| [00:09:32] |
* |
highwaychile has joined #pocoo |
| [00:20:53] |
* |
izibi has quit IRC |
| [00:29:37] |
* |
maix has quit IRC |
| [01:19:43] |
* |
highwaychile has quit IRC |
| [03:49:38] |
* |
Kaelten has joined #pocoo |
| [04:08:09] |
<prencher> |
Kaelten! you smell! |
| [04:08:19] |
<prencher> |
grumpy - ping |
| [06:43:18] |
<Kaelten> |
bah |
| [06:43:26] |
<Kaelten> |
wth to do you want prencher |
| [06:52:58] |
<prencher> |
wth do YOU want |
| [07:00:32] |
* |
davidcramer has quit IRC |
| [07:01:09] |
* |
davidcramer has joined #pocoo |
| [08:22:33] |
<asmodai> |
morning! |
| [08:30:54] |
* |
aconbere has joined #pocoo |
| [08:52:15] |
<aat> |
hi! |
| [09:07:50] |
* |
leche has joined #pocoo |
| [09:18:59] |
* |
davidcramer has joined #pocoo |
| [09:48:38] |
<saptah> |
morning! |
| [10:04:36] |
<asmodai> |
hey aat |
| [10:24:00] |
* |
leche has quit IRC |
| [10:25:15] |
<Leonidas> |
haha |
| [10:26:21] |
<Leonidas> |
asmodai: that muntenhuis is funny, they answer mails in dutch even when I say in the first sentence that I don't speak the language. |
| [10:31:51] |
<asmodai> |
jeez |
| [10:32:03] |
<asmodai> |
you want to get your hands on one of those coins? |
| [10:32:35] |
<asmodai> |
And weird thing is, English is required teaching in NL for at least 6 years |
| [10:38:24] |
<mitsuhiko> |
Leonidas: automated mail? |
| [10:40:20] |
<Leonidas> |
mitsuhiko: don't think so, I forwarded you the mail. |
| [10:41:30] |
<Leonidas> |
asmodai: now, provided "Hacking At Random" will be in NL maybe I can let them deliver to my tent :) |
| [10:42:25] |
<Leonidas> |
asmodai: in the muntenhuis it looks a bit like corporate policy, you cannot find an english word on their webseite ;) |
| [10:42:28] |
<mitsuhiko> |
Leonidas: could be automated yep |
| [10:42:38] |
<mitsuhiko> |
have you ever sent a message to microsoft.com? |
| [10:42:58] |
<mitsuhiko> |
but it's a little bit to little text for automated replies |
| [10:43:21] |
<Leonidas> |
mitsuhiko: no, I fear messages to/from microsoft could steal my soul, just like with indians and photographs %-) |
| [10:44:38] |
* |
izibi has joined #pocoo |
| [10:45:00] |
<Leonidas> |
mitsuhiko: """Wij leveren niet in het buitenland""" <- they need to parse the """deliver to germany""" part |
| [10:45:35] |
<Leonidas> |
mitsuhiko: """ook niet als er extra verzendkosten betaald worden.""" <- they need to parse the """how much would it cost""" part |
| [10:45:46] |
<mitsuhiko> |
Leonidas: doesn't that just mean "ins ausland"? |
| [10:46:32] |
<mitsuhiko> |
"wir liefern nicht ins ausland, selbst wenn extra versandkosten bezahlt worden sind" |
| [10:46:38] |
<mitsuhiko> |
that's what i think it says |
| [10:46:43] |
* |
leche has joined #pocoo |
| [10:46:51] |
<Leonidas> |
mitsuhiko: yeah, thats what google says, too |
| [10:47:04] |
<Leonidas> |
(google's translations are actually pretty decent) |
| [10:47:16] |
<mitsuhiko> |
they have dutch now? |
| [10:47:59] |
<Leonidas> |
mitsuhiko: yeah. I didn't know that before I tried. |
| [11:01:53] |
<asmodai> |
wij leevren niet in het buitenland = we don't deliver to foreign countries |
| [11:01:55] |
* |
davidcramer has quit IRC |
| [11:02:20] |
<asmodai> |
ook niet als er extra verzendkosten worden betaald = also not when extra shipping is paid |
| [11:02:51] |
<asmodai> |
I can probably always abuse my neighbours addresses |
| [11:03:08] |
<asmodai> |
they're not really into that sort of thing |
| [11:07:35] |
<Leonidas> |
that'd be very nice :) |
| [11:08:08] |
<Leonidas> |
it is a bit like mitsuhiko said yesterday. |
| [11:08:57] |
<birkenfeld> |
me too! |
| [11:13:49] |
<mitsuhiko> |
this year: asmodai in: the post interceptor |
| [11:15:42] |
<Leonidas> |
the advantage is, that birkenfeld lives not far away, so the coins can go in one single package. |
| [11:15:50] |
<birkenfeld> |
yep |
| [11:15:58] |
<birkenfeld> |
we can distribute them at the user group |
| [11:16:35] |
<birkenfeld> |
the question is how we pay you the 5€ |
| [11:16:41] |
<birkenfeld> |
send a banknote per post? |
| [11:18:27] |
<mitsuhiko> |
:D |
| [11:18:53] |
<Leonidas> |
birkenfeld: it's more than 5€, because it needs to be delivered to NL and thats costs another 4€ |
| [11:19:04] |
<birkenfeld> |
ah, sure |
| [11:19:44] |
<Leonidas> |
Let's see... how much does money transfer to NL cost? |
| [11:20:25] |
<apollo13> |
nothing? |
| [11:20:27] |
<apollo13> |
*runs* |
| [11:21:00] |
<Leonidas> |
