January 19, 2018 
< | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13 | 14 | 15 | 16 | 17 | 18 | 19 | 20 | 21 | 22 | 23 | 24 | 25 | 26 | 27 | 28 | 29 | 30 | 31 | >
January 19, 2018 [00:00:30] <oibot> Project oi-userland build #3725: SUCCESS in 19 min: https://hipster.openindiana.org/jenkins/job/oi-userland/3725/[00:19:12] *** Toasterson <Toasterson!~toast@2a02:aa10:e280:b880:e3d5:f722:2d0a:1d1a> has quit IRC (Ping timeout: 256 seconds)[00:36:08] *** jimklimov <jimklimov!~jimklimov@31.7.243.238> has quit IRC (Ping timeout: 255 seconds)[00:41:32] *** leoric <leoric!~leoric@188.232.89.90> has joined #oi-dev[00:42:19] <leoric> richlowe: it seems I have change to do 'RESOLVE_DEPS = REQUIRED_PACKAGES + COMPONENT_FMRIS' automatically[00:58:18] *** tsoome <tsoome!~tsoome@55-114-191-90.dyn.estpak.ee> has quit IRC (Ping timeout: 256 seconds)[01:07:54] <leoric> richlowe: what do you think about https://github.com/OpenIndiana/oi-userland/pull/3883/commits/74d889861090ab3c9fba53645b628145cf7a22c6 ?[01:10:56] <xenol> leoric: I think we can populate things over the weekend[01:11:41] <richlowe> leoric: don't know if I understand right now, give me a minute[01:11:41] <leoric> we shouldn't do it before gcc-6 migration[01:12:11] <xenol> what[01:12:15] <xenol> required packages?[01:12:18] <leoric> xenol: yes[01:12:22] <xenol> why?[01:12:48] <leoric> otherwise we'll have to do mass s/gcc-4/gcc-6/ immediately after this[01:13:01] <xenol> oh[01:13:06] <leoric> and this will make alarcher's life harder[01:13:11] <richlowe> leoric: the comment about the 'print' in the transform is wrong, right?[01:13:15] <xenol> but I have an idea how to get RED_PACKAGES for all the packages[01:13:20] <leoric> richlowe: yes, fixed it[01:13:25] <xenol> just need to do a lot of stuff on build server[01:13:27] <richlowe> k, I'm less confused now, I think that makes sense?[01:13:42] <leoric> I suppose :)[01:14:00] <richlowe> leoric: another thing I had trouble working out, was whether you had merged the automatic REQEUIRED_PACKAGES stuff in shared-rules fully, in a way that made sense for OI, or not.[01:15:15] <richlowe> that is, when done "properly" shared-macros.mk sets things up for PERL_VERSIONS PYTHON_VERSIONS and friends.[01:15:35] <richlowe> but I wasn't convinced I could get that right for hipster[01:15:43] <richlowe> it should be like, 3 lines of Makefile for someone who knows the answers. :)[01:15:56] <xenol> I ported just bits that included REDUIRED_PACKAGES[01:16:18] <xenol> I don't remmeber if there was some logic to populate variables you mentioned[01:17:00] <richlowe> it's basically a bunch of "if this version is in FOO_VERSIONS, add <this> to REQUIRED_PACKAGES[01:17:08] <richlowe> ifneq ($(findstring 5.10, $(PERL_VERSIONS)),)[01:17:08] <richlowe> REQUIRED_PACKAGES += runtime/perl-510[01:17:08] <richlowe> endif[01:17:11] <richlowe> etc, etc, etc.[01:17:22] <richlowe> with the joyous double-negative that gmake brings along :([01:17:31] <xenol> do you have a link to solaris-userland lines?[01:17:38] <richlowe> No, but I can find one[01:18:42] <richlowe> https://github.com/oracle/solaris-userland/blob/master/make-rules/setup.py.mk#L271[01:18:51] <richlowe> (I was wrong, they do it in the sub-makefiles I guess, not shared-macros)[01:19:24] <richlowe> if you grep REQUIRED_PACKAGES in make-rules there, you'll see a bunch of places where they flesh out some defaults a bit smarter than what you guys merged.[01:20:26] <leoric> well, we evidently miss them :) But we have REQUIRED_PACKAGES += metapackages/build-essential[01:20:52] <leoric> this helps with building packages, but not with resolving dependencies....