[17:15:07] *** echelog-2 <email@example.com> has joined #oi-dev [17:15:15] <alarcher> richlowe: I added them earlier today and have republished the components. testsuite still running... [17:17:35] *** merzo <firstname.lastname@example.org> has quit IRC (Ping timeout: 248 seconds) [17:17:47] *** Toasterson <Toastersonemail@example.com> has quit IRC (Ping timeout: 260 seconds) [17:26:43] *** tozhu <tozhu!~Thunderbi@22.214.171.124> has joined #oi-dev [17:27:25] *** tozhu <tozhu!~Thunderbi@126.96.36.199> has quit IRC (Client Quit) [17:46:38] <xenol> alarcher: I think so [18:01:21] *** ptribble <firstname.lastname@example.org> has joined #oi-dev [18:14:20] <alarcher> alp_: should we have mongodb use gcc-6 already? [18:14:40] <alarcher> alp_: then we could merge gcc 7.3.0 without being afraid of breaking deps [18:18:27] *** Toasterson <Toasterson!~toast@2a02:aa10:e280:b880:e3d5:f722:2d0a:1d1a> has joined #oi-dev [18:36:04] <alarcher> richlowe: according to bug reports most guality failures could be attributed to our gdb [18:38:52] <richlowe> people complain about gdb infrequently, but I have no actual idea what's wrong with it [18:38:54] <richlowe> (I don't use it) [18:40:06] <Agnar> *sigh* pkg(1M) planning for linked zones should be done parallel, not sequential [18:40:08] <igork> we are using gdb on dilos a lot [18:40:23] <richlowe> Agnar: isn't that a matter of telling pkg to do that? [18:41:01] <Agnar> richlowe: there's no flag for it afaik, so I think someone must touch the code [18:42:41] <richlowe> also, whoever packages pkg for OI left in the oracle code to remove the CDDL. [18:42:42] <richlowe> don't do that. [18:45:36] <alarcher> richlowe: I guess alp should know about it [18:47:57] <Agnar> oh and while we are chatting. 3 years ago I had a talk with ben rockwood who told me that he brought back the shared zones (inheriated-pkg-dir) to illumos/smartos. at least I can't find it in OI. does anyone know more about that? [18:51:07] <richlowe> smartos does it, tribblix does it. [18:51:10] <richlowe> there are good reasons pkg5 can't. [18:51:28] <Agnar> ah, so theres not support for IPS [18:54:53] <alarcher> uh just checking my mailbox I found a tarball named FSWfsmisc from 2006 [18:55:32] <richlowe> the greatest evil of svr4 packages, is that nobody ever had time to realize the limit on the length of their name got lifted [18:55:34] <richlowe> like, in pre-history [18:55:43] <alarcher> ext2fs, ntfs in userland for something that should have been OpenSolaris [18:55:49] <alarcher> or Solaris Express [18:55:56] <richlowe> and of course, alanc's fault for SUNWxplt [18:56:19] <alarcher> your "blame ala<TAB>" rule stands [18:56:48] <richlowe> alarcher: there's a bunch of fuse implementations, all separately sketchy. gwr knows about it, I think. [18:56:58] <richlowe> there was also an in-kernel ext2 project, but I don't know how far it got [18:57:03] <richlowe> I have the hg repo somewhere tohugh. [18:57:38] <alanc> SUNWxwplt was named when the limit was still 9 chars, and got baked into too many pkg dependencies to change until IPS [18:57:56] <Agnar> xenol: ping? did you got ansible built on hipster with pip install? [18:58:02] <alanc> at least Solaris no longer has a SUNWcs package... [18:58:15] <richlowe> alanc: every so often I plan to fix that [18:58:24] <richlowe> alanc: and then realize I don't know who I'd have to coordinate with. [18:59:26] <richlowe> luckily, then people get excited about compilers, or break ld [18:59:30] <richlowe> or have fun with streams [18:59:32] <richlowe> and I get distracted. [18:59:37] <richlowe> for very very specific values of "luckily" [18:59:59] <alanc> well, you wouldn't necessarily *have* to coordinate with Schillix, but you'd still hear about it from him [19:00:22] <richlowe> I think OI and omniti's installers would need to know, and probably tribblix [19:00:34] <richlowe> because he builds SVR4 packages based on an IPS image [19:01:23] <alanc> who would you rather have to listen to for hours? Schillix or Skrillex? [19:01:57] <alarcher> ouch [19:02:15] <alarcher> I found another nugget in my mailbox: https://imgur.com/a/9t3tA [19:02:48] <alarcher> I actually used it during my PhD :) [19:04:53] <richlowe> don't think I have many depressing screenshots [19:04:57] <richlowe> _many_ depressing emails though. [19:05:26] <richlowe> if there's a crying-poop-emoji, that's basically mbox circa 2010 [19:05:40] <alarcher> I can make screenshots everytime I break ld if you need :) [19:05:54] <alarcher> +\s [19:06:28] <alanc> have found that given enough time, depressing bug reports can turn into fun memories... [19:07:26] <alanc> like "omg, do you remember when Studio cc started segfaulting in November after working fine all through October, and it turned out the length of the name of the month caused a buffer overflow?" [19:07:37] <alarcher> ahaha [19:08:34] *** mnowak_ <email@example.com> has joined #oi-dev [19:08:54] *** Toasterson <Toasterson!