apollo13: yeah, as long as I use IBAN and BIC it's free. |
| [11:21:07] |
<apollo13> |
indeed |
| [11:21:11] |
<Leonidas> |
I'd go that way. |
| [11:22:12] |
<Leonidas> |
apollo13: but it is not in every EU country like this. In Poland you have to pay for it, because you get no of the EU-goodies for free there, you have to fight for each one. |
| [11:22:27] |
<apollo13> |
that's why I wasn't sure |
| [11:23:14] |
<Leonidas> |
http://www.postbank.de/-snm-0184330283-1221469788-029790000c-0000000104-1221470412-enm-privatkunden/pk_auslandszahlungen.html;jsessionid=F661933C6B378375BBD53AC30C856BC1F8F5.f086 |
| [11:23:34] |
<Leonidas> |
(hooray for stupid and long urls) |
| [11:25:09] |
<apollo13> |
the urls here are also something like this: http://raiffeisen.at/eBusiness/rai_template1/1006637000974-1006622331426_459441033793251264-459441033793251264-NA-1-NA.html |
| [11:25:22] |
<apollo13> |
but not longer, the just swap the digits in there |
| [11:25:37] |
<apollo13> |
and that for every url :/ |
| [11:25:55] |
<Leonidas> |
hehe :) |
| [11:26:10] |
<Leonidas> |
spupid, long and _pointless_ urls. |
| [11:26:39] |
<apollo13> |
I wonder why the need the .html |
| [11:26:48] |
<apollo13> |
to make it longer? |
| [11:27:44] |
<apollo13> |
http://www.python-forum.de/post-110228.html#110228 /me would use internal redirects |
| [11:29:27] |
<Leonidas> |
apollo13: the .html is probably for broken browsers or something. |
| [11:29:42] |
<apollo13> |
their site is broken, why care for broken browsers? |
| [11:30:36] |
<Leonidas> |
apollo13: I'd just generate the page into htdocs and return a redirect. Or return the output directly, if it is small enough. |
| [11:31:43] |
<Leonidas> |
Unfortunately, pdfLaTeX cannot really output PDF to a pipe, so you have to go the harddrive way in any case. |
| [11:32:07] |
<apollo13> |
why generate it into htdocs? |
| [11:32:12] |
<apollo13> |
that's an unneeded overhead |
| [11:32:26] |
<apollo13> |
you need to take care of deletion etc... |
| [11:32:55] |
<plaes> |
could someone make a small change to lodgeit? |
| [11:33:30] |
<plaes> |
when viewing the paste, there's this "Paste Details" header... |
| [11:34:17] |
<plaes> |
could someone change this to "View Paste Details" and also add straight "Download" link... |
| [11:34:29] |
<plaes> |
s/straight/direct |
| [11:45:52] |
<asmodai> |
Leonidas / birkenfeld / mitsuhiko : So I need to keep 3 in mind? |
| [11:45:55] |
<asmodai> |
Remember, there's 3 options |
| [11:46:01] |
<asmodai> |
the silver coated one, which is ~ EUR 5 |
| [11:46:15] |
<asmodai> |
the silver one which is EUR 18 (and which I can order multiple of) |
| [11:46:24] |
* |
izibi_ has joined #pocoo |
| [11:46:26] |
<Leonidas> |
apollo13: if it is LaTeX you have to generate it somewhere. I tried quite some time to get it work without writing to disk but failed. |
| [11:46:30] |
<asmodai> |
and the gold one, which is ~ EUR 30 and which I can also order multiple of |
| [11:47:29] |
<apollo13> |
Leonidas: I was speaking more generally, LaTeX might be a problem, yeah |
| [11:48:55] |
* |
Leonidas is considering the silver one instead of the silver coated, but OTOH I'm not really a numismatic. |
| [11:49:23] |
* |
izibi__ has joined #pocoo |
| [11:51:01] |
* |
izibi__ has joined #pocoo |
| [11:53:21] |
* |
izibi has quit IRC |
| [11:53:23] |
* |
izibi_ has quit IRC |
| [11:53:42] |
* |
izibi__ is now known as izibi |
| [12:04:30] |
* |
birkenfeld has quit IRC |
| [12:09:37] |
* |
grumpy_ has joined #pocoo |
| [12:23:09] |
* |
sebner has quit IRC |
| [12:30:29] |
* |
grumpy_ has quit IRC |
| [12:31:31] |
* |
djc has joined #pocoo |
| [12:32:41] |
<djc> |
why does this happen? http://hg.xavamedia.nl/mercurial/crew/annotate/d3f836bf6cc1/contrib/patchbomb#57 |
| [12:40:41] |
<djc> |
mitsuhiko: can you look at that? |
| [12:41:51] |
<mitsuhiko> |
djc: wrong lexer? |
| [12:42:20] |
<djc> |
ah, it doesn't have an extension |
| [12:42:24] |
<mitsuhiko> |
i wonder that there is a lexer at all, no extension |
| [12:42:36] |
<mitsuhiko> |
we should really start fixing the lexer guessing :-/ |
| [12:42:58] |
<djc> |
well, I think we turned off part of the lexer guessing |
| [12:43:12] |
<djc> |
since it was reading all of the files, which took too long |
| [12:43:17] |
<djc> |
also, have you seen this? |
| [12:43:18] |
<djc> |
http://hg.xavamedia.nl/mercurial/crew/file/0d513661d6c2/mercurial/commands.py |
| [12:43:34] |
<djc> |
