|
| |
| Join the live Webchat: Join #esp Join #sole | ||
| <@plexdev> | http://is.gd/6Ixeq by [tav] in ampify/misc/codereview/ -- A basic-but-should-work-fine JSON validator. -- [1 added] -- jsoncheck.py (A) as ~plexdev/297 in #esp · |
| <tav> | http://nampersands.com/?p=417 |
| <@plexdev> | http://is.gd/6IskM by [tav] in 2 subdirs of ampify/ -- Final set of renaming for the nite. -- [4 added, 4 removed] -- AUTHORS (A), LICENSE (A), TODO (A), README.md (A), CREDITS.md (D), LICENSE.md (D), TODO.txt (D), README.rst (D) as ~plexdev/296 in #esp · |
| <@plexdev> | http://is.gd/6IqVN by [tav] in ampify/misc/codereview/ -- Removing cpplint.py as we're not likely to C++. -- [1 removed] -- cpplint.py (D) as ~plexdev/295 in #esp · |
| <@plexdev> | http://is.gd/6IpID by [tav] in 9 subdirs of ampify/ -- Start process of cleaning up weblite. -- [12 added, 1 modified, 20 removed] -- config.py (A), model.py (A), remote.py (A), __init__.py (A), site.mako (A), site.py (A), weblite.py (A), crypto.py (A), paypalx.py (A), test_paypalx.py (A), cookie.py (A), validation.py (A), app.yaml (U), __init__.py (D), __init__.py (D), config.py (D), exception.py (D), model.py (D) ... as ~plexdev/294 in #esp · |
| <ipanuscan> | pretty diagram as ~ipanuscan/376 in #esp · |
| <ipanuscan> | A successful Git branching model as ~ipanuscan/375 in #esp · |
| <@tav> | http://nvie.com/archives/323 |
| <@tav> | social microcollaboration ? ;p |
| <@plexdev> | http://is.gd/6D3hf by [tav] in ampify/ -- Updated .codereview.cfg to the new settings. -- [1 modified] -- .codereview.cfg (U) as ~plexdev/293 in #esp · |
| <@plexdev> | http://is.gd/6CWpC by [tav] in 2 subdirs of ampify/ -- Moved jslint into codereview and updated nodelint. -- [1 added, 1 modified, 1 removed] -- jslint.js (A), nodelint.js (U), jslint.js (D) as ~plexdev/291 in #esp · |
| <@tav> | nn |
| <@tav> | my ab-based tests of node vs. python (even using libev) put node serving more req/s with lesser cpu load, but using more memory |
| <sbp> | I'm off. 'night! |
| <@tav> | end up reverting to good old ab |
| <sbp> | yeah, looks like it is |
| <@tav> | i've looked at tsung but always been too afraid to use it — it looked like a beast |
| <sbp> | http://gist.github.com/281449 |
| <@tav> | hmz |
| <sbp> | it's just an updated version of the counter script |
| <sbp> | uses both |
| <@tav> | node or redis ? |
| <@tav> | which load exactly is the test meant to verify? |
| <sbp> | yeah |
| <@tav> | tsung ? |
| <sbp> | he's using an http benchmarking suite |
| <sbp> | 60021 requests total |
| <@tav> | he's running multiple clients ? |
| <sbp> | server load is 0.9 |
| <sbp> | he's getting 1000 r/s |
| <@tav> | probably shouldn't focus on that, but figure once it's done, it's mainly done and provides a good foundation |
| <@tav> | my todo list is currently focussed on setting up review/build/test infrastructure btw |
| <sbp> | yeah :-) |
| <@tav> | (from phenny Makefile) |
| <@tav> | oooh, you do use git graph! =) |
| <@tav> | aha |
| <sbp> | nslater is just about to try to kill node.js + redis |
| <sbp> | :-) |
| <@tav> | sbp: i see you're making good use of gist |
| <sbp> | sup |
| <@tav> | hmzie |
| <sbp> | :-) |
| <@tav> | soz |
| <sbp> | L94, catch up |
| <@tav> | didn't i post that yesterday's gist? |
| <@tav> | client.close() |
| <@tav> | ehm |
| <sbp> | excellent |
| <sbp> | and yet numbers have gone up |
| <sbp> | 19 Jan 13:40:05 . 0 clients connected (0 slaves), 619150 bytes in use, 0 shared objects |
| <sbp> | http://github.com/fictorial/redis-node-client/blob/master/redisclient.js#L94 |
| <sbp> | oh wait, I'll bet I need to add an extra client disconnect call |
| <sbp> | when I close the tab, the number doesn't decrease |
| <sbp> | *connected |
| <sbp> | for some reason the clients stay connection. persistent connections over HTTP/1.1? |
| <sbp> | 19 Jan 13:37:04 . 33 clients connected (0 slaves), 628687 bytes in use, 0 shared objects |
| <sbp> | one problem is this: |
| <sbp> | hmm |
| <sbp> | don't go mad |
| <sbp> | damn people, it's up to 19 |
| <sbp> | heh, heh |
| <@tav> | 8, 9, 10! |
| <sbp> | check it out before I send a 2 signal to the process |
| <sbp> | that's my counter thing, running live |
| <sbp> | so yeah, http://helios.crschmidt.net:7210/ |
| <@tav> | hmz |
| <sbp> | 2 is INT |
| <sbp> | -- |
| <sbp> | ~/bin/node counter.js |
| <sbp> | ~/bin/redis-server & |
| <sbp> | trap cleanup 2 |
| <sbp> | } |
| <sbp> | exit |
| <sbp> | echo Quitting |
| <sbp> | kill -9 %1 |
| <sbp> | function cleanup() { |
