Switch to DuckDuckGo Search
   March 10, 2019  
< | 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 | >

Toggle Join/Part | bottom
[00:09:14] *** idodeclare <idodeclare!~textual@cpe-76-185-177-63.satx.res.rr.com> has joined #illumos
[00:13:57] *** alanc <alanc!~alanc@> has quit IRC (Remote host closed the connection)
[00:14:24] *** alanc <alanc!~alanc@> has joined #illumos
[00:16:18] <LeftWing> https://www.illumos.org/n/ -- thoughts? feelings?
[00:19:55] <wiedi> progress! :D
[00:20:22] <toasterson> LeftWing I love that the first link is to the dtrace book. Even I would have not found several reources without this page. Do we have more links we could add there?
[00:21:40] <LeftWing> I'd love to not drown the page in content
[00:22:10] <LeftWing> There's some work on a more comprehensive structured set of documentation -- https://illumos.github.io/docs/
[00:22:16] <toasterson> That is true aswell.
[00:22:18] <LeftWing> I just think we want a simpler front page
[00:22:31] <LeftWing> Most other projects with newer websites have a relatively uncluttered landing page
[00:23:05] <wiedi> it's great! The content is good
[00:23:10] <toasterson> But there is no link to the docs there right?
[00:24:26] <igork> https://illumos.github.io/docs/about/distro/
[00:24:34] <igork> what is it - Linux Comparison ?
[00:25:27] <igork> for DilOS - it should be Debian - because we are using Debian userland apps where are possible
[00:26:04] <igork> and have to add 'spac' column :)
[00:26:08] <igork> sparc
[00:26:12] <LeftWing> toasterson: The /docs/ stuff isn't quite ready yet
[00:26:20] <LeftWing> I need to trim out some old stuff
[00:26:37] <wiedi> there are a few tiny style issues (the logo doesn't have the right logo font and loading the google font is a bit slow, might be easy to just do the text in svg too. I think the header could be a bit smaller)
[00:27:22] <toasterson> ah ok planed is good from my perspective too. Just don't delay it too long. I like the new docs page.
[00:27:27] <richlowe> LeftWing: manual email?
[00:27:53] <LeftWing> richlowe: Eh?
[00:28:32] <richlowe> you never sent your ipd
[00:28:46] <LeftWing> richlowe: How astute of you
[00:28:53] <LeftWing> igork: There _is_ a SPARC column
[00:29:05] <richlowe> no need for sarcasm.
[00:29:06] <LeftWing> You may need to scroll. Like I said, it isn't finished yet.
[00:29:54] <igork> LeftWing: https://cl.ly/3dbef3553604
[00:30:18] <igork> it table with maintained distributions - have no SPARC
[00:30:33] <LeftWing> You may need to scroll. Like I said, it isn't finished yet.
[00:30:34] <richlowe> it does say _maintained_ ;)
[00:30:44] <richlowe> (I take back what I said about there being no need for sarcasm)
[00:30:45] <richlowe> seems I was wrong.
[00:32:07] <igork> LeftWing: APT - it is not packaging. correct packaging is DPKG
[00:32:17] <igork> apt is frontend to dpkg
[00:32:30] <LeftWing> igork: Mate, I'm not taking dictation right now.
[00:32:47] <igork> just correction
[00:32:58] <tomww> '15
[00:33:05] <toasterson> Pull Requests ftw
[00:34:12] <LeftWing> Anyway, I didn't mention the /docs stuff because I'm looking at it today. I'm just looking at replacing the front page of illumos.org with https://www.illumos.org/n/ instead of the wiki
[00:35:13] <richlowe> seems good
[00:35:23] <toasterson> ship it
[00:35:38] *** andy_js <andy_js!~andy@> has quit IRC (Quit: andy_js)
[00:36:01] <LeftWing> I'm looking for "Museo Sans" now
[00:36:44] <LeftWing> Oh good lord it's a non-free font
[00:38:06] <toasterson> Why Museo Sans. Is that the font in the illumos logo?
[00:38:22] <wiedi> yes
[00:38:26] <igork> LeftWing: link to CI=jenkins and zfs tests are on plans?