in my Firefox 3.0.1 on windows, that gets blank pages half-way down the page |
| [12:43:36] |
<mitsuhiko> |
the lexer guessing in pygments is notoriously broken. there is a long standing bug to fix that :-/ |
| [12:43:50] |
<mitsuhiko> |
djc: welcome to the wonderful world of firefox rendering bugs |
| [12:43:57] |
<mitsuhiko> |
it's slightly worse on webkit :-/ |
| [12:44:11] |
<mitsuhiko> |
it only happens from time to time and nobody has an idea why |
| [12:44:33] |
<mitsuhiko> |
proably too many dom elements |
| [12:44:44] |
<djc> |
is there an Fx bug that you know of? |
| [12:45:05] |
<djc> |
I do a little work for mozilla on hg stuff, so maybe I can get a little traction |
| [12:46:38] |
<mitsuhiko> |
djc: it's a problem all browsers have, that's why i never looked for it in deeper detail |
| [12:46:54] |
<mitsuhiko> |
i actually found that firefox 3 is doing a lot better than older versions and i haven't seen it for quite a long time |
| [12:47:17] |
<djc> |
it's not intermittent for me with that page |
| [12:47:23] |
<djc> |
I always get it |
| [12:47:31] |
<djc> |
commands.py is just pretty big |
| [12:49:37] |
<mitsuhiko> |
djc: works here :-/ |
| [12:49:40] |
* |
jinks has quit IRC |
| [12:49:42] |
<mitsuhiko> |
maybe it has to do with the graphics card |
| [12:50:09] |
<mitsuhiko> |
max texture size or something. i have no idea how cairo is rendering that |
| [12:51:45] |
* |
izibi_ has joined #pocoo |
| [12:59:09] |
<Leonidas> |
djc: no rendering problems here. |
| [12:59:37] |
<Leonidas> |
and I mean I have about the worst graphics drivers that you can gat for linux ;) |
| [12:59:41] |
<djc> |
mitsuhiko: https://bugzilla.mozilla.org/show_bug.cgi?id=455297 |
| [13:00:21] |
<aa_> |
Leonidas: man, i think I have one of those cards |
| [13:00:28] |
<aa_> |
Leonidas: like 2 seconds to switch desktop |
| [13:03:11] |
<mitsuhiko> |
Leonidas: that's a windows only problem as far as i can tell |
| [13:03:29] |
<mitsuhiko> |
a similar one existed on linux and os x a while back, but i haven't heard of that for quite a while |
| [13:04:21] |
<Leonidas> |
mitsuhiko: I sometimes had some rendering errors in firefox, but they were very minor and not reproducible |
| [13:05:10] |
* |
jinks has joined #pocoo |
| [13:06:19] |
* |
izibi has quit IRC |
| [13:37:04] |
* |
stifal has joined #pocoo |
| [13:50:09] |
<mitsuhiko> |
_why is great: http://hackety.org/2008/09/15/documentsRevealDjangoPonyTailOfLies.html |
| [13:52:53] |
<plaes> |
hehe.. just found it :D |
| [14:07:44] |
* |
birkenfeld has joined #pocoo |
| [14:09:05] |
* |
leche has quit IRC |
| [14:35:26] |
<CIA-49> |
Jinja 2: mitsuhiko trunk * 566:0c3ac202de59 / (CHANGES jinja2/compiler.py jinja2/ext.py): Fixed a confusing edge case (thanks apollo13) |
| [14:44:10] |
* |
grumpy_ has joined #pocoo |
| [14:44:44] |
<mitsuhiko> |
hi grumpy_ :) |
| [14:44:50] |
<mitsuhiko> |
i think prencher was looking for you |
| [14:45:10] |
<grumpy_> |
Yeah, I know. I saw the bat signal. |
| [14:45:48] |
<grumpy_> |
How is that ubuntu site going now. Any progress on getting it going on mod_wsgi again? |
| [14:46:02] |
* |
pete__c has quit IRC |
| [14:46:02] |
* |
Kaelten has quit IRC |
| [14:46:02] |
* |
maddiin has quit IRC |
| [14:46:02] |
* |
chrisberlin has quit IRC |
| [14:46:02] |
* |
hads has quit IRC |
| [14:46:02] |
* |
TheSheep has quit IRC |
| [14:46:02] |
* |
KAeL has quit IRC |
| [14:48:28] |
<mitsuhiko> |
grumpy_: turns out mod_wsgi has nothing to do with any of the problems there |
| [14:48:37] |
<mitsuhiko> |
cpickle seems to have concurrency problems or something |
| [14:48:56] |
<mitsuhiko> |
however we switched to fcgi because of policition reasons. (Don't ask) |
| [14:49:24] |
<grumpy_> |
I knew you had worked that out after you went to fastcgi. Hmmm, ..... okay will not ask. :-( |
| [14:50:08] |
<mitsuhiko> |
grumpy_: it was hard enough to convince the stupid part of the server team that python is an accepted and secure programming language |
| [14:50:10] |
* |
Kaelten has joined #pocoo |
| [14:50:10] |
* |
pete__c has joined #pocoo |
| [14:50:10] |
* |
TheSheep has joined #pocoo |
| [14:50:10] |
* |
hads has joined #pocoo |
| [14:50:10] |
* |
chrisberlin has joined #pocoo |
| [14:50:10] |
* |
maddiin has joined #pocoo |
| [14:50:10] |
* |
KAeL has joined #pocoo |
| [14:50:18] |
<mitsuhiko> |
but an extra apache module was too much |
| [14:50:49] |
<ronny> |
lol |
| [14:50:49] |
<grumpy_> |
