2006-08-06T00:02:59 <Kepplar> moin
2006-08-06T09:58:20 <Kepplar> moin
2006-08-06T12:47:52 <fpletz> ThomasWaldmann: could you please update the wikiconfig on xapian.wikiwikiweb.de? it's broken due to the multiconfig move
2006-08-06T13:33:19 <ThomasWaldmann> ok
2006-08-06T13:34:35 <ThomasWaldmann> fpletz: works now
2006-08-06T13:34:55 <fpletz> thanks
2006-08-06T13:35:33 * ThomasWaldmann noticed that fcgi uses the rather dumb thread module and not threading.
2006-08-06T13:37:10 <fpletz> mmh, pity, the htdocs of the xapian branch aren't aliased ;)
2006-08-06T13:37:24 <fpletz> http://thorus.nerdcamp.net/wiki/?action=fullsearch&context=180&value=test&fullsearch=Text
2006-08-06T13:37:28 <fpletz> looks like this now
2006-08-06T13:38:59 <ThomasWaldmann> wait, i change that
2006-08-06T13:43:06 <ThomasWaldmann> hehe
2006-08-06T13:43:16 <ThomasWaldmann> looks familiar
2006-08-06T13:44:34 <ThomasWaldmann> fpletz: the arrow gfx should be linked also
2006-08-06T13:45:26 <fpletz> ah, right ;)
2006-08-06T13:45:44 <ThomasWaldmann> rev: 1 (current) <- if you already implemented this, it should be either 0 for current OR the real rev. number
2006-08-06T13:45:57 <fpletz> it's t he real rev number
2006-08-06T13:46:04 <fpletz> current is just append if it's current
2006-08-06T13:46:06 <fpletz> +ed
2006-08-06T13:46:39 <fpletz> this will be useful for searches in the history of documents
2006-08-06T13:47:15 <ThomasWaldmann> oh, my fault (it was stuff from underlay, therefore always 1)
2006-08-06T13:47:41 <fpletz> ah, okay
2006-08-06T13:48:02 <ThomasWaldmann> did you contact starshine or thesheep?
2006-08-06T13:48:16 <ThomasWaldmann> i guess we really need gfx now :)
2006-08-06T13:48:29 <fpletz> starshine was away when I was here.. but I think she already started working on it
2006-08-06T13:48:32 <fpletz> so I'll wait for her ;)
2006-08-06T13:48:47 <fpletz> I contacted her last week, I think
2006-08-06T13:49:31 <ThomasWaldmann> I guess we just point her to current output, so she sees it's urgent now :)
2006-08-06T13:49:46 <fpletz> hehe, yeah ;)
2006-08-06T13:49:55 <ThomasWaldmann> starshine_away: see http://xapian.wikiwikiweb.de/?action=fullsearch&context=180&value=test&fullsearch=Text
2006-08-06T13:51:17 <ThomasWaldmann> until we have some better method, we could index a layer name (standard or underlay) to be able to exclude underlay pages from search
2006-08-06T13:52:12 <mitsuhiko> ThomasWaldmann: Google?
2006-08-06T13:52:58 <fpletz> yeah, will include that
2006-08-06T13:53:37 <ThomasWaldmann> mitsuhiko: it will be "< Moooooooooin >" soon
2006-08-06T13:53:54 <mitsuhiko> ^^
2006-08-06T13:54:29 <ThomasWaldmann> we are just waiting for some artist to jump on it
2006-08-06T13:56:10 <ThomasWaldmann> fpletz: that "similar pages" stuff, are you working on that?
2006-08-06T13:56:31 <fpletz> no, currently not.. just wanted to place some link there for cosmetic reasons ;)
2006-08-06T14:16:51 <xorAxAx> ThomasWaldmann: we can already differentiate system pages from normal pages
2006-08-06T14:16:57 <xorAxAx> and people want to search for system pages
2006-08-06T14:17:03 <xorAxAx> (to find config options etc.)
2006-08-06T14:17:15 <xorAxAx> underlay/data is not a good differentiation IMHO
2006-08-06T14:19:08 <xorAxAx> hmm, its rather slow - i need 3 to 8 seconds per request
2006-08-06T14:19:13 <fpletz> xorAxAx: but an option for that in the advanced ui would be nice anyway, I guess
2006-08-06T14:19:20 <xorAxAx> fpletz: sure
2006-08-06T14:19:34 <fpletz> xorAxAx: could you give me the timings?