[00:38:35] <wiedi> https://wiki.illumos.org/display/illumos/The+illumos+Logo
[00:39:45] <richlowe> LeftWing: http://sensi.org/~svo/glasstty/ surely
[00:40:07] <toasterson> uh hmm do we need a license then? Or do we have it?
[00:40:24] <igork> also, link to 'contribution process' on main page will be fine too
[00:43:28] <LeftWing> toasterson: I suspect the license allowed the SVG creator to use it in an SVG
[00:44:06] <igork> also, links will be better put between [] - because with next visit not easy found a place for links - https://cl.ly/8bee3681df58
[00:44:42] <igork> just feedback
[00:44:45] <LeftWing> I can adjust the visited colour to be brighter
[00:44:59] <igork> will be fine too
[00:46:01] <toasterson> Do we know who created the SVG's?
[00:46:40] <wiedi> I have the font and can make changes to the svgs
[00:47:19] <LeftWing> Do you have some pre-SVG format, or do you use an SVG editor
[00:47:43] <toasterson> the question is if we have something that would tell us what the license situation is.
[00:47:57] <wiedi> I'd use inkscape
[00:50:28] <wiedi> toasterson: https://www.exljbris.com/eula.html "This font is free for personal and commercial use"
[00:51:19] <toasterson> ah thanks that answers my questions
[00:51:31] <LeftWing> Oh
[01:14:39] *** snuff-work <snuff-work!~snuff-wor@202-161-112-134.tpgi.com.au> has quit IRC (Read error: Connection reset by peer)
[01:15:25] *** snuff-work <snuff-work!~snuff-wor@202-161-112-134.tpgi.com.au> has joined #illumos
[01:30:38] <LeftWing> Alright, due to popular demand I used the SVG
[01:30:47] <LeftWing> The link colours are all just blue now
[01:31:11] <wiedi> \o/ thanks! :D
[01:33:06] <LeftWing> I added a Contribution Guide link to the resources list, though the page we have clearly needs some cleaning up too; i.e., https://wiki.illumos.org/display/illumos/How+To+Contribute
[01:41:08] *** idodeclare <idodeclare!~textual@cpe-76-185-177-63.satx.res.rr.com> has quit IRC (Quit: My MacBook has gone to sleep. ZZZzzz…)
[02:15:09] <LeftWing> Alright, I've hopefully managed to put the site correctly at https://illumos.org/
[03:07:04] *** amrfrsh <amrfrsh!~Thunderbi@> has quit IRC (Ping timeout: 246 seconds)
[03:38:34] *** idodeclare <idodeclare!~textual@cpe-76-185-177-63.satx.res.rr.com> has joined #illumos
[04:42:19] <_Tenchi_> LeftWing: i saw your tweet bitching about 301 redirects and it turned out to be the missing piece of a puzzle in an authentication bug i've been dealing with for weeks... thanks for that ;)
[04:42:31] <LeftWing> Haha, you're welcome
[04:42:41] <_Tenchi_> blew away all browser history for localhost and everything started working haha
[04:42:59] <_Tenchi_> somewhere along the line a 301 was messing things up
[05:10:14] <bdha> LeftWing: You made monitoring so mad!
[05:10:27] <LeftWing> Not on purpose!
[05:10:50] * bdha tries to fix
[05:10:54] <bdha> Yay for new front page!
[05:11:03] <bdha> I should finish the rest? :(
[05:14:48] <bdha> LeftWing: Should clear in a minute, hopefully.
[05:15:06] <_Tenchi_> https://wiki.illumos.org/display/illumos/OmniTI+Jobs 👍
[05:15:51] <richlowe> bdha: yeah, finish the rest!
[05:16:08] <bdha> _Tenchi_: https://github.com/illumos/docs https://illumos.github.io/docs/
[05:16:19] <bdha> Pruned... needs more work.
[05:16:28] <bdha> richlowe: I'm gonna eat a popsicle first.
[05:17:10] <richlowe> now I want a popsicle.