~toast@2a02:aa10:e280:b880:e3d5:f722:2d0a:1d1a> has quit IRC (Ping timeout: 256 seconds) [19:09:52] <alanc> "and it was all because alanc had been excessively cute in putting date +'%e %B %Y' output in the mcs strings..." [19:11:35] <alanc> so you see, there are good reasons richlowe has a "blame alanc" hotkey 8-) [19:13:19] <alarcher> alanc: is it on a standard type-5 kbd? if so you are safe, it is about to be deprecated in illumos :P [19:14:13] <alanc> you still support those? we killed them years ago, but that's when we EOL'ed sun4u too... [19:17:45] *** MarcelT <MarcelTfirstname.lastname@example.org> has quit IRC (Ping timeout: 265 seconds) [19:21:30] <ptribble> richlowe: I don't really care what the package names are, as long as I know if they change [19:21:56] <richlowe> I did not intend, by any means, to talk my way into actually fixing it! [19:36:52] <alarcher> richlowe: vanilla gcc tests still running, yesterday before your email I had built/run with all your patches, then without msave-args only. I will upload the three tarballs later tonight when the vanilla testsuite is done. [19:37:12] <alarcher> I will also produce the compare_test output [19:38:02] <richlowe> well, if you run the compare_tests stuff (maybe I have a script somewhere?) it's mostly the diff that matters. [19:38:18] <richlowe> the major thing that always makes me worry is clean tests v. tests with -msave-args passed [19:38:38] <richlowe> but there's a variety of reasons they fail, because using dejagnu isn't the easiest thing [19:38:46] <alarcher> ok [19:39:02] <richlowe> I never did figure out how to pass -msave-args only in places it should go for a test run [19:39:11] <richlowe> I just passed it everywhere, and ignored the ones that failed because it couldn't ever work [19:43:53] *** Toasterson <Toastersonemail@example.com> has joined #oi-dev [20:01:17] *** mnowak_ <firstname.lastname@example.org> has quit IRC (Ping timeout: 256 seconds) [20:12:45] *** MarcelT <MarcelT!~marcel@p4FC24C88.dip0.t-ipconnect.de> has joined #oi-dev [20:15:35] *** mnowak_ <email@example.com> has joined #oi-dev [20:16:13] <Agnar> while doing pip install ansible, gcc tries to link some stuff and fails: https://paste.ec/paste/H1Gea51w#b3KMXia7UvlCyWOdOlF57bcinllTemwhdl99iQ9L9Lr - after googling I found that -fno-stack-protector should help, but it does not. any other ideas? [20:16:55] <richlowe> did you pass it when you built libsodium, too? [20:17:03] <richlowe> the other option is linking libssp (I think) [20:20:33] *** mnowak_ <firstname.lastname@example.org> has quit IRC (Ping timeout: 248 seconds) [20:22:13] *** mnowak_ <email@example.com> has joined #oi-dev [20:24:26] <Agnar> richlowe: good point, I'll clean up the complete pip cached foo [20:25:54] <Agnar> and luckily other OSses have their update issues as well. fedora disabled my explicit enabled kvm_intel nested=1 option again after an update. and I wonder why my vms fail now [20:26:11] <Agnar> took me just two weeks to figure out. :( [20:28:04] <Agnar> richlowe: fails again at the same point. having -lssp also in LDFLAGS [20:46:38] *** psydroid <psydroid!~psydroid@unaffiliated/psydroid> has quit IRC (Remote host closed the connection) [20:53:25] *** psydroid <psydroid!~psydroid@unaffiliated/psydroid> has joined #oi-dev [21:44:30] <richlowe> Agnar: then I don't have ideas, sorry :\ [21:44:43] <richlowe> I'd expect xenol to know, though. [21:59:09] <alarcher> richlowe: sorry I am stuck in a parallel mesh nightmare, not sure I can be active until tmrw. [21:59:57] <richlowe> I'm doing horrible things to illumos :) [22:00:29] <alarcher> richlowe: I am writing C++ with tr1 containers, fatality! :P [22:01:30] *** psydroid <psydroid!~psydroid@unaffiliated/psydroid> has quit IRC (Read error: Connection reset by peer) [22:01:32] *** psydroid_ <firstname.lastname@example.org> has joined #oi-dev [22:32:22] *** mnowak_ <email@example.com> has quit IRC (Ping timeout: 252 seconds) [22:38:22] *** merzo <firstname.lastname@example.org> has joined #oi-dev [22:50:57] *** Toasterson <Toastersonemail@example.com> has quit IRC (Ping timeout: 248 seconds) [23:01:11] *** leoric <firstname.lastname@example.org> has joined #oi-dev [23:29:08] *** Toasterson <Toasterson!~toast@2a02:aa10:e280:b880:e3d5:f722:2d0a:1d1a> has joined #oi-dev [23:46:43] <richlowe> anyone worked on a binutils update lately?