2006-08-06T14:19:45 <xorAxAx> _xapianSearch = 0.010s compile_huge_and_ugly = 0.335s getACL = 0.068s getPageCount = 0.008s run = 1.114s total = 7.649s
2006-08-06T14:20:03 <xorAxAx> the total vs. run difference is interesting
2006-08-06T14:20:11 <fpletz> ok, not my fault ;)
2006-08-06T14:21:09 <xorAxAx> my sync wiki on the same machine:
2006-08-06T14:21:11 <xorAxAx> _moinSearch = 0.981s getACL = 0.078s getPageCount = 0.000s getPageList = 0.061s run = 1.654s total = 3.284s
2006-08-06T14:21:27 <xorAxAx> showing all results for the same search :)
2006-08-06T14:21:43 <fpletz> mmh, okay, maybe my faul ;)
2006-08-06T14:21:45 <fpletz> +t
2006-08-06T14:21:51 <fpletz> what were your search terms?
2006-08-06T14:21:57 <xorAxAx> i used the query above
2006-08-06T14:22:00 <xorAxAx> "Test"
2006-08-06T14:23:00 <xorAxAx> its interesting to benchmark it in different scenario
2006-08-06T14:23:00 <xorAxAx> s
2006-08-06T14:23:15 <xorAxAx> xapian.www.de is the scenario "machine with high load" :)
2006-08-06T14:24:53 <fpletz> hehe, okay.. I'll introduce some more timers, I'm curious what code's causing that
2006-08-06T14:25:12 <fpletz> or is there an easy way for profiling moin?
2006-08-06T14:25:21 <xorAxAx> no
2006-08-06T14:25:37 <xorAxAx> but run vs. total means that it is losing the time in the request handler
2006-08-06T14:26:09 <xorAxAx> it could be e.g. flushing the request stream to the server/browser
2006-08-06T14:27:21 <xorAxAx> it is doing the following:
2006-08-06T14:27:25 <xorAxAx> a) loading the config
2006-08-06T14:27:40 <xorAxAx> b) loading the i18n stuff
2006-08-06T14:28:01 <xorAxAx> c) opening logs
2006-08-06T14:28:25 <xorAxAx> ThomasWaldmann: it looks like we need to profile that - b) might have introduced a performence regression.
2006-08-06T14:28:40 <xorAxAx> spending 66% of the time for those 3 tasks is a little bit frightening
2006-08-06T14:28:47 <xorAxAx> s/men/man/
2006-08-06T14:29:38 <xorAxAx> it could be the config as well
2006-08-06T14:29:41 <ThomasWaldmann> or the timers are just broken
2006-08-06T14:30:01 <xorAxAx> i dont think so - it is only broken for recursive calls
2006-08-06T14:30:07 <xorAxAx> neither run nor total are recursive
2006-08-06T14:30:25 <xorAxAx> but the server IO influences the timer
2006-08-06T14:30:29 <xorAxAx> on localhost, it might be different
2006-08-06T14:31:14 <xorAxAx> fpletz: so its not related to search :)
2006-08-06T14:31:26 <fpletz> hehe
2006-08-06T14:32:21 <fpletz> but I just saw that not all xapian code is between the timer calls
2006-08-06T14:32:26 <fpletz> will fix that ;)
2006-08-06T14:35:28 <fpletz> ok, pushed and tested.. no regression in my code ;)
2006-08-06T14:41:02 <fpletz> a friend wants to make the gfx. let's see what he manages to create
2006-08-06T14:44:57 <dino4k> moin :)
2006-08-06T14:47:21 <dino4k> nice :)
2006-08-06T14:47:34 <dino4k> ops
2006-08-06T14:48:35 <ThomasWaldmann> def start(self, timer):
2006-08-06T14:48:35 <ThomasWaldmann> self.timings[timer] = time.time() - self.timings.get(timer, 0)
2006-08-06T14:48:52 <ThomasWaldmann> sense?
2006-08-06T14:49:52 <ThomasWaldmann> sense for the case when time already was started?
2006-08-06T14:51:40 <ThomasWaldmann> ah, it is for the case when the timer was stopped. but it doesnt handle the erroneous case when timer is just started twice.
