinion | 13:22 |
rinigus | mal: yes, annoying and time consuming. instead of working on something else you are taken away the means that you used and suddenly have to start looking / creating replacements. | 13:24 |
rinigus | in principle, we can probably setup local obs instances to do it old(er) way. assuming that some simple runme.sh could be done that sets it up on my pc, for example | 13:25 |
rinigus | and we have reasonable ram consumption. don't know if it is possible | 13:25 |
rinigus | mal: but the issue with hosting is there as well. so, somehow we have to figure out how to replace whole chain code->compile->host | 13:26 |
mal | yeah, I think we should do some scripts for some things | 13:26 |
mal | in general we should have more scripts commonly done things | 13:28 |
mal | +for | 13:28 |
rinigus | taking into account recent developments, I will be volunteering much for doing it :( . which is a nasty divide now that I was not experiencing before. and not constructive either, as I would prefer to focus on things that cannot be pulled out by jolla management/lawyer | 13:29 |
rinigus | @NotKit: nice, will look into it. could see that you repacked it and worked through tar in service. I tried to go directly in https://build.opensuse.org/package/view_file/home:rinigus:maps/libpostal/_service?expand=1 which required different %setup | 13:32 |
T42 | <adampigg> Mal do u think porters will now be expected to self host? | 13:41 |
mal | @adampigg best would be to have some common hosting method for all ports | 13:46 |
mal | something like the suggested openrepos or something like that | 13:47 |
Nico[m] | Does anyone know, how much effort it would be to host something like that? Be it obs or some alternative? | 13:47 |
mal | only thing needed is some server with public address and a folder where you can upload the local repo | 13:48 |
mal | if you mean hosting adaptation | 13:48 |
Nico[m] | Well, I do have that. But I guess it would be more involved to provide that for others and something based on openrepos may make more sense then? | 13:49 |
T42 | <kabouik> mal: I'm not much of a developer but I did use OBS whenever I could write a proper .spec file for building packages and keeping them up to date easily, or rebuild with new deps; it was a much better workflow than building on device or on SDK for me. | 14:07 |
rinigus | I don't know whether openrepos has any api for uploads, never seen any. if we split hosting and building, some api would be needed (minio, scp) | 14:11 |
T42 | <vkn_n> OBS takes about ~166Gb. I downloaded the OBS backup recently. | 14:15 |
T42 | <adampigg> The fact we're having this discussion shows it hasnt been thought about | 14:24 |
mal | @adampigg which part? | 14:26 |
T42 | <adampigg> Hosting | 14:26 |
mal | changing to selfhosting repo was quite simple when I tried it | 14:27 |
T42 | <adampigg> On an individual basis maybe, but not as a common platform for all ports as slick as obs | 14:29 |
DylanVanAssche | rinigus (IRC): adampigg: I get that they want to move forward and improve the development setup but there wouldn't be so much problems if they just announced it upfront like: 'We're considering X, Y, Z which impacts the community by A, B, C for the following reasons: ... . How do you think as a community to move forward'. This would make clear that Jolla understands the consequences and so on. I | 14:29 |
DylanVanAssche | hope they learn a bit out of this in terms of communication. The current follow up post just reads like damage control IMHO. | 14:29 |
T42 | <vkn_n> Do we know exactly when OBS will be closed? | 14:31 |
T42 | <adampigg> Nope | 14:31 |
T42 | <adampigg> Probs friday :D | 14:32 |
DylanVanAssche | hahah | 14:32 |
T42 | <vkn_n> (( | 14:32 |
T42 | <vkn_n> I can allocate hard disk space on my personal server to transfer information. If you need to - write. We need to decide something... | 14:33 |
T42 | <adampigg> Atleast on the.forum it sounds like it wont go imediately | 14:34 |
mal | it's not going away anytime soon, there will be plenty of time for adaptations to adapt | 14:34 |
mal | because adaptations need to have at least one release before it goes down to make transition simple | 14:34 |
T42 | <adampigg> Maybe jolla should just provide server space, at ports.jolla.com, with some api to alllow uploads from a ci build | 14:37 |
T42 | <adampigg> Mal, what about.common and native-common repos? | 14:40 |
mal | that needs some thinking | 14:44 |
mal | @adampigg local build does build all dependencies by default | 14:44 |
T42 | <adampigg> Mal, obs also has the convenience of just goibg.down each package and triggering an update to get the latest updates, and them being available instantly | 14:48 |
pketo | I think providing server space would also have the legal issues regarding the content that is uploaded there | 14:56 |
pketo | I don't know the details about the legal issues, so don't ask :) | 14:57 |