[01:20:54] <xenol> I can add them tomorrow/saturday[01:22:16] <leoric> Sorry, guys, I was thrilled enough with this change to spend some time on it, but I still have to sleep a bit more than 2 hrs :_[01:22:56] <alanc> such luxury![01:23:08] <richlowe> awww, but you're gonna miss my super fun question![01:23:28] <richlowe> I was trying to find a way that didn't suck to fix wdiff to not conflict with the damn ON wdiff.[01:23:32] <richlowe> tempted to just rename ONs[01:23:39] <richlowe> because I bet I'm the only one who knows it's there, and does stuff[01:24:42] <richlowe> The results of running wdiff and getting wdiff have long ceased surprising me[01:24:45] <richlowe> I go straight to angry, now :)[01:25:54] <richlowe> alanc: you're not gonna give leoric the link to your stuff to make him have to sleep even less? :)[01:26:16] <alanc> which one? the check_rtime in Userland?[01:26:48] <alanc> oh, the sleep deprivation aid - clearly i'm too fried to understand[01:26:53] <richlowe> the one after.[01:27:21] <leoric> richlowe: have bad answer - move gnu wdiff to /usr/gnu/bin (and perhaps, mediate /usr/bin/wdiff)[01:27:26] <alanc> I posted the X-in-Userland link the other day[01:27:52] <leoric> (and we have to deliver gwdiff to make a bit more mess )[01:28:12] <richlowe> alanc: also, removing nethack is a travesty, and y'all should be ashamed.[01:28:38] <alanc> no one could ever replace danek though[01:28:50] <alanc> so how could we ever find a new maintainter?[01:29:15] <richlowe> oh, it's a pain in the ass to package[01:29:22] <richlowe> that's why I'm so annoyed I can just piggy-pack you anymore :)[01:29:29] <richlowe> piggy-back, even[01:30:20] <xenol> time for me to go the bed as well[01:30:22] <alanc> okay, piggy-pack may have surpassed maintainter as most amusing typo in this conversation[01:31:15] <alanc> though danek was always good about sticking close to upstream and not tainting our packages with dozens of patches to resync each update[01:31:54] <richlowe> except where that was the one true path.[01:32:03] <richlowe> isn't that just How Things Are with vim?[01:32:08] <richlowe> patch releases being _patch_ releases[02:16:34] *** v_a_b <v_a_b!~volker@p57A27E8F.dip0.t-ipconnect.de> has joined #oi-dev[02:19:05] *** v_a_b1 <v_a_b1!~volker@p57A27B73.dip0.t-ipconnect.de> has quit IRC (Ping timeout: 256 seconds)[02:39:06] *** alanc <alanc!~alanc@hqdc-proxy-mwg08-o.oracle.com> has quit IRC (Remote host closed the connection)[02:39:23] *** alanc <alanc!~alanc@hqdc-proxy-mwg08-o.oracle.com> has joined #oi-dev[02:39:23] *** ChanServ sets mode: +o alanc[02:58:23] <oibot> Project illumos-gate build #5115: SUCCESS in 1 hr 52 min: https://hipster.openindiana.org/jenkins/job/illumos-gate/5115/[03:12:29] <richlowe> leoric, xenol: https://gist.github.com/richlowe/df865c4bf286de91b60bad13bd97786c[03:12:37] <richlowe> alanc: ^^ you too, but just for sickness.[03:16:28] <richlowe> now I'm going to go to sleep, and pretend that never happened.[07:15:56] *** tsoome <tsoome!~tsoome@13-113-191-90.dyn.estpak.ee> has joined #oi-dev[07:18:43] <leoric> richlowe: assumption that COMPONENT_FMRI is the final one is wrong... We have something like https://github.com/OpenIndiana/oi-userland/blob/284b9129fdc8c867f71ca7ad23106cb0aa7d0470/components/sysutils/nut/nut-libs.p5m#L16[07:19:22] <leoric> we shouldn't rely on one set in Makefile[07:20:33] <leoric> otherwise looks useful[07:21:04] <leoric> The only thing is that pkg contents -rm $COMPONENT_FMRI for each package will take lots of time[07:53:59] *** tsoome_ <tsoome_!~tsoome@177-110-157-37.dyn.estpak.ee> has joined #oi-dev[07:54:18] *** tsoome <tsoome!~tsoome@13-113-191-90.dyn.estpak.ee> has quit IRC (Read error: Connection reset by peer)[07:55:54] *** tsoome <tsoome!~tsoome@13-113-191-90.dyn.estpak.ee> has joined #oi-dev[07:56:41] *** tsoome_ <tsoome_!~tsoome@177-110-157-37.dyn.estpak.ee> has quit IRC (Read error: Connection reset by peer)[08:00:05] *** tsoome <tsoome!