BTW, had a guy at work who needed to do some web stuff, got him to use Werkzeug. He was quite happy with it. |
| [14:51:14] |
<grumpy_> |
For the first time actually got to use mod_wsgi to host a web application. |
| [14:51:27] |
<mitsuhiko> |
cool :D |
| [14:52:11] |
<mitsuhiko> |
grumpy_: btw. i agreed with ben bangert to combine werkzeug.debug, dozer and some new code into a separate library as general purpose wsgi debugging system |
| [14:52:16] |
<grumpy_> |
I should really put up a wiki page on using mod_wsgi with Werkzeug. Had someone else ask me about a day or so ago as well. |
| [14:52:51] |
<mitsuhiko> |
it's currently designed to be used with a development server (non forking) only, but maybe we could work together to use some mod_wsgi hooks to provide ways to hook it into complex mod_wsgi setups at a later stage |
| [14:52:59] |
<grumpy_> |
Is dozer the profiling tool? |
| [14:53:45] |
* |
delroth has quit IRC |
| [14:53:53] |
* |
delroth has joined #pocoo |
| [14:54:22] |
<mitsuhiko> |
grumpy_: actually, it was a leak finder and got a profiler, but we try to move all that into a new project |
| [14:55:05] |
<prencher> |
mitsuhiko - yeah tux21b told some vivid stories about said server team |
| [14:55:06] |
<grumpy_> |
And werkzeug.debug is the in browser debugger, right? |
| [14:55:38] |
<ronny> |
yeah |
| [14:55:49] |
<mitsuhiko> |
grumpy_: yes |
| [14:55:51] |
<grumpy_> |
Remember my odd ideas about being able to park the debugger state and accessing it through a protected admin rather than back to the client browser? |
| [14:56:05] |
<mitsuhiko> |
yep |
| [14:56:24] |
<mitsuhiko> |
grumpy_: the idea behind the new debugging system is that it opens a separate http server for the debugging interface |
| [14:56:37] |
<prencher> |
grumpy_ - hey, got the chance to user the worker in a more realistic capacity recently with mod_wsgi (no goddamned php), works quite well; still doesn't go up quite as far as nginx, but it's close.. slightly heavier but it does cut down on configuration.. main concern is the abilty of a ddos to exhaust all threads |
| [14:56:38] |
<mitsuhiko> |
and maybe we could write a firefox extension similar to firebug later too |
| [14:57:37] |
<grumpy_> |
But why does it need to be separate web server? With WSGI application composition it surely could host out of same server? |
| [14:58:00] |
<prencher> |
grumpy_ - oh it's not a seperate server in this case, that's the whole point |
| [14:58:27] |
<mitsuhiko> |
grumpy_: the reason why a separate server is because i want a pdb in there |
| [14:58:49] |
<mitsuhiko> |
grumpy_: so you can step through the application and in a separate browser window you have the debugger, the logging output handler, a shell and the profiler |
| [14:58:50] |
<prencher> |
oh for mitsuhiko, my bad :p |
| [14:59:22] |
<mitsuhiko> |
grumpy_: it's actually a separate wsgi application. if the server supports multithreaded operations we could host it at http://server/_debuginterface or something |
| [14:59:50] |
<grumpy_> |
Still don't see why a separate server. Anyway, I really got to find some time to actually look at Werkzeug and tyhe debugger stuff. Been caught up playing with chroot environment support for daemon mode. |
| [14:59:55] |
<mitsuhiko> |
but at least for local development environment it would make sense to have it as a separate server because wsgiref is singlethreaded and singlethreaded things help the interactive debugger a lot |
| [15:03:30] |
<grumpy_> |
prencher: brain not thinking well enough to even remember ddos issues and looks like being called to sleep now |
| [15:04:20] |
<prencher> |
grumpy_ - hehe well it's quite simple.. if you have N threads, and files to download..if you can open N connections server dies |
| [15:04:54] |
<grumpy_> |
as far as bringing a server down, one I worry about is mod_cgid as there is a bug in there which makes it easy to lock the server up, brought it up on httpd list couple of times and no response |
| [15:05:48] |
<prencher> |
grumpy_ - still, it's not all that realistic an issue these days, what with hundreds if not thousands of threads being feasible.. but event polling really shines because it can just keep going until you exhaust connection limits on the server (more or less) |
| [15:07:15] |
<grumpy_> |
you don't have to sell me on event driven systems, one of my projects, now open source for 15 years was all about that sort of stuff |
| [15:07:53] |
<prencher> |