2006-08-06T14:52:09 <xorAxAx> ThomasWaldmann: yes, it should simply do nothing in that case
2006-08-06T14:52:23 <xorAxAx> or create a stack internally and create a new dict entry
2006-08-06T14:52:29 <ThomasWaldmann> i fix that stuff
2006-08-06T14:52:55 <xorAxAx> can you add a timer around load_multiconfig and load_i18n as well?
2006-08-06T15:01:49 <dino4k> fpletz showed me his work on Moin [searching stuff] and now i'm working on the gfx for it ;) could you please comment on: http://phidev.org/~dino/moin/Moin.png
2006-08-06T15:02:28 <xorAxAx> dino4k: sans-serif would look nicer IMHO
2006-08-06T15:02:52 <xorAxAx> while google's letters are bolded, these just have shadows
2006-08-06T15:03:40 <dino4k> ok, i'll try out ;)
2006-08-06T15:19:40 <dino4k> xorAxAx: i updated http://phidev.org/~dino/moin/Moin.png ;)
2006-08-06T15:19:47 <dino4k> just for you :p
2006-08-06T15:22:18 <ThomasWaldmann> looks unsharp
2006-08-06T15:22:31 <ThomasWaldmann> (the arrows)
2006-08-06T15:23:21 <ThomasWaldmann> dino4k: and the letters look rather differently sharp
2006-08-06T15:24:11 <ThomasWaldmann> maybe due to AA?
2006-08-06T15:25:21 <dino4k> hm
2006-08-06T15:25:52 <dino4k> wait a moment, i'm working on another pair of arrows ;)
2006-08-06T15:37:20 <dino4k> first the arrows... http://phidev.org/~dino/moin/Moin.png
2006-08-06T15:38:16 <dino4k> ThomasWaldmann: i like the last one ;)
2006-08-06T15:42:52 <xorAxAx> it is still with serifs on the left side :-p
2006-08-06T15:43:03 <xorAxAx> and its a bit small
2006-08-06T15:43:32 <xorAxAx> and i wonder if it is legal
2006-08-06T15:43:41 <dino4k> hmm
2006-08-06T15:44:25 <fpletz> should we ask google? ;)
2006-08-06T15:44:49 <dino4k> its just like writing "sponsored by google"
2006-08-06T15:46:39 <xorAxAx> note that it is not allowed to reproduce the google logo if you are not doing a verbatim screenshot
2006-08-06T15:46:49 <xorAxAx> and this is quite near to the google logo
2006-08-06T15:47:17 <fpletz> there's this thread on summer-accepted about using google logos where lh indicated we should send her such requests.. maybe drop her a mail?
2006-08-06T15:47:54 <dino4k> fpletz: sounds like a good idea :D
2006-08-06T15:52:07 <ThomasWaldmann> maybe it would look better if it was slightly larger
2006-08-06T15:52:27 <ThomasWaldmann> (use same size as google)
2006-08-06T15:53:17 <ThomasWaldmann> for the sans-serif variant, the M and i look different from the o and n
2006-08-06T15:53:46 <ThomasWaldmann> maybe due to not having roundness
2006-08-06T15:54:00 <ThomasWaldmann> and the o and n look better
2006-08-06T15:58:03 <ThomasWaldmann> and we could swap the blue and red usage, so it looks less similar to google
2006-08-06T15:58:33 <ThomasWaldmann> (swap M and o colour, keep i)
2006-08-06T16:01:46 <ThomasWaldmann> xorAxAx: look at timing on mm wiki now
2006-08-06T16:19:51 <Kepplar> moin
2006-08-06T16:19:58 <Kepplar> British win \o/
2006-08-06T16:21:07 <xorAxAx> ThomasWaldmann: which one?
2006-08-06T16:24:51 <ThomasWaldmann> http://moinmoin.wikiwikiweb.de/
2006-08-06T16:27:16 <Kepplar> hey ThomasWaldmann
2006-08-06T16:27:47 <Kepplar> fpletz: Yes, that's my thread :P
2006-08-06T16:32:31 <xorAxAx> ThomasWaldmann: thats not the 1.6 branch :)
2006-08-06T16:34:09 <Kepplar> base["core/page"] = Rubik(os.path.join(self.str_location, self.str_PAGEDIR, self.ITEM_NAME_PATTERN))
2006-08-06T16:34:13 <Kepplar> File "/usr/lib/python2.4/posixpath.py", line 62, in join
2006-08-06T16:34:15 <Kepplar> any idea?