~tsoome@13-113-191-90.dyn.estpak.ee> has quit IRC (Ping timeout: 240 seconds)[08:05:41] *** tsoome <tsoome!~tsoome@177-110-157-37.dyn.estpak.ee> has joined #oi-dev[08:16:50] *** tsoome <tsoome!~tsoome@177-110-157-37.dyn.estpak.ee> has quit IRC (Ping timeout: 276 seconds)[08:17:36] *** leoric <leoric!~leoric@188.232.89.90> has quit IRC (Remote host closed the connection)[08:17:55] *** tsoome <tsoome!~tsoome@177-110-157-37.dyn.estpak.ee> has joined #oi-dev[08:22:47] *** tsoome_ <tsoome_!~tsoome@13-113-191-90.dyn.estpak.ee> has joined #oi-dev[08:24:05] *** tsoome <tsoome!~tsoome@177-110-157-37.dyn.estpak.ee> has quit IRC (Ping timeout: 240 seconds)[08:24:05] *** tsoome_ is now known as tsoome[08:28:04] *** tsoome <tsoome!~tsoome@13-113-191-90.dyn.estpak.ee> has quit IRC (Quit: tsoome)[08:32:15] *** mnowak_ <mnowak_!~mnowak_@94.142.238.232> has joined #oi-dev[08:47:45] *** tomww <tomww!~tom@unaffiliated/tomww> has quit IRC (Ping timeout: 264 seconds)[09:42:58] *** jimklimov <jimklimov!~jimklimov@89.24.55.249> has joined #oi-dev[09:47:50] *** jimklimov <jimklimov!~jimklimov@89.24.55.249> has quit IRC (Ping timeout: 276 seconds)[09:48:48] *** jimklimov <jimklimov!~jimklimov@89.24.55.249> has joined #oi-dev[09:54:23] *** jimklimov <jimklimov!~jimklimov@89.24.55.249> has quit IRC (Ping timeout: 248 seconds)[09:54:43] *** tomww <tomww!~tom@tom.bn-ulm.de> has joined #oi-dev[09:54:43] *** tomww <tomww!~tom@tom.bn-ulm.de> has quit IRC (Changing host)[09:54:43] *** tomww <tomww!~tom@unaffiliated/tomww> has joined #oi-dev[09:54:43] *** ChanServ sets mode: +o tomww[09:55:25] *** tomww <tomww!~tom@unaffiliated/tomww> has quit IRC (Client Quit)[09:55:43] *** tomww <tomww!~tom@tom.bn-ulm.de> has joined #oi-dev[09:55:43] *** tomww <tomww!~tom@tom.bn-ulm.de> has quit IRC (Changing host)[09:55:43] *** tomww <tomww!~tom@unaffiliated/tomww> has joined #oi-dev[09:55:43] *** ChanServ sets mode: +o tomww[10:01:13] *** jimklimov <jimklimov!~jimklimov@89.24.55.249> has joined #oi-dev[10:12:51] *** tsoome <tsoome!~tsoome@177-110-157-37.dyn.estpak.ee> has joined #oi-dev[10:13:45] *** mnowak_ <mnowak_!~mnowak_@94.142.238.232> has quit IRC (Ping timeout: 256 seconds)[10:26:17] *** ptribble <ptribble!~ptribble@cpc92716-cmbg20-2-0-cust138.5-4.cable.virginm.net> has joined #oi-dev[10:26:35] *** tsoome <tsoome!~tsoome@177-110-157-37.dyn.estpak.ee> has quit IRC (Quit: tsoome)[10:26:42] *** Toasterson <Toasterson!~toast@83.150.4.98> has joined #oi-dev[10:36:27] *** merzo <merzo!~merzo@46-24-133-95.pool.ukrtel.net> has quit IRC (Ping timeout: 240 seconds)[10:44:55] *** jimklimov <jimklimov!~jimklimov@89.24.55.249> has quit IRC (Ping timeout: 260 seconds)[10:48:20] <xenol> morning[10:49:26] *** wiedi <wiedi!~wiedi@ip5b40ee10.dynamic.kabel-deutschland.de> has quit IRC (Read error: Connection reset by peer)[10:51:32] *** jimklimov <jimklimov!~jimklimov@ip-86-49-249-141.net.upcbroadband.cz> has joined #oi-dev[11:07:00] *** jimklimov <jimklimov!~jimklimov@ip-86-49-249-141.net.upcbroadband.cz> has quit IRC (Ping timeout: 268 seconds)[11:07:59] *** jimklimov <jimklimov!~jimklimov@ip-86-49-249-141.net.upcbroadband.cz> has joined #oi-dev[11:15:57] *** jimklimov <jimklimov!~jimklimov@ip-86-49-249-141.net.upcbroadband.cz> has quit IRC (Ping timeout: 264 seconds)[11:15:59] *** wiedi <wiedi!~wiedi@ip5b40ee10.dynamic.kabel-deutschland.de> has joined #oi-dev[11:16:05] *** jimklimov <jimklimov!~jimklimov@ip-86-49-249-141.net.upcbroadband.cz> has joined #oi-dev[11:23:14] *** asx <asx!~asx@93-34-145-252.ip50.fastwebnet.it> has quit IRC (Remote host closed the connection)[11:26:13] *** tsoome <tsoome!~tsoome@177-110-157-37.dyn.estpak.ee> has joined #oi-dev[11:26:22] *** asx <asx!~asx@93-34-145-252.ip50.fastwebnet.it> has joined #oi-dev[11:38:11] *** jimklimov <jimklimov!