[05:21:11] *** rsully <rsully!~rsully@unaffiliated/rsully> has quit IRC (Quit: rsully)
[05:23:31] *** hemi770_ <hemi770_!~hemi666@unaffiliated/hemi770> has joined #illumos
[05:26:17] *** hemi770 <hemi770!~hemi666@unaffiliated/hemi770> has quit IRC (Ping timeout: 245 seconds)
[05:30:13] *** hemi770_ <hemi770_!~hemi666@unaffiliated/hemi770> has quit IRC (Ping timeout: 246 seconds)
[05:32:04] *** hemi770 <hemi770!~hemi666@unaffiliated/hemi770> has joined #illumos
[05:53:23] <bdha> hm.
[05:56:50] <bdha> wiedi's changes break nav weirdly. Keeps current element in the url when you click around.
[05:57:09] <bdha> Going to cache mkdocs back in. :\
[06:07:04] <bdha> https://imgur.com/a/kNqSrsu
[06:15:25] <bdha> Ok, fixed that bug. Going to merge weidi's bits in.
[06:15:57] <richlowe> bdha: wait, what happened to my popsicle?
[06:20:36] <bdha> Yo, I'm unemployed, I can't just be sharing with the whole class.
[06:23:23] <bdha> Updated: https://illumos.github.io/docs/
[06:23:30] <bdha> (Though caching may be weird)
[06:27:56] <bdha> Confluence diff. >:\
[06:34:11] <bdha> LeftWing: Can we lock the wiki?
[06:35:34] <LeftWing> Maaaaybe?
[06:36:36] <bdha> Resyncing content from it. I guess it's not a big deal if I keep on it. And I still don't have a clear idea of what needs to be _removed_, really.
[06:36:39] <bdha> shrug
[06:36:44] <LeftWing> The big thing to update is the build instructions and I'm going to rewrite that all anyway
[06:36:55] <LeftWing> So... Don't worry about that one particularly
[06:37:22] <LeftWing> I think we're carrying a lot of .. shall we say "fluff"
[06:37:25] <richlowe> why is software? :\
[06:37:50] <bdha> Because software is.
[06:38:06] <LeftWing> Thankfully it is, rather than isn't, or I'd have to farm or whatever.
[06:39:42] <bdha> I wonder if we can tag/pin headsup and similar in topicbox. And then just link to that.
[06:39:53] <bdha> Rather than maintaining the list.
[06:40:45] <LeftWing> Unclear.
[06:40:51] <LeftWing> But that'd be cool.
[06:41:39] <LeftWing> The nav tree is currently very busy -- I wonder if we might be better turning this into the Handbook, a la FreeBSD
[06:42:30] <bdha> That was the original intention like 10 years ago!
[06:42:32] <LeftWing> Like we can have a front page, with a prominent Documentation link that goes into this tree. Then make sure it has only relevant up-to-date documentationy things
[06:42:36] <LeftWing> Ha
[06:42:40] <LeftWing> I think it's still a good plan
[06:43:17] <LeftWing> I was going to review all this today btw, until I realised how much there was and decided to patch the burning lack of a real front page first
[06:43:37] <LeftWing> With ma Cee Ess Ess
[06:43:38] <bdha> Real front page was a pretty straight-forward win.
[06:43:49] <LeftWing> I used a framework!
[06:43:56] <bdha> Did you _write_ that framework?
[06:44:01] <LeftWing> I didn't even!
[06:44:08] * bdha dubious
[06:44:09] <LeftWing> This isn't the fucking terminal man come on
[06:44:34] <LeftWing> I actually started by doing View Source on rust-lang.org if you must know
[06:44:47] <bdha> I started working a book last week and rjbs was like "do you need to write a LiquidPlanner clone so you can manage working on the book?"
[06:44:50] <bdha> I know how this works.
[06:45:02] <richlowe> sounds legit.
[06:45:09] <LeftWing> Ha. I wrote vtmc when I was supposed to be writing slides for a talk
[06:58:48] <bdha> Resync'd the changes from the last year. Not much. Ignored the build instructions, mostly.
[06:59:05] <bdha> updating the GH issues, adding a "blocker" label.