2006-08-06T16:34:17 <Kepplar> elif path == '' or path.endswith('/'):
2006-08-06T16:34:20 <Kepplar> AttributeError: 'NoneType' object has no attribute 'endswith'
2006-08-06T16:38:19 <xorAxAx> Kepplar: path is None
2006-08-06T16:38:38 <Kepplar> i got that bit :P
2006-08-06T16:39:29 <fpletz> what about "elif not path or path.endswith('/'):"?
2006-08-06T16:39:38 <ThomasWaldmann> xorAxAx: it is a bug fix :) and if you want to see it under load ...
2006-08-06T16:39:48 <Kepplar> ah
2006-08-06T16:39:48 <xorAxAx> fpletz: its in the stdlib, he incorrectly called join
2006-08-06T16:39:51 <Kepplar> i think i may have found it
2006-08-06T16:40:01 <Kepplar> xorAxAx: No, os path did :P
2006-08-06T16:40:05 <xorAxAx> ThomasWaldmann: yeah, but i think it might be the i18n code
2006-08-06T16:40:17 <xorAxAx> ThomasWaldmann: so its not visible in the 1.5 wiki :)
2006-08-06T16:40:37 <xorAxAx> ThomasWaldmann: loading the config is done quickly anyway because you introduced caching
2006-08-06T16:40:38 <ThomasWaldmann> I'll port it to 1.6 also.
2006-08-06T16:40:48 <xorAxAx> yeah
2006-08-06T16:41:13 <Kepplar> awww
2006-08-06T16:41:14 <Kepplar> crap
2006-08-06T16:41:29 <Kepplar> I'm joining the path on the creation of the Rubik
2006-08-06T16:41:54 * Kepplar moves to to be created on open()
2006-08-06T17:15:33 <ThomasWaldmann> ok, new clock code also on 1.6 now
2006-08-06T17:23:00 <xorAxAx> getACL = 0.002s getPageList = 0.069s loadLanguage = 0.004s load_multi_cfg = 0.072s run = 0.378s send_page = 0.275s send_page_content = 0.209s total = 1.950s
2006-08-06T17:23:06 <xorAxAx> it is something else ...
2006-08-06T17:25:17 <Kepplar> i have to say I do like how the timers work
2006-08-06T17:47:00 <Kepplar> anyone home?
2006-08-06T17:47:09 <Kepplar> File "/home/kepplar/dev/moinmoin/storage/MoinMoin/request/__init__.py", line 151, in __init__
2006-08-06T17:47:12 <Kepplar> self.rootpage = Page(self, rootname, is_rootpage=1)
2006-08-06T17:47:22 <Kepplar> request is way outside my area
2006-08-06T17:47:34 <Kepplar> but for some reason its passing an empty page name for rootname
2006-08-06T17:48:14 <Kepplar> this is causing me to be held up so I'm gonna take a break until someone is able to answer :)
2006-08-06T17:56:13 <ThomasWaldmann> you can ignore that
2006-08-06T17:57:32 <ThomasWaldmann> self.rootpage is equivalent to using the data_dir for the pages
2006-08-06T17:58:12 <Kepplar> yes, but its breaking tests :P
2006-08-06T17:58:51 <ThomasWaldmann> then you have to fix something, maybe the tests
2006-08-06T17:59:22 <xorAxAx> request.rootpage is the singleton used for all operations that dont belong to a specific page, Kepplar
2006-08-06T17:59:25 <Kepplar> for some reason because of it the flatfile object itself tries to get encoded
2006-08-06T17:59:29 <xorAxAx> you can refactor that as well
2006-08-06T17:59:39 <xorAxAx> "encoded"?