grumpy_ - oh i know, just saying.. it seems the one potential weakness of worker mpm when you dont actually run stuff inside apache |
| [15:08:42] |
<prencher> |
all that the workers are doing in our case is delegating to wsgi daemons and serving files.. only uses around 4mb per instance running 25-50 threads, it's quite nice |
| [15:09:25] |
<grumpy_> |
yeah but still limited number of threads per process, anyway I gotta sleep, at least baby sleeping well tonight, hasn't made a noise in four hours |
| [15:09:43] |
<prencher> |
grumpy_ - indeed :) that's the main one |
| [15:09:46] |
<prencher> |
nn mate |
| [15:10:14] |
<grumpy_> |
I really gotta get irc working through our work firewall, presuming that is that timezones overlap that it even helps |
| [15:13:23] |
<prencher> |
grumpy_ - hehe, i should imagine irc with a bunch of eurotrash is hard as an aussie ;) |
| [15:15:57] |
<grumpy_> |
enough, I gotta sleep now :-) |
| [15:16:07] |
* |
grumpy_ has quit IRC |
| [15:17:45] |
<mitsuhiko> |
birkenfeld: http://github.com/chrisdone/lojban-haskell-utilities/tree/master/Language/Lojban/Jbovlaste.hs |
| [15:17:56] |
<mitsuhiko> |
seems like a bug in pygments |
| [15:19:47] |
<birkenfeld> |
mitsuhiko: multiple bugs in there... |
| [15:19:49] |
<birkenfeld> |
not nice |
| [15:20:05] |
<mitsuhiko> |
is ' part of identifiers? |
| [15:21:37] |
<birkenfeld> |
yes |
| [15:21:46] |
<birkenfeld> |
if it occurs in identifiers |
| [15:22:15] |
<mitsuhiko> |
so 'foo' is a valid identifier? |
| [15:22:24] |
<birkenfeld> |
no |
| [15:22:27] |
<birkenfeld> |
but foo' is |
| [15:22:31] |
<birkenfeld> |
'a' is a char litera |
| [15:22:32] |
<birkenfeld> |
l |
| [15:22:40] |
<birkenfeld> |
'foo' is a syntax error |
| [15:22:55] |
<birkenfeld> |
it's like digits in other languages, they can't start identifiers too |
| [15:23:06] |
<mitsuhiko> |
ah |
| [15:40:06] |
<Leonidas> |
github does not like me :/ |
| [15:41:15] |
<empty> |
that's ok because I don't like github :) |
| [15:41:48] |
<CIA-49> |
Pygments: gbrandl trunk * 666:b2f26f55404e / (CHANGES pygments/lexers/functional.py): Some fixes in the Haskell lexer. |
| [15:46:00] |
<mitsuhiko> |
Leonidas: as long as you like bitbucket :D |
| [15:50:55] |
* |
kaelten_ has joined #pocoo |
| [15:52:04] |
* |
Kaelten has quit IRC |
| [15:52:27] |
* |
kaelten_ has quit IRC |
| [15:52:38] |
* |
Kaelten has joined #pocoo |
| [15:52:44] |
* |
djc has left #pocoo |
| [16:07:49] |
* |
highwaychile has joined #pocoo |
| [16:14:19] |
<ronny> |
empty: you got any idea how to interpret the output of git ls-files? |
| [16:15:08] |
<ronny> |
its haunting me the last few days, trying to figure how to get the informations that git status will output |
| [17:07:04] |
<asmodai> |
Kaelten / prencher : talk about being official now: http://www.worldofwarcraft.com/splash-wrathdate.htm |
| [17:14:24] |
* |
murphy_karasu has joined #pocoo |
| [17:14:32] |
<murphy_karasu> |
hello! |
| [17:15:53] |
<murphy_karasu> |
*nudges birkenfeld and mitsuhiko |
| [17:27:20] |
<empty> |
ronny: I'm not sure what you're asking? |
| [17:29:55] |
<empty> |
ronny: we recently added an is_dirty flag and the branch_name info to the repo |
| [17:40:48] |
* |
jpcw2002 has joined #pocoo |
| [17:41:22] |
<Leonidas> |
oh hi murphy_karasu |
| [17:41:47] |
* |
murphy_karasu waves |
| [17:41:52] |
<ronny> |
empty: i need to see differences betwen workdir, index and repo in order to infer the 'state' if a file in the workdir |
| [17:42:25] |
<ronny> |
empty: it seems like git ls-files is supposed to do that, but i cant make sense out of the different outputs/output modes |
| [17:42:50] |
<empty> |
ah |
| [17:42:59] |
<empty> |
ronny: I don't know if we're parsing that anywhere but I can look |
| [17:43:26] |
<ronny> |
empty: the parsing is not the issue, correctly interpreting is |
| [17:43:32] |
<ronny> |
its nearly undocumented |
| [17:43:38] |
* |
highwaychile has quit IRC |
| [17:43:59] |
<empty> |
ok |
| [17:44:13] |
<mitsuhiko> |
hi murphy_karasu :) |
| [17:44:30] |
<mitsuhiko> |
murphy_karasu: long time no see :) |
| [17:44:40] |
* |
birkenfeld waves back |
| [17:48:41] |
<murphy_karasu> |
true |
| [17:48:55] |
<murphy_karasu> |
how are you? |
| [17:49:08] |
<birkenfeld> |
mitsuhiko is busy, while I am fine |
| [17:50:41] |
<mitsuhiko> |
hahaha |
| [17:51:28] |
<murphy_karasu> |
have you thought about the idea to re-use pygments lexers in CodeRay? |
| [17:52:27] |
<mitsuhiko> |
thought yes. i have no idea how to accomplish that though |