| <sbp> | -- |
| <sbp> | okay, this works: |
| <sbp> | chuckle |
| <archels> | your bash ninja rank is: Beverly Hills as ~archels/54 in #esp · |
| <sbp> | meh |
| <sbp> | this is in a script, and ^C is terminating the whole script |
| <sbp> | oh, I see |
| <sbp> | but it ain't working |
| <sbp> | %1 should be the first job |
| <archels> | %1 replaces to abc? as ~archels/53 in #esp · |
| <sbp> | why doesn't that work? |
| <sbp> | well, I was trying $ kill -9 %1 |
| <archels> | piped if you're feeling adventurous? as ~archels/52 in #esp · |
| <archels> | ps, find pid, kill pid? as ~archels/51 in #esp · |
| <sbp> | what is the best method to kill the backgrounded abc once pqr exits? |
| <sbp> | $ pqr |
| <sbp> | $ abc & |
| <sbp> | bash question |
| <archels> | It's not like they were missing an essential tool to go beyond one or two decimals. as ~archels/50 in #esp · |
| <archels> | Hm, curious that decimal approximation of pi took so long. as ~archels/49 in #esp · |
| <@tav> | what you say? |
| <oierw> | in #esp · |
| * | silenius prints it for reading on his way home as ~silenius/704 in #esp · |
| <silenius> | "Conversion to base 10: 12 days" ^^ as ~silenius/703 in #esp · |
| <silenius> | great paper :) as ~silenius/702 in #esp · |
| <@tav> | spotify:track:6FvtSHS3hzjS8itHJPRzoJ |
| <@tav> | quad-core i7 cpu, 6 gigs of ram and 5 x 1.5 TB drives |
| <@tav> | that's quite impressive — especially given the machine specs |
| <@tav> | 2700 billion! |
| <sbp> | http://bellard.org/pi/pi2700e9/pipcrecord.pdf |
| <sbp> | Fabrice Bellard! |
| <sbp> | Fabrice Bellard holds the current record for computing decimal places of pi! |
| <sbp> | wow! |
| <@phanny> | 62 as ~phanny/224 in #esp · |
| <sbp> | .o py len('71828182845904523536028747135266249775724709369995957496696763') |
| <@tav> | Phi = 1.61803398874989484820458683436563811772030917980576286213544862 // A001622 |
| <@tav> | Pi = 3.14159265358979323846264338327950288419716939937510582097494459 // A000796 |
| <@tav> | E = 2.71828182845904523536028747135266249775724709369995957496696763 // A001113 |
| <@tav> | go defines a math.Phi constant |
| <@tav> | you backoff by a factor of 2? well, we backoff by a factor of phi! |
| <@tav> | see if that makes things better! |
| <@tav> | aside: we should have a golden ratio-based backoff algorithm ;p |
| <silenius> | that's what the forward claims and I agree as ~silenius/701 in #esp · |
| <silenius> | re "suited for teaching mathematics" - yes and no. not much material to use 1:1 in a course, but lots of background-information that should enable teachers to teach math in a better way as ~silenius/700 in #esp · |
| <@tav> | that'd put him at 4 decimal places around 100 AD? |
| <@tav> | "Request to over zealous 'deleters' not to delete." <--- from summary of Timeline image |
| <sbp> | hmm, Ptolemy isn't included though |
| <sbp> | also nice: http://en.wikipedia.org/wiki/File:Timeline_of_approximations_for_pi.jpg |
| <@tav> | nice |
| <sbp> | “It's probably bad form to write a review of a review, but I was a little chapped at reading a previous reader's complaint that this book wasn't well suited for teaching mathematics. That's sort of like complaining that your microwave oven doesn't also play DVD's” |
| <sbp> | this is great: |
| <@plexdev> | http://is.gd/6BCGd by [tav] in 2 subdirs of ampify/ -- A bunch of codereview-related changes. -- [4 added, 2 modified, 3 removed] -- .review.py (A), breakpad.py (A), review.py (A), review_builtins.py (A), git-review (U), upload.py (U), .presubmit.py (D), presubmit.py (D), presubmit_builtins.py (D) as ~plexdev/290 in #esp · |
| <sbp> | hmm, thanks |
| <silenius> | by the way - my math+history recommendation, for those interested in errors committed a long time ago and lessons learned from them: http://www.amazon.com/History-Mathematics-Introduction-2nd/dp/0321016181/ref=sr_1_11?ie=UTF8&s=books&qid=1263922297&sr=8-11 as ~silenius/699 in #esp · |
| <sbp> | heheh |
| <@tav> | hrhr |
| <silenius> | must have been the Babylonians, then as ~silenius/698 in #esp · |
| <@tav> | to be fair, that probably was written in ancient times |
| <sbp> | see, for example, javascript:alert(Math.PI) |
| <sbp> | pretty much sums it up |
| <sbp> | ]]] — http://jsdoc.sourceforge.net/ |
| <sbp> | in #esp · |
| <sbp> | */ |
| <sbp> | Circle.PI = 3.14; |
| <sbp> | * A very rough value for pi |
| <sbp> | /** |
| <sbp> | [[[ |
| <sbp> | yeah, that's the spirit |
| <silenius> | at least i think so (what else do we need users for?) as ~silenius/697 in #esp · |
