Monday, 2018-11-26

*** feodoran_ is now known as feodoran00:27
tortoisedoclbt :07:06
tortoisedoclbt: obs does not allow to build against older versions of a same package?07:06
tortoisedoc(for example rust-1.26.0 depending on 1.25.0, but if i have 1.26.0 built already, 1.26.0 will be "unresolvable")07:07
tortoisedoc(the same applies for any package)07:07
CoolgeekHey, I have a crash/reboot yesterday while transfering via MTP a big file (phone->PC). Where can I check the log (if any) ?09:04
CoolgeekI transfered this file OK with developer mode and scp09:04
*** Claudybuns is now known as Claudy09:25
tortoisedocach is there an update being built on obs :S11:06
r0kk3rztortoisedoc: you cant have multiple versions in the same project no11:08
r0kk3rzwe've been through this before :P11:08
r0kk3rzthats why the older rusts are named rust-126 or whatever11:08
tortoisedocr0kk3rz : I don't understand. What do you mean by "multiple versions in the same project"? a "versions" as an rpm as such is a deliverable artifact, and should affect the building of the project in the way that if its required, and not there, then it doesnt build.11:09
r0kk3rzso if you have a project home:tdoc:rust, with many rust builds11:10
r0kk3rzobs, as it currently works, cant handle that11:11
tortoisedocalso the way rust is named on obs is wrong, in that it crosses the streams (obs package name like rust026 being referenced in spec, which is rpm)11:11
r0kk3rzperhaps newer versions of obs can handle that better11:11
tortoisedocr0kk3rz : my bet is its a bug of obs ;)11:11
tortoisedoc(on obs == on mer obs)11:11
r0kk3rzyeah11:11
tortoisedoclbt can probably confirm this11:12
r0kk3rzim not sure its a 'bug' its more a consequence of how it does dependency resolution11:12
tortoisedocr0kk3rz : I dont think its a dependency problem, as in my spec files i ask for a specific version, so it should be easy for obs no?11:13
r0kk3rzit is a dependency problem :P because you can specify it in the spec and then it says later 'nah i cant find that one'11:13
tortoisedocits not :), in that the dependency is there, and specifying a given version only works for the most recent version (and not the previous ones)11:14
r0kk3rzwe're arguing semantics, anyway11:14
tortoisedocI have to disagree on this one :)11:14
tortoisedoc(as well)11:14
r0kk3rzthats why i said 'dependency *resolution*'11:15
tortoisedocwe are arguing about obs not doing it's job :PL11:15
r0kk3rzwell my dear sir i must retort11:15
r0kk3rzwe're arguing about how terrible rust is11:16
tortoisedocI think we can all agree that rust is a rolling turd :P11:16
r0kk3rz:D11:16
tortoisedocit's impressive the effort they have put in to make sure people use only the latest & greatest11:16
r0kk3rzanyway. from my experience thats just how MerOBS acts, whether correct or not11:17
tortoisedocwith total disregard of older systems and backportability11:17
tortoisedocobswizard lbt could shed some light with one of his spells..11:17
r0kk3rzthere is an update being worked on, and maybe that takes advantage of newer libsolv magic11:18
tortoisedoclibsolv?11:19
r0kk3rzyeah11:22
r0kk3rzthe library used for dependency resolution in zypper and such11:23
jmlichhi, I have some trouble with sailfish os web browser. It crashes when loading 'www.root.cz', can anyone reproduce this?13:45
jmlichI am using Jolla 1.13:45
fledermausstarts loading, stalls at about 70%13:53
fledermausyeah, deadsies13:54
fledermauslocked up, black screen, red flashing light, the works14:10
r0kk3rzoh it crashes lipstick, nice14:11
tortoisedochmm root.cz what website is that?14:18
fledermausseems to be some free software news/magazine site14:20
tortoisedocfull of trackers certainly :P (kidding)14:20
fledermausprobably is, lots of sites are.14:21
fledermaustortoisedoc: take a look at the second cook declaration here: https://mastodon.social/@ersatzmaus/10103728712122040314:23
tortoisedoc:D14:24
tortoisedocconsidering all of them  is most likely a tcp connection, no wonder the browser jams14:25
fledermausoh, that wasn't this site.14:34
fledermausI dunno what this site does.14:34
leszekjmlich: are you the one that created the tjc entry? Please use the search. There is a complete thread for sites that crash the browser and everything :P15:04
jmlichoh, i am sorry. So next time15:04
leszeki closed the thread for now. Please add it to the linked one in the comment if it isn't there yet15:05
oltortoisedoc (IRC), r0kk3rz (IRC): OBS explicitly excludes previous binary packages build from the same OBS package from dependency resolution to build new binary packages from OBS package. This is not a bug, this is an important feature. OBS is about reproducible builds, so build result can depend on other packages, but never on whatever was built previously from this package.15:20
r0kk3rzol: in this case, there are seperate packages in the same project for previous versions15:22
olImagine that you've just created a new project and populated it with packages with exactly the same sources. The result should be the same as if you came to this project's state through multiple package revisions.15:22
r0kk3rzbecause rust_n needs to be built by rust_n-115:23
olThen you need N packages with different names for that.15:23
r0kk3rzyeah it was handled by changing the name in the spec, but i think tortoisedoc wants it to work through versioning15:24
olNot in OBS. This breaks main OBS purpose: reproducible builds.15:24
r0kk3rzi figured it was working as it was supposed to15:24
olIn case of package bootstrap from the same sources (like you first build minimal package from the same source and then build a full version using different build flags) you can use multibuild feature introduced in OBS 2.8.15:26
r0kk3rzyeah thats the issue, this isnt the same source either...15:27
r0kk3rzrust seems to have decided to do things differently15:27
olThen N packages with different names producing binary packages with different names as well.15:28
pketo_https://build.opensuse.org/package/view_file/devel:languages:rust/rust/rust.spec?expand=115:30
pketo_that seems to bootstrap it with the prebuilt binaries from rust-lang.org, if I understood correctly15:31
r0kk3rznot a bad idea really15:31
r0kk3rzit beats trying to build everything from v1 onwards15:32
olpketo_ (IRC): This is cheating!15:32
r0kk3rzalso practical :)15:32
pketo_:)15:32
pketo_I kind of assumed that you have looked into how others do it :)15:32
olPrevious version is not built correctly, just binaries are used. This defeats the whole purpose of reproducible builds. This way you can build current version as well, by downloading binaries and packaging them.15:33
pketo_well, it uses specific prebuilt previous version, so isn't that reproducible?15:48
olRepackaging downloaded binaries is also reproducible. Why bother building it at all?15:57
pketo_to build it with different options?16:09
olThat was a rethorical question. :-(16:58
*** Colgate_ is now known as Colgate17:17
deicide-wheres the best place to read about sailfish os? u know guide and shit17:20
leszeksailfishos.org , together.jolla.com and youtube "SailfishOS Podcast" & "SailfishOS App Podcast"17:29
depscriberelated: someone asked me for the d/l link for the community sailfish repository (the non-paid-for, no-android version) i had no idea what to say. is there a grand repository?17:32
deicide-https://jolla.com/sailfishx-linux-instructions/ this one right?17:40
deicide-do i need to know anything before doing it?17:40
leszekdeicide-: you will loose the drm keys forever when unlocking the bootloader17:57
r0kk3rzthe sony page should make that clear18:01
leszekquit18:07
deicide-drm keys?18:12
deicide-@les18:17
deicide-fuck im too used to discrod18:17
aalmiSabre users hands up, need help18:39
deicide-never mind19:04
deicide-fuck my life, phone is defective19:04

Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!