| [17:52:37] |
<mitsuhiko> |
there are basically two types of lexers in pygments |
| [17:52:44] |
<mitsuhiko> |
regular expression based and custom lexers |
| [17:52:52] |
<mitsuhiko> |
the former being the most common one |
| [17:53:12] |
<mitsuhiko> |
that type of lexer however mostly depends on regular expressions that are not that regular any more |
| [17:53:16] |
<murphy_karasu> |
those are the ones I thought about |
| [17:53:18] |
<birkenfeld> |
... of course many regex lexers also have custom contstructors etc. |
| [17:53:25] |
<murphy_karasu> |
is there a list of custom lexers? |
| [17:53:30] |
<birkenfeld> |
which is however neglectable in the first iteration |
| [17:53:42] |
<mitsuhiko> |
last time i looked the ruby regular expression support was very weak and 1.9 is not very common currently |
| [17:53:48] |
<mitsuhiko> |
murphy_karasu: sec |
| [17:54:05] |
<murphy_karasu> |
it only lacks lookbehind |
| [17:54:14] |
<birkenfeld> |
does it have atomic grouping? |
| [17:54:20] |
<murphy_karasu> |
yes |
| [17:54:23] |
* |
Baumichel has joined #pocoo |
| [17:54:26] |
<birkenfeld> |
that's good |
| [17:54:33] |
<birkenfeld> |
might speed things up compared to Python's |
| [17:54:34] |
<murphy_karasu> |
but we could work around those problems |
| [17:55:22] |
<mitsuhiko> |
>>> [x.__name__ for x in lexers._iter_lexerclasses() if x.__base__ is Lexer] |
| [17:55:22] |
<mitsuhiko> |
['DelphiLexer', 'RubyConsoleLexer', 'SqliteConsoleLexer', 'ErbLexer', 'MatlabSessionLexer', 'MakefileLexer', 'PythonConsoleLexer', 'MakefileLexer', 'RawTokenLexer', 'TextLexer', 'LiterateHaskellLexer'] |
| [17:55:26] |
<birkenfeld> |
anyway, the sane thing would be to share the lexer defs in a neutral format and then generate python or ruby from it |
| [17:55:46] |
<birkenfeld> |
assuming you don't want to parse Python with Ruby :) |
| [17:55:55] |
<murphy_karasu> |
exactly |
| [17:56:15] |
<murphy_karasu> |
I think the lexer definitions look almost like a decriptive language |
| [17:56:19] |
<mitsuhiko> |
birkenfeld: maybe that could become pygments 2 where we also have the chance to fix lexer nesting |
| [17:56:28] |
<murphy_karasu> |
that subset of Python is easy to parse, I think |
| [17:56:35] |
<birkenfeld> |
mitsuhiko: why 2? there is not 1 yet :) |
| [17:56:46] |
<mitsuhiko> |
birkenfeld: two as in "we break api" |
| [17:56:52] |
<mitsuhiko> |
at least internally |
| [17:57:01] |
<birkenfeld> |
okay |
| [17:57:07] |
<birkenfeld> |
I don't worry much about internally |
| [17:57:18] |
<birkenfeld> |
I don't think many people have custom lexers running |
| [17:57:21] |
* |
mitsuhiko eagerly waits for the next bitbucket deployment to see his changes in action |
| [17:57:25] |
<mitsuhiko> |
gnaa. wrong channel |
| [17:57:31] |
<birkenfeld> |
most people who write them also do contribute them |
| [17:57:58] |
<mitsuhiko> |
birkenfeld: i'm actually quite annoyed by the number of pygments bug reports that get lost in different tracker |
| [17:58:00] |
<mitsuhiko> |
+s |
| [17:58:19] |
<mitsuhiko> |
i'm now grepping the github one more frequently |
| [17:58:47] |
<birkenfeld> |
well, as far as I'm concerned there is only one tracker :) |
| [17:59:11] |
<birkenfeld> |
people complaining about highlighting issues in e.g. the github tracker should be forwarded |
| [18:00:04] |
<murphy_karasu> |
mitsuhiko: the console lexers are not based on regexps, it seems |
| [18:00:19] |
<mitsuhiko> |
murphy_karasu: btw. since when does ruby have atomic grouping? |
| [18:00:24] |
<murphy_karasu> |
I understand why simple ones like Text and RawToken are special |
| [18:00:36] |
<murphy_karasu> |
I thought it had, let me check |
| [18:00:41] |
<mitsuhiko> |
murphy_karasu: because as said above, lexer nesting is currently pretty ugly |
| [18:01:01] |
<mitsuhiko> |
at least for the typical repl scenario |
| [18:01:25] |
<birkenfeld> |
yes. |
| [18:01:40] |
<birkenfeld> |
there's also the old problem with "</script>" in the JSHTML lexer |
| [18:01:56] |
<murphy_karasu> |
/(?>a*)a/ === 'aaa' #=> nil |
| [18:01:58] |
<mitsuhiko> |
birkenfeld: "</script>"? |
| [18:02:05] |
<prencher> |
asmodai - guess thye made the year after all |
| [18:02:41] |
<birkenfeld> |
mitsuhiko: http://dev.pocoo.org/projects/pygments/ticket/242 |
| [18:02:54] |
<mitsuhiko> |
birkenfeld: that's not even valid html |
| [18:03:06] |
<mitsuhiko> |
there is a reason why json outputs "<\/script>" ;) |
| [18:03:06] |