[07:00:21] *** TheGreekOwl <TheGreekOwl!~Arkomeda@2a02:587:4832:3800:c875:22de:2d43:c47f> has quit IRC (Quit: I must go, my planet needs me.)
[07:01:56] <LeftWing> Ok cool
[07:02:23] <bdha> https://github.com/illumos/docs/issues # PRs and crap welcome.
[07:21:58] <alanc> Popsicle Requests?
[07:22:32] <bdha> Not currently taking popsicle requests.
[07:24:53] <alanc> well that's no fun, though it would be melted by the time it crossed the country anyway
[07:27:10] <alanc> the new front page does look very nice though
[07:27:14] <bdha> LeftWing: I kind of feel like the Projects pages should be ported to IPDs?
[07:27:54] <bdha> alanc: It does!
[07:28:02] <bdha> LeftWing++
[07:28:06] <LeftWing> Thanks!
[07:41:57] *** idodeclare <idodeclare!~textual@cpe-76-185-177-63.satx.res.rr.com> has quit IRC (Quit: My MacBook has gone to sleep. ZZZzzz…)
[07:49:27] *** hadfl_ <hadfl_!~hadfl@omniosce.ee.ethz.ch> has joined #illumos
[07:57:57] *** hadfl <hadfl!~hadfl@omniosce.ee.ethz.ch> has quit IRC (*.net *.split)
[07:57:57] *** jhot[m] <jhot[m]!jhotmatrix@gateway/shell/matrix.org/x-bxijyohtxiojulul> has quit IRC (*.net *.split)
[07:57:58] *** xar- <xar-!xar-@2600:3c00::f03c:91ff:fe24:4661> has quit IRC (*.net *.split)
[08:45:34] <tsoome> good stuff:)
[08:54:22] <tsoome> btw, I think we can drop the loader pages from wiki
[09:45:11] *** andy_js <andy_js!~andy@> has joined #illumos
[10:53:01] *** jimklimov <jimklimov!~jimklimov@> has joined #illumos
[10:56:41] *** jimklimov1 <jimklimov1!~jimklimov@> has joined #illumos
[11:00:09] *** jimklimov <jimklimov!~jimklimov@> has quit IRC (Ping timeout: 246 seconds)
[11:02:32] *** Zakureth <Zakureth!~IceChat9@66-190-248-86.dhcp.astr.or.charter.com> has quit IRC (Ping timeout: 245 seconds)
[11:12:02] *** snuff-work <snuff-work!~snuff-wor@202-161-112-134.tpgi.com.au> has quit IRC (Read error: Connection reset by peer)
[11:49:46] *** AndrewG <AndrewG!~andrew@a20.cucumber.me.uk> has quit IRC (Ping timeout: 268 seconds)
[12:17:19] *** AndrewG <AndrewG!~andrew@a20.cucumber.me.uk> has joined #illumos
[13:02:02] *** amrfrsh <amrfrsh!~Thunderbi@> has joined #illumos
[13:25:35] <jimklimov1> kinda forgot: after all the partitioning magic, is there some simple way to verify that zpool I/O is well aligned with hardware sector/page sizes? :-)
[13:55:12] <igork> tsoome: https://paste.dilos.org/?88f424e3a2922d13#Nvy4ssTKusy9aaVa5WJlRvh2lwCzisevYMaoAaZuFbc=
[13:55:16] <igork> panic on sparc
[13:55:30] <igork> with font update
[13:56:13] <tsoome> yep, you need to create rule to build 12x22 font from Gallant19
[13:56:30] <tsoome> unless peter has it done already
[13:56:43] <igork> huh
[13:56:49] <tsoome> as he was already bitten bi it.
[13:56:55] <tsoome> by*
[13:57:07] <igork> can we use different rules for intel and sparc ?
[13:57:31] <igork> i do not want change fonts and console on sparc
[13:58:17] <tsoome> well, 12x22 is for sparc. we use 8x16 for x86 (because of vgs text)
[13:58:20] <tsoome> vga*
[13:58:38] <gitomat> [illumos-gate] 10119 audit(1) gets NULL check wrong -- John Levon <john.levon at joyent dot com>
[13:59:02] <igork> tsoome: can you point me where i can try to update rule for sparc ?
