zenspider

Activity Graph

Page 1 of 47 | Next »

2018-11-14

[22:21:01] zenspider: they're ALSO being recorded and will be on confreaks
[22:21:24] zenspider: pavelz: did you figure it out?
[22:23:31] zenspider: anyone around who is willing to test this out for me?
[22:23:43] zenspider: curl -L https://tinyurl.com/graphics-setup | bash
[22:23:56] zenspider: obviously... remove "| bash" and read it first

2018-11-12

[09:54:47] zenspider: arfnir: unexpected colon? is this really old ruby code?
[09:55:24] zenspider: it probably tells you what file/line it was near when it had the problem
[09:56:16] zenspider: that sounds like < 1.9 syntax tho... eg using ":" instead of "then" in case/when/then

2018-11-01

[20:35:26] zenspider: eam: wow. I used kerberos once... in 1997. (really) I assume it hasn't gotten easier.

2018-10-31

[21:01:03] zenspider: has joined #ruby

2018-10-29

[20:58:19] zenspider: "modular programming" is a term from the 70s.
[20:58:29] zenspider: ala dikstra and folk
[20:59:06] zenspider: ruby Modules have (roughly) nothing to do with "Modular programming"
[21:01:21] zenspider: Inside: what is it?
[21:02:16] zenspider: Inside: I meant what code are you spinning on?
[21:51:15] zenspider: Radar: I took it to be confident ruby... amazon seems to agree
[22:27:27] zenspider: YAY! I think I'm actually done with the graphics gem changes for my talk!
[22:37:07] zenspider: baweaver: use that as your default from now on... I haven't seen a 4x3 conf in a very long time. And keynote (probably others) will automatically adjust as best they can.
[22:37:35] zenspider: Eiam: faster. not easier. and less about cheaper as it is about profit
[22:38:38] zenspider: luckily some publishers are more savvy than others... prag for example
[22:39:06] zenspider: totally... going bigger is always easier than trying to go smaller. I did that once and it sucked
[22:39:26] zenspider: tho... maybe that should be a signal that I already had too much on a slide?? I'll have to think about that
[22:41:07] zenspider: Radar: I generally agree. I'm a huge fan of Tahakashi method
[22:41:50] zenspider: Pragprog, learn enough (hartl) are both really streamlined

2018-10-24

[21:44:58] zenspider: doesn't mean it's a good idea...
[22:48:42] zenspider: juria_roberts: not sure your words/question make sense...
[22:49:00] zenspider: phaul: I tend to agree
[22:49:48] zenspider: I try as much as possible to stick to "do the simplest thing that can possibly work" and that pushes back heavy on all of the dry-* packages (last I looked)
[22:51:00] zenspider: phaul: calibrate accordingly... I'm an edge case. :P
[23:29:59] zenspider: juria_roberts: ...
[23:31:56] zenspider: havenwood: wget -O ruby-install-0.6.0.tar.gz https://github.com/postmodern/ruby-install/archive/v0.6.0.tar.gz ?!?
[23:32:19] zenspider: a tarball? oh... because bootstrap issue if it were a gem?
[23:32:37] zenspider: still... why not a git checkout?
[23:33:53] zenspider: man... I'm trying to profile and optimize this code and while the profile is pointing me at things worth working on and my microbenchmarks are definitely showing improvements... the actual improvement on the top level benchmark is ... zero

2018-10-17

[03:04:10] zenspider: zxq2: that's not your problem. that's your proposed solution.

2018-10-09

[09:33:05] zenspider: Don't you love it when someone asks a question with seemingly no desire for an answer?