| <@tav> | heh |
| <silenius> | ack@sbp - that's what users are for? as ~silenius/696 in #esp · |
| <@plexdev> | http://is.gd/6BzKx by [tav] in 2 subdirs of ampify/ -- Adding a pyutil package with format_traceback module. -- [1 added, 1 removed] -- format_traceback.py (A), format_traceback.py (D) as ~plexdev/289 in #esp · |
| <@tav> | heh |
| <sbp> | oh well, who needs tests |
| <@tav> | i never finished with doctest — got distracted by deployment + the general continuous integration blah |
| <@tav> | whoever thought it helped with documentation needs to have their knee capped |
| <sbp> | yes |
| <@tav> | i fucking hate javadoc and it's various bastard children |
| <sbp> | no, quite |
| <@tav> | please to be not using any of those pls, k, thx! |
| <sbp> | well, same thing basically |
| <sbp> | actually, didn't you make a doctest module? |
| <sbp> | understand why you rolled your own now, heh |
| * | sbp peers at http://www.straw-dogs.co.uk/12/22/5-ways-documenting-javascript/ |
| <sbp> | :-) |
| <@tav> | ah, yes, hehe, that got a good laugh at the riva |
| <sbp> | and ah, bummer |
| <sbp> | Abyssinian Moon |
| <@tav> | so even though my gandi identifier starts with TE, tav espian, it has my offical name |
| <@tav> | thus i had to change my Gandi full name |
| <@tav> | unfortunately, the Italian registrar requires you to use your official passport name |
| <@tav> | sbp: A.M. ? |
| <sbp> | which I just visited for the first time via http://amp.espians.com/ |
| <sbp> | (I ask because I'm surprised to find you used your full birthname to register ampify.it) |
| <sbp> | tav: by the way, did you notice who I sent A.M. to? |
| <@tav> | meaning that the capabilities should be dependent on a different layer from that |
| <@tav> | openid, with all its flaws, still (sort of) suffices as an authentication mechanism |
| <@tav> | "my argument is that, the capabilities of a identified user (however poorly) should be separate to the authentication process" |
| <silenius> | tav: but openid is not a "trust network" (though it requires lots of trust when you consider using it..) as ~silenius/695 in #esp · |
| <@tav> | forgive my ranting |
| <@tav> | i.e. subjective evaluations and verifications as opposed to stupid-arsed "objective" ones |
| <@tav> | the nice thing about trust networks is that i could've been trusted as an "adult" when i was say 12 as opposed to having to wait till 18 or whatever |
| <@tav> | but i'm sure there are other solutions too |
| <@tav> | of course, i'm enamoured with trust networks in this regard |
| <@tav> | ... to the authentication process |
| <@tav> | so my argument is that, the capabilities of a identified user (however poorly) should be separate |
| <@tav> | in the end, even the most rigorous super-verified-identity-provider will be gamed |
| <silenius> | ignore my last statement as ~silenius/694 in #esp · |
| <@tav> | no, i'm saying how only trusting a few providers has historically always been a stupid idea |
| <silenius> | ah. so it's NOT my provider after all, since I need a verification from somebody else? as ~silenius/693 in #esp · |
| <@tav> | in theory that kind of works — but in practise, we find that they give out certs to fraudsters claiming to be whatever |
| <@tav> | that of CAs for SSL certs — with verisign and friends being exclusively trusted |
| <silenius> | tav: you require that my provider provides certain SLAs. Since I run my own provider, I just say: "sure!" as ~silenius/692 in #esp · |
| <@tav> | e.g. let's take the biggest organised MITM attack ever |
| <@tav> | but, let's take a lesson from history on that one? |
| <@tav> | for example, you might require that a provider provide certain SLAs about their validation process |
| <silenius> | and if everybody runs his own provider, I will soon start to only accept a few trustworthy providers that have some measures against spammers abusing their service as ~silenius/691 in #esp · |
| <@tav> | well, that really is a separate issue |
| <@tav> | and it's definitely verifying that |
| <silenius> | so after 3 spammers used it, I will block shitprovider.com as ~silenius/690 in #esp · |
| <@tav> | all the user is claiming is to be user1797219731 |
| <silenius> | it just verified that a user-record in the spammer's provider exist as ~silenius/689 in #esp · |
| <@tav> | hmz, how not ? |
| <silenius> | tav: yes, it's a valid authentication, but no, it has not verified the identity of the user as ~silenius/688 in #esp · |