~jimklimov@ip-86-49-249-141.net.upcbroadband.cz> has quit IRC (Quit: Leaving.)[11:38:27] *** jimklimov <jimklimov!~jimklimov@ip-86-49-249-141.net.upcbroadband.cz> has joined #oi-dev[12:04:12] *** jimklimov <jimklimov!~jimklimov@ip-86-49-249-141.net.upcbroadband.cz> has quit IRC (Quit: Leaving.)[12:04:43] *** jimklimov <jimklimov!~jimklimov@ip-86-49-249-141.net.upcbroadband.cz> has joined #oi-dev[12:06:06] *** merzo <merzo!~merzo@185.39.197.205> has joined #oi-dev[12:11:40] *** jimklimov <jimklimov!~jimklimov@ip-86-49-249-141.net.upcbroadband.cz> has quit IRC (Read error: Connection reset by peer)[12:11:58] *** jimklimov <jimklimov!~jimklimov@ip-86-49-249-141.net.upcbroadband.cz> has joined #oi-dev[12:20:03] *** jimklimov <jimklimov!~jimklimov@ip-86-49-249-141.net.upcbroadband.cz> has quit IRC (Read error: Connection reset by peer)[12:40:38] *** mnowak_ <mnowak_!~mnowak_@94.142.238.232> has joined #oi-dev[12:49:04] <alp_> hi[13:08:49] *** ptribble <ptribble!~ptribble@cpc92716-cmbg20-2-0-cust138.5-4.cable.virginm.net> has quit IRC (Quit: Leaving)[13:12:15] *** merzo <merzo!~merzo@185.39.197.205> has quit IRC (Ping timeout: 256 seconds)[13:15:50] *** merzo <merzo!~merzo@185.39.197.205> has joined #oi-dev[13:37:07] <tomww> someone mentioned that there will be a laptop at the illumos table running openindiana[13:37:42] <tomww> I would really like to make sure we have LinbreOffice running on that laptop...[13:38:25] <tomww> who is preparing that laptop? any ideas?[13:54:49] <Woodstock> i will certainly bring my laptop, which is running OI and has LO installed[13:55:16] <Woodstock> maybe i'll even bring a 2nd laptop for demo purposes[14:00:03] <alp_> I'm dropping library/g++/icu and system/library/g++/boost* from OI repo[14:00:12] <alp_> ( so it's not available now :) )[14:21:34] *** tsoome <tsoome!~tsoome@177-110-157-37.dyn.estpak.ee> has quit IRC (Quit: tsoome)[14:39:34] <alp_> tomww: ready[14:40:38] <richlowe> alp_: it'll take lots of time, but less time then rebuilding everything[14:41:02] <richlowe> alp_: dunno what I can do about finding the package(s) better based on the source, it was just meant to maybe speed tings up.[14:41:17] <richlowe> rather than needing 2 full rebuilds of everything.[14:42:28] <alp_> we need to discuss process with xenol and alarcher, but 1) I don't see alarcher now and 2) have some urgent things to do in real life.[14:46:14] <richlowe> oh, not a huge deal, I don't understand how you guys do things enough for "process" parts. I just threw that together 'cos it seemed like it might help[15:10:00] <oibot> Project oi-userland build #3726: SUCCESS in 29 min: https://hipster.openindiana.org/jenkins/job/oi-userland/3726/[15:14:13] *** wiedi <wiedi!~wiedi@ip5b40ee10.dynamic.kabel-deutschland.de> has quit IRC (Quit: ^C)[15:22:49] *** tsoome <tsoome!~tsoome@13-113-191-90.dyn.estpak.ee> has joined #oi-dev[16:33:37] *** tsoome_ <tsoome_!~tsoome@177-110-157-37.dyn.estpak.ee> has joined #oi-dev[16:35:35] *** tsoome <tsoome!~tsoome@13-113-191-90.dyn.estpak.ee> has quit IRC (Read error: Connection reset by peer)[16:35:56] *** tsoome_ <tsoome_!~tsoome@177-110-157-37.dyn.estpak.ee> has quit IRC (Read error: Connection reset by peer)[16:43:31] *** tsoome <tsoome!~tsoome@13-113-191-90.dyn.estpak.ee> has joined #oi-dev[16:50:54] *** jollyd <jollyd!~alarcher@2001:700:300:4104:863a:4bff:feae:6226> has joined #oi-dev[16:51:03] <jollyd> hello[16:53:08] <jollyd> xenol: here now[16:53:15] <xenol> hello[16:54:11] *** merzo <merzo!~merzo@185.39.197.205> has quit IRC (Ping timeout: 268 seconds)[16:54:19] <jollyd> sorry about the delay, got stuck after the presentation[17:03:39] *** pjama <pjama!~pjama@c211-31-53-64.kelvn3.qld.optusnet.com.au> has quit IRC (Ping timeout: 256 seconds)[17:07:13] *** pjama <pjama!