<murphy_karasu> |
so atomic grouping works in Ruby 1.8. lookbehind does not, sadly, but mostly it's possible to find a workaround |
| [18:03:10] |
<mitsuhiko> |
*simplejson |
| [18:03:11] |
<birkenfeld> |
mitsuhiko: oh, okay |
| [18:03:24] |
<mitsuhiko> |
i'll close it |
| [18:03:24] |
<birkenfeld> |
mitsuhiko: well, then I guess the bug is invalid :) |
| [18:04:30] |
<mitsuhiko> |
closed |
| [18:04:34] |
* |
leche has joined #pocoo |
| [18:04:48] |
<murphy_karasu> |
mitsuhiko, birkenfeld: so there are no license related or political problems with me stealing your lexers? |
| [18:04:55] |
<murphy_karasu> |
;) |
| [18:05:06] |
<mitsuhiko> |
murphy_karasu: of course not. but i think it would be a good idea to cooperate on that |
| [18:05:31] |
<mitsuhiko> |
even though i have no idea by now how exactly this will happen |
| [18:05:44] |
<mitsuhiko> |
i have already problems getting the jinja2 lexer pass the .net regular expressions :-/ |
| [18:05:45] |
<murphy_karasu> |
I would make sure that any bug reported on that in CodeRay trackers would be send to yours |
| [18:06:47] |
<birkenfeld> |
mitsuhiko: jinja2/.net? |
| [18:06:55] |
<murphy_karasu> |
mitsuhiko: it's just an experiment, it would be beta for at least a year before I release such a thing |
| [18:07:18] |
<mitsuhiko> |
birkenfeld: ironpython implements re as thin wrapper over System.Text.RegularExpressions |
| [18:07:33] |
<birkenfeld> |
mitsuhiko: ah, jinja2 uses regexes? |
| [18:07:45] |
<mitsuhiko> |
which is the last hurdle required to get it running on ironpython |
| [18:07:55] |
<mitsuhiko> |
birkenfeld: yep. the jinja2 lexer does |
| [18:07:57] |
<murphy_karasu> |
mitsuhiko: but I like the idea of a Ruby project using Python code anyway ;) and it would be interesting to get benchmarks from both engines |
| [18:08:08] |
<mitsuhiko> |
it's really hideous what the lexer is doing there, but i just took it unchanged from 1.0 |
| [18:08:26] |
<mq> |
mitsuhiko: why the hell do they do that? That's clearly a big bug. |
| [18:08:39] |
<birkenfeld> |
murphy_karasu: so would you try parsing the python regexlexer definitions for a start? |
| [18:08:42] |
<mitsuhiko> |
murphy_karasu: if you want to parse the python files, python's compile() takes a forth parameter that when set to 0x400 gives you the ast |
| [18:09:05] |
<murphy_karasu> |
birkenfeld: yeah, after I got the next stable version out |
| [18:09:06] |
<mitsuhiko> |
mq: it is, certainly. it's one of the great ironpython wtfs |
| [18:09:06] |
<birkenfeld> |
ah yes. that's of course elegant |
| [18:09:18] |
<birkenfeld> |
using the Python AST |
| [18:09:38] |
<mq> |
mitsuhiko: and they dont't intend to fix it? |
| [18:09:41] |
<mitsuhiko> |
mq: they are lazy. some people showed them _sre from pypy but they refuse to implement it |
| [18:09:47] |
<mitsuhiko> |
that's why there is a community edition currently |
| [18:09:55] |
<mq> |
... |
| [18:10:06] |
<mq> |
idiots |
| [18:10:14] |
<mitsuhiko> |
mq: they also insist on changing python semantics |
| [18:10:18] |
<mitsuhiko> |
type.__call__ is type |
| [18:10:19] |
<mitsuhiko> |
True |
| [18:10:28] |
<mitsuhiko> |
isn't that hilarious? |
| [18:10:38] |
<prencher> |
mitsuhiko - and no deterministic gc.. |
| [18:10:41] |
<prencher> |
which is just..retarded |
| [18:10:45] |
<birkenfeld> |
mitsuhiko: well, one can argue these are implementation-defined semantics |
| [18:10:58] |
<mitsuhiko> |
prencher: which is what all python impls except of cpython offer |
| [18:11:04] |
<mq> |
why the hell do they even implement python in the first place, then? |
| [18:11:11] |
<mitsuhiko> |
birkenfeld: type.__call__ being type? doubt it |
| [18:11:15] |
<prencher> |
mitsuhiko - that doesnt make it any less retarded |
| [18:11:25] |
<mitsuhiko> |
i can't pass the type as first argument to type but to type.__call__ |
| [18:11:32] |
<prencher> |
deterministic is a *good thing*.. it makes for a lot cleaner code |
| [18:11:38] |
<murphy_karasu> |
mitsuhiko, birkenfeld: 'll inform you about my progress :) but I'm off now to visit a friend. thanks for the chat! |
| [18:11:39] |
<mitsuhiko> |
try to subclass a metaclass and override the default call behavior |
| [18:11:48] |
<mitsuhiko> |
murphy_karasu: cya :) |
| [18:11:50] |
<birkenfeld> |
murphy_karasu: our pleasure |
| [18:11:53] |
<birkenfeld> |
mitsuhiko: well, I find "type(type) is type" equally interesting |
| [18:11:55] |
<prencher> |
mitsuhiko - you end up with shit like C#'s using() without it |