2018-10-08

[19:35:41] zenspider: has joined #ruby-offtopic
[19:47:59] zenspider: hrm... anyone know off the top of their head if I can set an ivar on a RTypedData ?
[19:48:22] zenspider: I now need to return 2 things where before there was only 1
[19:51:38] zenspider: mroutis: yes... it is ruby related.
[19:51:45] zenspider: ivar == instance variable
[19:51:56] zenspider: RTypedData is an internal struct
[20:00:25] zenspider: hrm... this might not be the right way to do this anyways... I dunno
[20:13:21] zenspider: A wrapper around SDL2. My code currently wraps SDL1 and the changes between them are a bit much
[21:29:46] zenspider: holy shit... I don't just have pixels... it's actually sorta kinda working.
[21:30:01] zenspider: and... it went from 30-60 fps to ~260fps ?!?!?!
[21:30:14] zenspider: that has to be wrong
[21:31:24] zenspider: cout: hey hey! how goes?
[21:36:01] zenspider: I think there's a pretty big difference between "the stove is hot" and "C++ is a fine language"
[21:41:45] zenspider: just hadn't done it. The RTypedData is a stand alone struct... has an internal RObject but I wasn't sure if it was gonna behave differently
[21:41:51] zenspider: turns out it seems fine. I think.
[21:42:03] zenspider: I'm juggling too many balls atm tho... who knows?
[21:43:58] zenspider: appears that way
[21:44:37] zenspider: my window sizes are entirely wonky. There must be some scaling thing I am unaware of and all the numbers are wrong... but my demos are almost all running. it's crazy
[21:47:16] zenspider: well... considering I'm doing a complete rewrite about a month before my talk...
[21:48:47] zenspider: ah shit. I just introduced a segfault somewhere... huh
[21:49:00] zenspider: I'm giving a talk on my graphics gem.
[21:51:58] zenspider: I was there. :P
[22:01:43] zenspider: thanks. dunno about the rake newb thing. that sounds odd but fairly low priority atm.
[22:01:55] zenspider: hopefully after this revamp not everything will be broken

2018-10-06

[02:54:29] zenspider: !connection schneider
[20:39:18] zenspider: ogotai: you might prefer split

2018-10-05

[11:23:38] zenspider: working on a emacs mode to color methods based on flog scores... I'm undecided on how to do the score -> shading at this point. Suggestions?
[11:24:12] zenspider: (right now it goes from white through pink all the way up to full red linearly
[21:25:05] zenspider: havenwood: maaaannn... I love rinda
[21:25:09] zenspider: I really need more reasons to use it
[21:26:46] zenspider: built on top of it, yeah
[21:26:55] zenspider: apeiros: bot still brokey?
[21:27:47] zenspider: that said... I've probably carried this ssh key a tad too far.
[21:29:34] zenspider: if the image is signed, I believe
[21:29:56] zenspider: oh. you mean there is no "skip" button? that seems new
[21:30:10] zenspider: I'm holding off on mojave for a while... until at least .1
[21:35:46] zenspider: haha. oh no

2018-10-04

[09:16:19] zenspider: Pupeno: it's 2am pacific time... so...
[09:31:40] zenspider: thebigj: have you at least googled it?
[10:38:14] zenspider: za1b1tsu: similar to earlier today: if you haven't measured, you don't care.
[21:27:40] zenspider: za1b1tsu: why are you reinventing the wheel?
[21:29:03] zenspider: ah. ok. you're not. You're just abstracting unnecessarily.
[21:29:51] zenspider: here's one of mine: https://github.com/seattlerb/flog/blob/master/lib/flog_cli.rb#L69-L154
[21:30:24] zenspider: here's one that's more complicated (so plugins can add options)
[21:30:26] zenspider: https://github.com/seattlerb/minitest/blob/master/lib/minitest.rb#L169-L233
[21:31:00] zenspider: but they're both fairly straightforward applications of OptionParser w/o any unnecessary abstractions.
[21:33:44] zenspider: DTSTTCPW... mostly. YAGNI a bit too.
[21:34:02] zenspider: do the simplest thing that can possibly work.
[21:34:08] zenspider: and you ain't gonna needit
[21:34:19] zenspider: aka: stop overthinking and complicating things. :P
[22:46:52] zenspider: it "should" be: gem install rails; rails new myproject
[22:47:04] zenspider: where "should" is some function of linux dogma changing things