~pjama@c211-31-53-64.kelvn3.qld.optusnet.com.au> has joined #oi-dev[17:20:35] <tomww> alp_: thank you very much! :)[17:21:58] <tomww> alp_: do you know users who tried installing libreoffice on the recent hipster snapshot? I would really like to make sure that it is installable there. I'm looking towards FOSDEM conference with the prominently placed Laptop Screen displaying the OpenIndiana Desktop...[17:24:07] <tomww> (who is bringing the laptop to FOSDEM? Does anyone remeber who wrote about that?)[17:25:56] <tomww> Woodstock: sorry, only saw your reply right now. Now I understand :)[17:26:00] *** jimklimov <jimklimov!~jimklimov@ip-86-49-249-141.net.upcbroadband.cz> has joined #oi-dev[17:26:43] <tomww> Woodstock: do you think LibreOffice will install on that demo laptop? If you could try that early, that would help me tune the existing packages accordingly.[17:28:32] <Woodstock> tomww: i'm not sure, and i can't test. and i'm not even sure i'll bring it, as i'm flying without checked baggage this time[17:39:25] <jimklimov> you can't check in laptops anyway ;)[17:39:43] <tomww> two arms, two laptops .) (sorry)[17:40:21] <jimklimov> and a humpy camel back!?[17:40:54] <tomww> In the airport: "Look those people there, they must be going to FOSDEM"[17:41:01] *** leoric <leoric!~leoric@188.232.89.90> has joined #oi-dev[17:42:23] <jimklimov> don't point fingers, chile, ait's rude! :)[17:42:39] <tomww> :)[17:43:15] <jimklimov> hope them ain't aggressive, or protective of their gear and those little ones, the lappies![17:43:35] <tomww> I'll rework my question. Is there anyone out ther with fairly recent Hipster installed who already has libreoffice 5.2 installed, or yould just try it with "dryrun"?[17:44:10] <jimklimov> from localhostoi?[17:44:59] <tomww> pfexec pkg install -nv libreoffice52-desktop-int[17:45:02] <tomww> yes[17:45:57] <jimklimov> I'll see how my internet goes now... maybe can try :)[17:46:04] <tomww> if dryrun gets happy, then I think it'l just work in real. Interesting part will be if the libsmb downgrade with the version facet is still necessary.[17:46:07] <tomww> thank you![17:46:45] <jimklimov> i'm in the middle of relocation, so got a modem just today[17:46:55] <tomww> if libsmb version facet is still necessary, then the focus shifts to samba and libiconv[17:47:01] <tomww> no hurry please! :)[17:47:01] <jimklimov> or so they called it... a several-hundred-megabit modem :\[17:47:20] <jimklimov> well, gotta test what works here and what needs a chisel ;)[17:47:54] <tomww> yeah, you are in luck if your location has offerst of that kind. even in germany you sometimes you get only 1 or 3 Mbits downstream[17:48:16] <jimklimov> I lived for 3 years in 40kbyte/s on good weather[17:48:24] <jimklimov> and nothing when the hatches got wet[17:49:15] <jimklimov> dryrun is upset about oi and sfe both providing libglew[17:49:21] <tomww> oh, so you got enough time for programming because loading a webpage gave you enough time between clicks, right[17:49:37] <jimklimov> otherwise no complaints :)[17:49:41] <tomww> ah, very goot catch.[17:49:59] <jimklimov> half a dozen of pairs like[17:49:59] <jimklimov> The following packages all deliver file actions to usr/lib/libGLEWmx.so.1.13.0:[17:49:59] <jimklimov> pkg://localhostoih/sfe/library/libglew at 1 dot 13.0,5.11-0.0.151.1.8:20160108T184333Z[17:49:59] <jimklimov> pkg://openindiana.org/x11/library/libglew at 1 dot 13.0,5.11-2017.0.0.0:20170306T133728Z[17:50:23] <tomww> maybe the complaints appear in a later phase. I'll check for glew ... most likely just a rename of the package on the SFE_OI side.