2006-08-06T18:00:01 <Kepplar> I moved to flatfile Page's methods to encode and code wikinames
2006-08-06T18:00:08 <Kepplar> (2f) business I think
2006-08-06T18:00:18 <Kepplar> to make it fs friendly
2006-08-06T18:01:02 <Kepplar> File "/home/kepplar/dev/moinmoin/storage/MoinMoin/storage/formats/flatfile.py", line 391, in int_name_to_fs_name
2006-08-06T18:01:06 <Kepplar> print "RRRRRRRRRRRRRRRRRR" + file_name
2006-08-06T18:01:08 <Kepplar> TypeError: cannot concatenate 'str' and 'FlatFileProxy' objects
2006-08-06T18:01:17 <Kepplar> i put two print statements in one that repr()'ed the filename other just printed it
2006-08-06T18:02:05 <xorAxAx> file_name doesnt sound like a name for a flatfileproxy object
2006-08-06T18:02:23 <Kepplar> it shouldnt be
2006-08-06T18:02:30 <Kepplar> ive traced it back, its the pagename that gets there
2006-08-06T18:02:31 <Kepplar> supposingly
2006-08-06T18:02:42 <Kepplar> but because its empty string flatfile object it sent itself O_o
2006-08-06T18:03:31 * ThomasWaldmann fixed wrong path in i18n_init, cgi is 1.5s faster now
2006-08-06T18:10:37 <Kepplar> strange
2006-08-06T18:11:02 <Kepplar> on the last function hop, when a blank string is sent, the resulting object from that argument inside the function is the object
2006-08-06T18:11:13 <Kepplar> for example say in class foo we have methods bar()
2006-08-06T18:11:24 <Kepplar> bar(self, string)
2006-08-06T18:11:27 <Kepplar> is the prototype
2006-08-06T18:11:30 <Kepplar> if we call it
2006-08-06T18:11:33 <Kepplar> bar("")
2006-08-06T18:11:38 <Kepplar> then inside bar
2006-08-06T18:11:46 <Kepplar> string is object foo
2006-08-06T18:11:49 <Kepplar> er
2006-08-06T18:12:03 <Kepplar> object of class foo (or class if your calling class but this is besides thepoint)
2006-08-06T18:12:15 <Kepplar> does anyone have a clue why?
2006-08-06T18:13:36 <xorAxAx> Kepplar: well
2006-08-06T18:14:14 <Kepplar> all previous hops on the traceback send the blank (unicode) string
2006-08-06T18:14:14 <xorAxAx> Kepplar: do some tracing
2006-08-06T18:14:23 <Kepplar> i've done that
2006-08-06T18:14:54 <xorAxAx> did you define __unicode__, __str__ or __repr__?
2006-08-06T18:15:07 <Kepplar> ?
2006-08-06T18:15:16 <Kepplar> the object not one of mine
2006-08-06T18:15:25 <Kepplar> its coming from outside storage
2006-08-06T18:15:39 <xorAxAx> flatfileproxies dont come from outside
2006-08-06T18:15:43 <Kepplar> oh
2006-08-06T18:15:45 <Kepplar> yes
2006-08-06T18:15:52 <Kepplar> that as __str__
2006-08-06T18:16:12 <Kepplar> but it goes through 5 methods within it
2006-08-06T18:16:21 <Kepplar> each stage it prints as a blank unicode (which is fine0
2006-08-06T18:16:24 <Kepplar> except the last "hop"
2006-08-06T18:16:30 <Kepplar> when it becomes the damn object itself
2006-08-06T18:16:33 <Kepplar> for no reason
2006-08-06T18:17:37 <Kepplar> and its passed around as a local (ie no self.whatever) in the traceback
2006-08-06T18:24:35 <xorAxAx> maybe the wand "reveal correct object" helps, it can be found in MoinMoin.magic.wands
2006-08-06T18:38:21 <Kepplar> xorAxAx: can't see a folder or module called magic O-o
2006-08-06T18:54:35 <ThomasWaldmann> fpletz: merge main and it'll get faster
2006-08-06T18:54:50 <fpletz> ThomasWaldmann: ok
2006-08-06T18:55:07 <Kepplar> ThomasWaldmann: I dont understand root pages at all, are they treated as page with no name?
2006-08-06T18:56:49 <fpletz> ThomasWaldmann: wow, great ;)
2006-08-06T18:57:18 <ThomasWaldmann> Kepplar: i told you to ignore them. we don't need that concept for 1.5 compatible storage as it is flat.
2006-08-06T18:57:44 <Kepplar> thats fine, except it erroring me
2006-08-06T18:57:59 <ThomasWaldmann> that virtual rootpage concept was just made to be able to extend to hierarchical page storage, but we never did.
2006-08-06T18:58:19 <Kepplar> so why is something trying to use storage with an empty pagename?