[13:59:22] <tsoome> somewhere in sparc/Makefile* ?
[13:59:58] <tsoome> I can only guess, I have no sparc build environment.
[14:12:57] *** myrkraverk <myrkraverk!~chatzilla@unaffiliated/myrkraverk> has quit IRC (Ping timeout: 245 seconds)
[14:24:02] *** myrkraverk <myrkraverk!~chatzilla@unaffiliated/myrkraverk> has joined #illumos
[14:50:44] *** myrkraverk <myrkraverk!~chatzilla@unaffiliated/myrkraverk> has quit IRC (Ping timeout: 255 seconds)
[14:58:13] *** sarvet <sarvet!~sarvet@dslb-092-073-105-207.092.073.pools.vodafone-ip.de> has joined #illumos
[15:04:28] *** myrkraverk <myrkraverk!~chatzilla@unaffiliated/myrkraverk> has joined #illumos
[16:01:00] *** tomoyat1 <tomoyat1!~tomoyat1@> has quit IRC (Quit: ZNC 1.6.5 - http://znc.in)
[16:01:32] *** tomoyat1 <tomoyat1!~tomoyat1@tomoyat1.com> has joined #illumos
[16:20:31] *** idodeclare <idodeclare!~textual@cpe-76-185-177-63.satx.res.rr.com> has joined #illumos
[16:44:30] * jeffpc tries to figure out how to send & receive a zfs dataset which has corruption
[17:15:40] <LeftWing> jeffpc: There's a magic variable you can whack to "skip over" certain kinds of panics, I think, if that's what you're hitting
[17:16:04] <LeftWing> "zfs_recover" I guess
[17:51:30] <tomww> what about failmode? maybe zfs send would stop at faults even if failmode and /etc/system settings are set to continue if errors occur. but I could think of using rsync and try syncing out most if the data?
[18:08:35] <jeffpc> LeftWing: it's not a panic, it's 'zfs send' erroring out with EIO
[18:09:13] *** rsully <rsully!~rsully@unaffiliated/rsully> has joined #illumos
[18:09:15] <LeftWing> jeffpc: Oh that's no good
[18:09:20] <jeffpc> and yeah, rsync is the ultimate fallback
[18:09:37] <LeftWing> Maybe you can identify the object it's trying to read and inspect it with zdb
[18:10:06] <jeffpc> I know the source of corruption, it was an accidental mkfs a while back
[18:10:20] <jeffpc> and I'm slowly cleaning things up
[18:10:30] <jeffpc> but I want to move to a different disk first
[18:10:50] <jeffpc> it'd be nice if zfs send sent "ECKSUM holes"
[18:10:59] <jeffpc> or even the bad data
[18:11:18] <LeftWing> Well I'm saying maybe you can identify the specific objects it's trying to read that are corrupt and ... doctor up a replacement
[18:11:33] <jeffpc> zpool status lists them
[18:11:43] <jeffpc> but they exist on snapshots
[18:11:49] <jeffpc> so doctoring is too much work
[18:12:01] <LeftWing> Can you destroy the snapshots?
[18:12:33] <jeffpc> in theory, yes
[18:12:46] <jeffpc> in practice, I want to make sure they are actually useless
[18:12:48] <jeffpc> which takes time
[18:12:51] <LeftWing> Aye.