[17:50:44] <jimklimov> or "pkg depend either of ... " :)[17:50:47] <tomww> it'l take some few days until I can make that[17:51:21] <jimklimov> ok, good luck[17:51:37] <tomww> I think I'll find out soon. Last question, what is your system's version of those packages right now:[17:52:02] <tomww> pkg info name osnet-incorporation entire | grep FMRI[17:52:26] <tomww> so I can replay here on the build-VM for SFE-OI package[17:52:41] * jimklimov called back to the cupboards ;)[17:52:43] <jimklimov> root@jimoi:/root# pkg info name osnet-incorporation entire | grep FMRI[17:52:43] <jimklimov> pkg: info: no packages matching the following patterns you specified are[17:52:43] <jimklimov> installed on the system. Try querying remotely instead:[17:52:43] <jimklimov> entire[17:52:43] <jimklimov> FMRI: pkg://openindiana.org/consolidation/osnet/osnet-incorporation at 0 dot 5.11-2017.0.0.16923:20171223T005150Z[17:52:43] <jimklimov> FMRI: pkg://openindiana.org/release/name at 0 dot 5.11-2017.0.0.5:20171030T214925Z[17:53:09] <tomww> entire is uninstalled?[17:53:09] <jimklimov> I guess I nuked entire back when that was fashionable :)[17:53:24] <tomww> entire is not not important to me.[17:53:34] <tomww> thank you, I'm now perfectly prepared.[17:53:36] <jimklimov> not unimportant?[17:53:38] <jimklimov> :)[17:53:46] <jimklimov> good luck nailing this :)[17:54:08] <tomww> haha, thanks. It this is only glew, than it is manageable :)[17:54:17] <tomww> *If[17:55:13] *** wiedi <wiedi!~wiedi@ip5b409739.dynamic.kabel-deutschland.de> has joined #oi-dev[17:58:41] <leoric> jollyd: is anybody in there?[18:00:42] <leoric> can you add richlowe's changes to gcc-6 branch, run gmake REQUIRED_PACKAGES and test full rebuild? ;)[18:00:55] <leoric> or should we do it in a different way?[18:01:56] <leoric> I can create new build zone, update it to gcc-6 packages and participate in the fun[18:27:11] *** Toasterson <Toasterson!~toast@83.150.4.98> has quit IRC (Ping timeout: 276 seconds)[18:27:23] <richlowe> leoric: the crappy script I provided should nevertheless be faster/almost-better than 'make required_packages', even if it leaves some mess to clean up by hand.[18:27:47] <richlowe> (I can't imagine what a bulk-build of oi-userland looks like right now, so I'm perhaps misjudging which bits make it suck)[18:28:13] <leoric> perhaps[18:28:57] <leoric> but in any case these changes should be made in gcc-6 branch , as otherwise it'll mean one more oi-userland rebuild[18:29:12] <tomww> jimklimov: well, solution is already there since a while, need to rebuild SFE libglew which is relocated to /usr/gnu/ offset. That is the smallest change I can do at the moment. Package will be updloaded after "autobuild" has finished[18:29:12] <richlowe> he _advantage_ of 'make REQUIRED_PACKAGES' is you can compare the time that takes, and the time the following build to test its output takes.[18:30:32] <jacobs> yes, and there is a big difference in the time that 'pkgdepend resolve' takes when you limit the scope of resolution.[18:30:35] <jimklimov> jollyd: leoric: do your builds use ccache by the way?[18:30:57] <jollyd> sorry got caught in a discussion with a colleague[18:31:09] <jimklimov> if yes, then the bulk rebuild does eat a lot of time in configure, unpack, make... but not in compile at least :)[18:31:34] <jimklimov> and it is often half or more of the time in raw builds from scratch[18:33:04] <richlowe> jimklimov: do you actually have ccache working? (in general, for anything)[18:33:13] <jimklimov> our project at work enjoys a 93% hit rate over the past half a year as we iterate, chopping product rebuilds 3 to 10 times for different components[18:33:14] <leoric> I suppose resolve takes several minutes... Let's say, 2 mins in average. I suppose, we have >1000 components, winning at least 1.5 minutes per component is a huge win[18:33:41] <jimklimov> yes, colleagues before me in the project got me on this needle ;)[18:34:04] <leoric> jollyd: are you free now?[18:34:16] <jollyd> leoric: should I pick your latest pushes to rich's branch as well?