2018-10-03

[14:04:21] zenspider: Ping timeout: 260 seconds
[14:06:00] zenspider: has joined #ruby
[19:50:41] zenspider: glad you had this conversation...
[19:52:18] zenspider: I'm not so sure about that part...
[19:53:03] zenspider: to clarify, File.open and File.exist? use whatever you give them, whether that's releative or absolute, they don't care. File.expand_path, by definition, results with an absolute path. So, you need to either use it on both or neither to have an apples-to-apples comparison
[19:53:30] zenspider: don't push it...
[19:54:18] zenspider: nope. it just uses whatever string you hand it
[19:55:38] zenspider: File.open opens a path
[19:55:42] zenspider: it doesn't parse anything
[19:56:57] zenspider: baweaver: doesn't matter. File.open and File.exist? don't parse anything
[19:57:04] zenspider: they just make system calls to do the thing
[19:57:15] zenspider: path is a path
[19:57:36] zenspider: and it is NOT consistent with require...
[19:58:46] zenspider: and the answer is and has been: it doesn't matter. it's a path
[19:59:04] zenspider: I get it just fine
[19:59:21] zenspider: you don't get the ramifications of what I'm saying
[19:59:29] zenspider: I've answered that. three times
[19:59:41] zenspider: you're inability to listen to my answer is beside the point
[20:00:33] zenspider: no no and no
[20:01:23] zenspider: I could say the same to you. My inability to distinguish you from a basic troll right now isn't helping your case to get help
[20:02:14] zenspider: !kick cthu|
[20:03:31] zenspider: nope. I actually wasn't.
[20:06:23] zenspider: read up on require
[20:06:25] zenspider: it's very clear
[20:07:16] zenspider: well... it was... looks like rubygems rewrote it?
[20:07:41] zenspider: if you do a relative path for require, you need to supply the extension as well
[20:11:57] zenspider: heh. where "modern" is the last... what... 10+ years?
[20:12:33] zenspider: 1.8 had "."... 1.9 did not iirc
[20:12:47] zenspider: broke some of my stuff. 1.9 was a disaster in so many ways
[20:13:06] zenspider: tho I wish they'd still do the shim thing for upcoming versions
[20:14:25] zenspider: za1b1tsu: what are your goals?
[20:15:14] zenspider: za1b1tsu: use real / ordinary objects wherever possible.
[20:15:36] zenspider: I wouldn't recommend struct for normal use. They have too many gotchas
[20:15:51] zenspider: I use struct for quick prototyping... and avoid it otherwise
[20:16:39] zenspider: za1b1tsu: to what end? what are you trying to avoid?
[20:18:25] zenspider: it doesn't matter if your object subclasses BasicObject or Object. They'll hold onto whatever you make them hold onto. That's not bloat, that's need. If you don't make them hold onto superfluous stuff, and nothing extra holds onto them (like closures), they'll be fine
[20:18:48] zenspider: stick to the simplest thing that could possibly work, until you have evidence/measurements that say you should do otherwise
[20:19:46] zenspider: za1b1tsu: one way to look is to try something like this: /usr/bin/time -l ruby -e 't0 = Time.now; 10_000_000.times do BasicObject.new end; p Time.now - t0'
[20:20:22] zenspider: on my laptop, on battery, I take just about 1 second to produce 10m objects whether it is BasicObject or Object. The maximum resident set size doesn't change either.
[20:22:17] zenspider: hash literals need to be copied on every instantiation, but they shouldn't be that slow
[20:23:04] zenspider: doing the above cmdline replacing Object.new with {} results in it being 60% faster
[20:23:27] zenspider: but adding stuff to the hash? makes it not an apples-to-apples comparison. It slows down to ~3.7 seconds on my side
[20:24:57] zenspider: za1b1tsu: Hash.new is a const lookup + method call... takes 4.19 seconds on my side. {} is a static hash in the bytecode + dup
[20:25:19] zenspider: Hash.new will be closer to reality as you'll make some class of your own and instantiate on it
[20:26:03] zenspider: havenwood: and have nasty gotchas like to_a
[20:26:16] zenspider: really really really dislike that. it's bit me on many occasions
[20:26:46] zenspider: I want a superclass BasicStruct that is pretty much just initialize + attr_accessor
[20:26:54] zenspider: I'd use the fuck out of that
[20:27:55] zenspider: it can... but it won't have the C speeds I want :P
[20:28:13] zenspider: well... and no. I don't want it off of BasicObject... that's TOO basic
[20:28:29] zenspider: BasicObject is for when you want something that's not really object-like... usually proxy objects and the like
[20:52:23] zenspider: Eiam: oooonly for rapid prototyping. And I will use Struct 99% of the time for that. OpenStruct is overly magical
[21:12:43] zenspider: Eiam: Some certainly do. And in general, their code is bloated and slow.