[18:13:02] <jeffpc> hence the "slowly cleaning things up"
[18:13:57] <jeffpc> I've concluded that zfs can't do what I want it to do, and so I'm about to dd the pool over
[18:14:10] <jeffpc> which will preserve all the badness
[18:17:16] <andyf> LeftWing - If you have a chance, I'd appreciate you testing my wsdiff patches with your Jenkins thing. Hopefully fixes the speed issue and the encoding problem you saw with non-C locale. https://illumos.org/rb/r/1554/
[18:17:24] <LeftWing> andyf: I was just looking at it haha
[18:22:43] *** idodeclare <idodeclare!~textual@cpe-76-185-177-63.satx.res.rr.com> has quit IRC (Quit: My MacBook has gone to sleep. ZZZzzz…)
[18:23:16] *** idodeclare <idodeclare!~textual@cpe-76-185-177-63.satx.res.rr.com> has joined #illumos
[18:27:05] *** myrkraverk <myrkraverk!~chatzilla@unaffiliated/myrkraverk> has quit IRC (Read error: Connection reset by peer)
[18:30:08] *** myrkraverk <myrkraverk!~chatzilla@unaffiliated/myrkraverk> has joined #illumos
[18:51:24] *** jimklimov1 <jimklimov1!~jimklimov@> has quit IRC (Quit: Leaving.)
[19:02:38] *** atrius <atrius!~atrius@99-160-139-232.lightspeed.nsvltn.sbcglobal.net> has quit IRC (Quit: ZNC - http://znc.in)
[19:06:34] <LeftWing> andyf: wsdiff was definitely much quicker! 12 minutes instead of ~2 hours
[19:06:53] <tsoome> 2 hours?!
[19:11:24] <LeftWing> tsoome: There was a slight performance regression
[19:11:45] <LeftWing> Python is, as near as I can tell, a system for producing them
[19:12:27] <tsoome> indeed. I can see that in Solaris 11:)
[19:12:28] *** eki <eki!~eki@dsl-hkibng41-567327-143.dhcp.inet.fi> has quit IRC (Quit: leaving)
[19:12:42] <LeftWing> ha
[19:13:20] <tsoome> and still they are rewriting it into python… .P
[19:13:21] <tsoome> :P
[19:23:39] <LeftWing> andyf: https://www.illumos.org/issues/10525
[19:24:10] <LeftWing> I noticed this as well
[19:24:20] <LeftWing> I suspect it might be encoding related?
[19:35:31] *** atrius <atrius!~atrius@2600:1700:22d1:518:5156:ff24:fc96:2f1a> has joined #illumos
[19:42:35] <LeftWing> Also I gave python3.5 a go and I hit https://www.illumos.org/issues/10526
[19:51:13] *** rsully <rsully!~rsully@unaffiliated/rsully> has quit IRC (Quit: rsully)
[19:57:54] <andyf> LeftWing - I can look at those two as well.. at least one of them should be trivial :D
[19:58:03] <LeftWing> :D
[19:58:15] <LeftWing> Thanks, and sorry for piling on so many things
[19:59:04] <andyf> np.. we need to get it back on track
[20:05:27] *** pmooney <pmooney!~pmooney@50-93-251-200.fttp.usinternet.com> has quit IRC (Ping timeout: 245 seconds)
[20:06:28] <andyf> sysconf("SC_NPROCESSORS_ONLN") is 40 on my test system so the bug doesn't show up there - one line fix though
[20:07:42] *** pmooney <pmooney!~pmooney@50-93-251-200.fttp.usinternet.com> has joined #illumos
[20:11:51] <LeftWing> andyf: If you tell me what it is I can hack it in and make sure it works here
[20:12:09] <andyf> + max_threads += int(max_threads/5)
[20:12:12] <LeftWing> Ah ha
[20:12:23] <LeftWing> Does that round or floor/ceil?
[20:13:20] <LeftWing> I guess it floors()
[20:13:43] <andyf> yes, towards zero
[20:13:57] <andyf> "truncates towards zero"
[20:14:17] <LeftWing> Alright that works for me thanks
[20:15:28] <LeftWing> So my rough guess for 10525 is ... if we were somehow combining two bytes of input into one byte of output in one side of the diff (e.g., two not-UTF-8 bytes becomes one "?" say)
[20:15:43] <LeftWing> That there'd then be a pretty serious diff in the output
[20:16:58] <andyf> The whole python2/3 incompatibility thing is painful
[20:17:13] <LeftWing> There are definitely way more paper cuts than I was expectingb
[20:17:18] <LeftWing> I can see why people resisted the switch for so long
[20:17:59] <LeftWing> I guess we'll be turfing out 2.7 some time soon anyway
[20:18:03] <andyf> It looks like mercurial are almost ready to release a beta that works with python3
[20:18:07] <LeftWing> So it'll just be 3 then?