| [18:12:01] |
<mitsuhiko> |
birkenfeld: but that doesn't break code |
| [18:12:04] |
<birkenfeld> |
mitsuhiko: no |
| [18:12:05] |
<prencher> |
which is ugly and means you can "temporarily" leak stuff if you forget |
| [18:12:09] |
<birkenfeld> |
of course not, since it's CPython |
| [18:12:11] |
<prencher> |
RAII has no such issue |
| [18:12:17] |
* |
birkenfeld is out with the dogs |
| [18:12:18] |
* |
murphy_karasu has quit IRC |
| [18:12:30] |
<mitsuhiko> |
birkenfeld: so you argue that ironpython's type.__call__ is type behavior is good because it makes it impossible to subclass it? |
| [18:12:32] |
<mitsuhiko> |
if yes, okay :D |
| [18:12:45] |
<mitsuhiko> |
prencher: there is no raii in python except for with |
| [18:13:00] |
<mitsuhiko> |
and cpython's gc isn't deterministic either. refcounting is, but that's not a gc |
| [18:16:13] |
<prencher> |
mitsuhiko - http://paste.pocoo.org/show/WXLuuRazx00NYlwqBgZf/ . deterministic |
| [18:17:00] |
<prencher> |
which means you can do RAII stuff like getting rid of file handles reliably in del; god help you in IP, Jython, .NET (other than C++/CLI) if you forget to explicit release in those |
| [18:17:06] |
<prencher> |
then it'll just dissappear eventually |
| [18:17:17] |
<prencher> |
which can be in a second or in 20 minutes. |
| [18:18:10] |
<mitsuhiko> |
prencher: that's just because the refcount drops to zero |
| [18:18:18] |
<prencher> |
mitsuhiko - what' |
| [18:18:20] |
<prencher> |
er |
| [18:18:24] |
<mitsuhiko> |
that's not the gc |
| [18:18:47] |
<prencher> |
mitsuhiko - you can say it's not the gc all you want, it's what manages memory in python other than cyclic |
| [18:18:56] |
<prencher> |
and you can rely on that behaviour |
| [18:19:08] |
<prencher> |
i really dont care if you want to argue semantics over whether its a gc or not |
| [18:19:11] |
<mitsuhiko> |
prencher: raise an exception |
| [18:20:51] |
<mitsuhiko> |
prencher: http://paste.pocoo.org/show/6YMZgsxJ4Vn770c6NC30/ |
| [18:22:13] |
<prencher> |
mitsuhiko - yes, im aware in that case it wont immdiately clear it, but that doesnt change that in the normal case python gets rid of resources much faster.. it's certainly not perfect (need full on RAII like C++ for that), but in C# & co you're just SOL if you dont wrap everything in using, basically |
| [18:22:37] |
<birkenfeld> |
mitsuhiko: it's not good, but I assume there is some reason |
| [18:22:44] |
<birkenfeld> |
perhaps it's just laziness though |
| [18:22:53] |
<prencher> |
in those languages you basically need the 'exceptional' handling code *always*, in python you dont |
| [18:23:58] |
<mitsuhiko> |
d has the best raii support of all those languages, even though it has a non deterministic gc |
| [18:24:13] |
<mitsuhiko> |
you just have to put "scope" or how the keyword is in front of the var declaration |
| [18:24:33] |
<prencher> |
mitsuhiko - there's no reason .NET can't.. it in fact can |
| [18:24:37] |
<prencher> |
C++/CLI does it |
| [18:25:01] |
<prencher> |
they gate their bytecode with generated code to clean up objects, just like in native code |
| [18:25:08] |
<mitsuhiko> |
that is not garbage collected unless you're using gcnew |
| [18:25:09] |
<prencher> |
it's just unfortunate the CLI doesnt provide something builtin to handle it |
| [18:26:17] |
<prencher> |
mitsuhiko - the behaviour is still C++, and its still gc'ed .. it's complicated |
| [18:26:46] |
<prencher> |
but what they did is wrap their code like they normally do, just with bytecodes (try finally basically) to clean objects, instead of native code |
| [18:27:11] |
<prencher> |
raii makes stuff A LOT cleaner, that's the thing.. it makes me wonder what braindead fuck decided not to have it as a native feature of the CLI and C# |
| [18:27:45] |
<prencher> |
if you're not using `using' all over the place, you can essentially leak like crazy, only difference is its temporary and much more subtle than oldschool leaks |
| [18:27:56] |
<prencher> |
file handles, sockets and other such fun limited resources |
| [18:30:55] |
<prencher> |
ah well, at least its not as bad as the jvm |
| [18:31:34] |
<mitsuhiko> |
i like using. it's simple |
| [18:31:49] |
<prencher> |
mitsuhiko - not simpler than not using it |
| [18:32:04] |
<prencher> |
its verbose, and its "easy" to forget |
| [18:32:18] |
<mitsuhiko> |
using doesn't kill you and it makes you think if you really want to clean up now |
| [ |