2018-09-14

[21:17:24] zenspider: leitz: what did you figure out?
[21:32:25] zenspider: for non-predicate version, this seems huge
[21:32:28] zenspider: >> a = [(1..5),(1..4),(1..3)].map(&:to_a); a.first.product(*a[1..-1])

2018-09-07

[08:52:33] zenspider: man... color math is just grotesque...
[08:55:24] zenspider: rlawrence: still there?
[09:02:56] zenspider: apropos of nothing, check out a cleaner way to format your templates: https://dpaste.de/AZeG
[09:07:39] zenspider: so your code is getting up to the first null terminated string in the struct and then bombing out?
[09:11:37] zenspider: I don't have access to the structure or header you're grabbing... it might be nice to isolate whether this is a problem with the pack/unpack side of things or the actual call to ioctl.
[09:12:08] zenspider: I found https://wisesciencewise.wordpress.com/2017/07/06/c-program-to-send-identify-controller-command-to-an-nvme-storage-drive/ which has some equivalent C code... maybe run that or the python, grab the blob in question, and just try to unpack that?
[09:12:44] zenspider: it _could_ be that you found a bug... but that's pretty unlikely. that said, if there is going to be a dangling free, it's probably in someplace obscure like ioctl
[09:15:28] zenspider: this doco is so strange: https://nvmexpress.org/wp-content/uploads/NVM_Express_1_2_Gold_20141209.pdf the byte ranges are literally written backwards
[09:17:38] zenspider: rlawrence: yeah. I'd use the python to do the same action, save it off into a binary file, then use ruby to read it in and test your template against. assuming that works fine, then try to isolate the ioctl stuff more
[09:18:01] zenspider: mongodb?? really? for your thing?
[09:20:19] zenspider: !connection alem0lars
[09:20:35] zenspider: rlawrence: good luck. sorry I couldn't be more help
[09:21:02] zenspider: I do find that removing noise and lining stuff up so the human brain can pattern match more does help a lot
[09:22:13] zenspider: leitz: look at line 3
[09:23:30] zenspider: compare it to line 16, which works...
[09:27:22] zenspider: what is the arg on line 16?

2018-09-05

[00:00:14] zenspider: leitz: probably don't need a full rebuild? at least, it looks like you're already using packages...
[00:28:06] zenspider: leitz: why are you compiling ruby?
[00:29:56] zenspider: seems a bit extreme
[00:31:19] zenspider: "couldn't"? seems like you can.
[00:31:34] zenspider: ya know what? nevermind... I don't think any understanding is going to come of this

2018-09-04

[06:31:51] zenspider: Read error: Connection reset by peer
[19:46:04] zenspider: has joined #ruby
[19:51:07] zenspider: meh. that won't affect that much tbh
[19:52:40] zenspider: that ball's been rollin' for a while now
[19:54:36] zenspider: Target version changed from 2.2.0 to Next Major
[19:54:41] zenspider: that was 1 year ago
[20:31:34] zenspider: has joined #ruby
[23:06:00] zenspider: has joined #ruby
[23:07:18] zenspider: Client Quit
[23:07:28] zenspider: has joined #ruby
[23:19:50] zenspider: the repl has been off and on for a while
[23:20:00] zenspider: might be down for now