[20:18:08] <andyf> that's the last blocker in OmniOS anyway
[20:18:11] <LeftWing> Fair
[20:18:31] <andyf> I'm not sure about soon.. maybe a couple of years.
[20:18:31] <LeftWing> Wow I didn't realise hg wasn't yet officially supported on 3
[20:18:40] <tsoome> OI insttall scripts need update still:)
[20:18:50] <LeftWing> The caiman consolidation?
[20:18:54] <tsoome> ye
[20:18:57] <LeftWing> Mmm.
[20:19:03] <tsoome> its all written in 2...
[20:19:21] <LeftWing> Time to rewrite it in rust I guess! :D
[20:19:37] <andyf> Dan switched OmniOS to ksh :)
[20:19:45] <LeftWing> I wonder what the rust-on-SPARC situation is
[20:20:07] <tsoome> i think it is stunning:D
[20:30:24] <LeftWing> Allegedly the std part of rust works on sparcv9-sun-solaris at least
[20:31:24] <LeftWing> Well, _builds_ perhaps
[20:31:32] <LeftWing> The magic of llvm I guess.
[20:51:35] *** clapont <clapont!~clapont@unaffiliated/clapont> has joined #illumos
[21:00:30] <jbk> yeah, the issue i think is more about other rust libraries assuming stuff in libc is there when it's not
[21:00:36] <jbk> i've been working on some of that for illumos
[21:00:54] <jbk> (cfmakeraw, cfsetspeed, openpty, etc)
[21:09:44] <andyf> @LeftWing - Yep, it's things with the high-bit set.. working on a patch
[21:14:19] <igork> jbk: i have cfmakeraw on dilos tree for reduce portability issues of userland apps - no issues with it
[21:15:01] <igork> just FYI - it can se safely ported to illumos too, i think
[21:27:28] <andyf> Oh, I think I can speed it up a bit more too
[21:34:38] *** amrfrsh <amrfrsh!~Thunderbi@> has quit IRC (Ping timeout: 245 seconds)
[22:23:12] <LeftWing> igork: You should get it reviewed and integrated!
[22:28:37] <igork> LeftWing: you are welcome with: https://bitbucket.org/dilos/dilos-illumos/commits/8c119cbc80dabec1b079fda750bb5ca67e32dc72
[22:29:08] <igork> i do not know illumos needs and ported it for dilos needs
[22:29:44] <LeftWing> If it's a correct and conforming implementation of cfmakeraw() I daresay it would meet our needs
[22:31:01] <LeftWing> We'd want it in a new libc version, though we'd catch that during review.
[22:32:22] <LeftWing> I look forward to seeing it up for review!
[23:00:54] <richlowe> andyf: how in the world is subprocess managing to be so slow?
[23:01:54] <andyf> richlowe - see https://stackoverflow.com/questions/10888846/python-subprocess-module-much-slower-than-commands-deprecated
[23:02:36] <andyf> Even after fixing all of these problems (which I think I've done now), I'll leave an issue open to address wsdiff performance under python3
[23:02:42] <richlowe> I wouldn't expect even that to stretch so much.
[23:02:49] <richlowe> andyf: is one using a pipe and the other a tempfile or something?
[23:03:24] <richlowe> andyf: https://github.com/jrfonseca/gprof2dot may make your life much better in that regard
[23:05:23] <richlowe> I mean, so would wsdiff having been written properly, too, but...
[23:05:40] <richlowe> pretty annotated callgraphs are easier to achieve. :)
[23:06:33] <andyf> I'll have a look. I have flamegraphs of it running under py2 and py3, but everything just seemed stretched out because so much of wsdiff is running external commands
[23:34:01] *** snuff-work <snuff-work!~snuff-wor@202-161-112-134.tpgi.com.au> has joined #illumos
[23:46:32] *** amrfrsh <amrfrsh!~Thunderbi@> has joined #illumos

   March 10, 2019  
< | 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 | >