[18:34:18] <jollyd> yes[18:34:21] <leoric> yes[18:34:25] <jimklimov> for pretty deterministic cases, where result of preprocessor is the same as built earlier, it is quite safe to take the cached object file[18:34:37] <leoric> it'll help such components as mariadb[18:34:54] <richlowe> jimklimov: yeah, the problem has been most solaris-y builds have been using a compiler it didn't like (Studio), and done in ways it liked even less.[18:35:04] <richlowe> I think a GCC-built userland maybe the first (and only) real opportunity for it.[18:35:19] <jimklimov> and guess what we have for years now?[18:35:31] <jimklimov> a gcc kernel and gcc userland, both speedupable by ccache (tested)[18:35:40] <richlowe> getting it through cw for ON has never worked as properly as it should.[18:35:41] <leoric> we also should think what can we do about 'depend' actions in manifests[18:35:52] <leoric> will they break current resolve logic?[18:36:04] <richlowe> leoric: no.[18:36:30] <jimklimov> illumos-gate does benefit from ccache, but spending hours in Make it is not quite visible that a few minutes of actual compilation were not spent :)[18:37:23] <leoric> richlowe: if package has manual depend fmri=X type=require, should X be added to REQUIRED_PACKAGES?[18:37:25] <jimklimov> will you be at FOSDEM? I can just show then ;)[18:38:51] * jimklimov offliing now[18:39:10] <richlowe> leoric: I think so, yes.[18:39:18] <jollyd> leoric: is gmake REQUIRED_PACKAGES supposed to work from the components top directory?[18:39:29] <richlowe> leoric: though if pkgdepend is going to find it, you have to wonder why it's also specified manually.[18:39:30] <leoric> I think, yes[18:39:49] <richlowe> leoric: so more likely, in cases where it's noticed it should _only_ be in REQUIRED_PACKAGES[18:39:50] <leoric> richlowe: well, checking that is eady[18:40:10] <richlowe> you can't prove the opposite, unfortunately.[18:40:37] <richlowe> and I suspect components aren't so diligent with comments about whether things are specified because pkgdepend can't find them, or not.[18:41:04] <jollyd> leoric: it does not seem to be the case[18:42:02] <leoric> yes[18:43:11] <leoric> jollyd, I'm sorry, but it seems we've found the main actor in this show :)[18:43:38] <leoric> So, the plan.[18:43:52] <jollyd> sorry after teaching and coding my head is pretty messed up, you'll need to talk to me in a very simple way :)[18:43:54] <leoric> Add resolve change to gcc-6 branch[18:43:57] <jollyd> done[18:44:18] <leoric> add REQUIRED_PACKAGES to packages wich miss them[18:44:44] <leoric> (either global gmake REQUIRED_PACKAGES or richlowe's makefile)[18:45:17] <leoric> look through result to fix possible issues[18:45:19] <leoric> commit[18:45:43] <leoric> rebuild everything and fix broken things[18:45:55] <jollyd> leoric: where is rich's makefile?[18:46:11] <leoric> https://gist.github.com/richlowe/df865c4bf286de91b60bad13bd97786c[18:46:33] <leoric> On the last stage I can help if subtasks are clearly set :)[18:47:26] <leoric> Should we fill and announce gcc-6 repository now or when the task is finished?[18:47:43] <jollyd> when it is finished would be best[18:48:06] <richlowe> jollyd: Note that that makefile is a travesty of badness.[18:48:07] <leoric> when are you going to dive into this?[18:49:51] <jollyd> I am trying to run now[18:49:57] <jollyd> before going home[18:50:17] <leoric> what assistance do you need?[18:53:02] <jollyd> hmm actually I do not see any change so far with fix-stuff.mk[18:54:10] <leoric> what do you mean ? does it still run pkgdepend?