rinigus | mal: I am not even sure whether local build would work for my current setup, for example. as I can push all bits to obs and work from there. as a result, we don't need to maintain many local copies of the same / similar thing | 14:59 |
rinigus | pketo: I would expect hosting is the part that has legal implications | 15:00 |
mal | good example of hosting legal issues for ports would be vendor blobs | 15:02 |
rinigus | DylanVanAssche: I would not count that those responsible for decisions and communication would learn much from it, unfortunately | 15:03 |
rinigus | mal: exactly that's why sony hosts them separately. do we have anyone pushing blobs to obs? | 15:03 |
DylanVanAssche | rinigus (IRC): :( | 15:04 |
rinigus | as piggz mentioned before, it is a question of pasting what can be done and cannot be done publicly. aka education | 15:04 |
rinigus | to my knowledge, blobs were always no-no | 15:05 |
T42 | <adampigg> @rinigus [as piggz mentioned before, it is a question of …], Yes, put clear guidelines on the front page, and notify offenders if found....but...has any actually been found? | 15:19 |
T42 | <adampigg> Pketo^^ are there any.actual.cases of infringing packages? | 15:20 |
pketo | I don't know | 15:27 |
T42 | <adampigg> Can we call it obs-gate from now on? :D | 17:55 |
DylanVanAssche | lol | 17:56 |
attah | How useful or worthless would it be to contribute a Xeon E3-1246v3 machine with 32GB ram to "the cause"? | 18:05 |
T42 | <adampigg> I think its too early to think about things loike that. | 18:11 |
T42 | <adampigg> Need to know the proposed solitions | 18:11 |
attah | indeed, that will be interesting... although it did sound a bit like "it will be easier to run locally" | 18:12 |
T42 | <adampigg> Maybe we neeed to start drafting requirements on the post? | 18:25 |
T42 | <adampigg> I can conteibute tomorrow | 18:25 |
attah | As an outsider it is all bit vague exactly what capabilities are hardest to replace | 18:26 |
attah | Repo hosting seemingly being one of the easier, and the most problematic to keep | 18:26 |
T42 | <adampigg> Adaptation package building, hosting, updating and auto-rebuilding when a dependancy changes | 18:27 |
T42 | <adampigg> Is most of it for ports | 18:27 |
T42 | <adampigg> There is other stuff for complex apps | 18:28 |
rinigus | @adampigg: posted my reply, maybe should have waited. but it was not friday-type of reply either | 18:28 |
attah | What's the key reason for not doing it locally, apart from not really being suitable for automation? | 18:28 |
rinigus | doing locally means you have to have each developer spending time on admin task as well | 18:29 |
rinigus | having setups diverging. as for automation, don't underestimate it for ports and larger apps | 18:30 |
rinigus | network cost could be significant, btw. for service you have to pay each mb going out | 18:30 |
attah | to me the risk of setups diverging or not being equal enough is kinda orthogonal to a central service being able to work with a minimum of maintenance... | 18:32 |
attah | but i haven't tried, so... | 18:33 |
attah | yeah, hosting is a pain | 18:33 |
rinigus | attah: hosting is not a pain as such. it's a cost. unless you mean from legal pov | 18:34 |
attah | both legality and cost | 18:36 |
T42 | <adampigg> Porters shouldnt be expected to pay for hosting | 18:37 |
attah | especially porters *alone* | 18:38 |
attah | the wider community however might, i guess | 18:39 |
attah | but as per before, that's a hassle to coordinate | 18:40 |
T42 | <adampigg> But only.as a single.collective,.imagine the pain in admin otherwise | 18:40 |
rinigus | maybe someone out of users should organize it then. but legality has to be ensured somehow | 18:41 |
rinigus | as through review, for example | 18:41 |
attah | yeah, clear responsibilities, admins and so on... decentralizing would be a whole other headache | 18:42 |
attah | although basically darkwebbing the repos would fix the legality part xD | 18:43 |
rinigus | or hosting somewhere with more relaxed view on it | 18:43 |
rinigus | but it is easier to keep all legal anyway | 18:44 |
attah | because unless people engage in active abuse, there is little chance anybody will go after a firmware blob or two | 18:44 |
rinigus | if we keep it blob free, there will be no issues either | 18:44 |
attah | hopefully, but IP is a sticky mess | 18:45 |
rinigus | then host sony only. and pinephone and others who keep blobs on their sites | 18:46 |
vknecht | or, change laws, so that individuals and community can't be bothered with business rules, especially when devices aren't supported anymore so why the f*ck should we be sued for distributing dead files ? | 22:34 |
vknecht | it's we, the good people, who should have power and sue them for letting their security-bug-riddled devices rot and put consumers at risk :-P | 22:37 |
Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!