2018-08-31

[07:21:25] zenspider: *cough* webrick *cough*
[07:22:19] zenspider: then you have two problems
[07:26:17] zenspider: node? now you have ALL the problems :P
[09:39:42] zenspider: short answer: yes
[09:39:55] zenspider: nowhere in the doco does it say anything about "()" in the syntax
[09:43:41] zenspider: looks like you just need to apply the looping yourself. *shrug*

2018-08-29

[20:29:58] zenspider: man... trying to get spring working on this client's app... it CHEWS through battery like it's nothing
[20:30:10] zenspider: ruby process taking almost a whole CPU
[20:30:42] zenspider: I can't even configure it to fall back to polling sucessfully.. ugh. ruby tooling has just gotten FAR too complex and buggy
[20:32:46] zenspider: pretty much
[20:33:44] zenspider: I'm really only using it to try to drive a test to generate a PDF to see WTF asciidoctor-pdf is fucking up now. It's entirely convoluted but I don't need some BS process killing my battery in the process
[20:33:54] zenspider: it's pretty terrible IMHO
[20:50:56] zenspider: anyone have experience with asciidoc + pdf?

2018-08-20

[19:58:08] zenspider: Ping timeout: 260 seconds

2018-08-17

2018-08-16

[00:20:53] zenspider: fuck me. sorted the data before I generated the graph and now it isn't crashing. totally non-random data driven crashes. gotta love it

2018-08-15

[21:22:02] zenspider: holy shit. I figured it out.
[21:22:04] zenspider: >> a = [-75, 78, 103, -51, 31, -56, 80, -22]; a.pack("C*").unpack("Q>*")
[21:22:20] zenspider: what?!? screw you bot
[21:22:31] zenspider: => [13064493699747565802]
[21:22:59] zenspider: last * is not needed either
[21:23:29] zenspider: translation: big endian unsigned long long
[21:24:01] zenspider: iirc it's arbitrary... it happens to fit in this case
[21:25:10] zenspider: I don't think this code will work with all input... but I don't think there IS a way to make it work with all input. bigint is not explicitly supported since it is internal to ruby
[21:25:22] zenspider: unsigned long long
[21:25:54] zenspider: yeah. I think if you did "Q>*" you can then math them up with a fold
[21:26:30] zenspider: bot... I will murder you
[21:28:01] zenspider: whatever happened to _why's sandbox code? IIRC it was too messy to make it back into ruby
[21:29:22] zenspider: >> a = [-75, 78, 103, -51, 31, -56, 80, -22]; a.pack("C*").unpack("N*").inject { |n,m| (n << 32) + m }
[21:29:32] zenspider: WHY DID THAT WORK?!?
[21:29:42] zenspider: anyhow. there's a portable way to do bigint
[21:30:49] zenspider: Sandbox.disable :disk, :network # would be nice
[22:48:05] zenspider: is there some way to build / install graphviz on OSX that doesn't segfault like mad??
[22:48:18] zenspider: seems data dependent for sure... but totally innocuous stuff
[23:02:50] zenspider: SeepingN: that addresses how to install. it doesn't address segfaults at all. the homebrew build is pretty terrible
[23:07:44] zenspider: it's disheartening that graphviz has ~1000 issues open
[23:08:23] zenspider: it's such bitrot. and my favorite competitor to it, which had much better (but slower) layout was killed off by graphviz popularity
[23:09:04] zenspider: I have a LOT of scripts that generate graphs tho... this sucks
[23:19:28] zenspider: fuck that. no.
[23:19:44] zenspider: HELL no. I don't have a PhD in this shit.
[23:22:48] zenspider: doesn't even remotely do the same thing. I have some old code for that too somewhere...
[23:26:10] zenspider: hrm. I don't know gruff. url?
[23:27:52] zenspider: wait. you mean ... yeah. no. that's also doing charting, not graphing.
[23:28:27] zenspider: I'm really surprised to see it has semi-recent commits. that thing is ancient. very early seattle.rb days
[23:30:14] zenspider: ugh. everything either conflates graphs with charts or is java or D3