[18:55:39] <leoric> the idea was to use this makefile and run gmake fix[18:56:35] <jollyd> I did[18:57:07] <leoric> also it requires some work[18:57:14] <jollyd> I am missing some packages in the zone[18:57:39] <jollyd> so pkg contents does not succeed for some components[18:57:48] <leoric> COMPONENT_FMRI=$$(pkgmogrify /tmp/name-mog *.p5m | grep -v '^#') seems to be wrong[18:58:07] <jollyd> so I get "pkg contents: contents: must request remote contents for specific packages"[18:58:53] <jollyd> I do get a few REQUIRED_PACKAGES on the terminal output that seem missing from the Makefile[18:59:19] <jollyd> for perl packages the makefile throws "pkg: Illegal FMRI 'library/perl-5/mysql-dbi-$(PLV)': Invalid Package Name: library/perl-5/mysql-dbi-$(PLV)"[18:59:32] <leoric> this was expected[18:59:48] <leoric> pkgmogrify misses args[19:00:22] <leoric> I'd used gmake REQUIRED_PACKAGES, as your zone likely already have all packages built and published[19:00:54] <leoric> perhaps, somehow filter out packages which already have REQUIRED_PACKAGES declared in Makefile[19:01:14] <leoric> yes, gmake components.mk[19:01:28] <leoric> find component's Makefile[19:01:35] <leoric> grep for REQUIRED_PACKAGES[19:02:01] <leoric> if they are missed, gmake REQUIRED_PACKAGES in component's dir[19:06:54] <leoric> for file in $(cat components.mk| awk '{ print $NF"/Makefile" }' ) ; do grep -q REQUIRED_PACKAGES $file && continue; dir=$(dirname $file); gmake -C $dir REQUIRED_PACKAGES; done[19:07:08] <leoric> something like this should work[19:08:02] <jollyd> thank you[19:15:08] <jollyd> for some reason the network is crap tonight....[19:22:04] <jollyd> running now[19:31:36] <jollyd> leoric, xenol: basically the stuff in common-lisp gets an empty generated list[19:31:44] <jollyd> they need manual addition of deps[19:41:33] <jollyd> leoric, xenol: commits landing there https://github.com/alarcher/oi-userland/commits/gcc6[19:42:51] <xenol> what can I do?[19:47:42] <xenol> btw running gmake REQUIRED_PACKAGES is slow[19:47:58] <xenol> generated build logs already contain the information about dependencies[19:49:09] *** jimklimov <jimklimov!~jimklimov@ip-86-49-249-141.net.upcbroadband.cz> has quit IRC (Quit: Leaving.)[19:55:45] <jollyd> damn last illumos-gate compile with gcc-6 does not boot[19:55:46] <jollyd> ...[19:55:52] <jollyd> what did I do wrong...[20:15:03] <jollyd> nothing... reboot and good[20:15:06] <jollyd> weird[20:15:11] *** jollyd <jollyd!~alarcher@2001:700:300:4104:863a:4bff:feae:6226> has quit IRC (Quit: Leaving.)[20:42:03] *** jollyd <jollyd!~alarcher@188.113.95.225> has joined #oi-dev[20:42:59] <jollyd> back[21:13:51] *** wiedi_ <wiedi_!~wiedi@ip5b409739.dynamic.kabel-deutschland.de> has joined #oi-dev[21:17:45] *** wiedi <wiedi!~wiedi@ip5b409739.dynamic.kabel-deutschland.de> has quit IRC (Ping timeout: 264 seconds)[21:46:28] *** jollyd <jollyd!~alarcher@188.113.95.225> has quit IRC (Quit: Leaving.)[21:48:59] *** mnowak_ <mnowak_!~mnowak_@94.142.238.232> has quit IRC (Quit: Leaving)[21:51:21] *** leoric_ <leoric_!~leoric@176.212.32.36> has joined #oi-dev[21:52:57] *** leoric <leoric!~leoric@188.232.89.90> has quit IRC (Ping timeout: 240 seconds)[21:59:48] <tomww> yeah, suddenly I can see 7 downloads if the LO package Hipster LO 52 5.2.5.1[22:01:52] <tomww> btw, libglew SFE package is now updated in the repo, so no longer in conflict for recent hipster versions. If someone needs linglew from SFE for hioster version before system/kernel at 0 dot 5.11-2017.0.0.16925 system/library at 0 dot 5.11-2017.0.0.16925, then please let me know I'll make the package for older installes then.[22:34:13] *** Toasterson <Toasterson!~Thunderbi@2a02:1205:34f9:26c0:71be:3eaa:2b11:65e3> has joined #oi-dev[22:54:42] <tomww> 3[23:28:07] *** jollyd <jollyd!~alarcher@188.113.95.225> has joined #oi-dev[23:33:36] *** mnowak_ <mnowak_!~mnowak_@94.142.238.232> has joined #oi-dev[23:42:20] *** asx <asx!~asx@93-34-145-252.ip50.fastwebnet.it> has quit IRC (Quit: Leaving)[23:48:58] *** asx <asx!~asx@93-34-145-252.ip50.fastwebnet.it> has joined #oi-dev