2006-08-06T18:58:41 <ThomasWaldmann> maybe it's the rootpage? :-P
2006-08-06T18:58:48 <xorAxAx> Kepplar: i gave you the explanation already
2006-08-06T18:58:48 <Kepplar> well yes i know that:P
2006-08-06T18:58:55 <Kepplar> xorAxAx: that wasn't sufficiant
2006-08-06T18:59:05 <xorAxAx> Kepplar: feel free to pose any question
2006-08-06T18:59:22 <Kepplar> xorAxAx: I would like to know the detailed architecture of moinmoin
2006-08-06T18:59:37 <ThomasWaldmann> if you manage to just access one flat level of pages stored in the data_dir, it is enough for 1.5 compatibility
2006-08-06T18:59:38 <xorAxAx> that is not a question
2006-08-06T18:59:42 <Kepplar> xorAxAx: point being you can't answer that - its soemthing that should be extensively documented
2006-08-06T19:00:02 <xorAxAx> Kepplar: answer what?
2006-08-06T19:00:28 <Kepplar> 17:59 < Kepplar> xorAxAx: I would like to know the detailed architecture of moinmoin
2006-08-06T19:00:40 <Kepplar> i know it wasn't phrased as a question - but we don't need to be pedantic
2006-08-06T19:00:42 <xorAxAx> as i said, that is not a question :)
2006-08-06T19:00:53 <ThomasWaldmann> i like those questions after 2 months
2006-08-06T19:01:11 <xorAxAx> yeah, normally you can gather the architecture by reading the code
2006-08-06T19:01:24 <Kepplar> xorAxAx: thats also known as insufficant documentation :)
2006-08-06T19:01:40 <xorAxAx> Kepplar: sure, i havent seen any docs from you yet :)
2006-08-06T19:01:46 <xorAxAx> this is an opensource project :)
2006-08-06T19:01:58 <Kepplar> xorAxAx: Rubik's documented very well
2006-08-06T19:02:09 <xorAxAx> i mean about moinmoin
2006-08-06T19:02:11 <Kepplar> xorAxAx: plus I'm going to be uploaded UML and stuff after moin
2006-08-06T19:02:26 <xorAxAx> rubik is a complicated way to store tuple-keyed objects in a dictionary
2006-08-06T19:02:35 <xorAxAx> Kepplar: thats cool
2006-08-06T19:03:11 <Kepplar> xorAxAx: does much more than that:P
2006-08-06T19:03:44 <Kepplar> so whats this magic wand?
2006-08-06T19:03:57 <xorAxAx> it was a bad jok
2006-08-06T19:03:57 <xorAxAx> e
2006-08-06T19:04:07 <Kepplar> oh
2006-08-06T19:04:20 <ThomasWaldmann> Kepplar: the 1.5 storage you need to implement is rather obvious if you just look at some data/pages directory
2006-08-06T19:04:30 <Kepplar> this problem makes *no* logical sense
2006-08-06T19:04:55 <Kepplar> ThomasWaldmann: thats not the issue
2006-08-06T19:05:10 <Kepplar> Page *should* be working now
2006-08-06T19:05:13 <ThomasWaldmann> and i already told you multiple time to ignore rootpage stuff
2006-08-06T19:05:24 <Kepplar> im not talking about that directly
2006-08-06T19:05:33 <Kepplar> Im talking about the problems that it causes
2006-08-06T19:05:39 <Kepplar> I can't ignore it because make test fials
2006-08-06T19:05:41 <Kepplar> fails*
2006-08-06T19:05:59 <ThomasWaldmann> then look at why they fail and fix the problem
2006-08-06T19:06:08 <Kepplar> ....
2006-08-06T19:06:13 <Kepplar> Thats what this convosation is about!
2006-08-06T19:07:06 <Kepplar> As i said. Something to do with Rootpage is trying to create a Page with no name. This is then drilling down into flatfile where the WikiName encoding to filesystem safe
2006-08-06T19:07:28 <Kepplar> and suddently on the last function hop the blank unicode string coming in turns into Flatfile object
2006-08-06T19:07:31 <Kepplar> for no apparent reason
2006-08-06T19:08:03 <ThomasWaldmann> remove rootpage code
2006-08-06T19:09:13 <Kepplar> ok, but thats *way* out of storage and if it causes any other errors removing it I simply don't have time to refactor that
2006-08-06T19:09:32 <ThomasWaldmann> there are 2 ways:
2006-08-06T19:10:14 <ThomasWaldmann> either you remove the rootpage stuff and implement a much simpler, but sufficient storage (you may look at old code if you want)
2006-08-06T19:10:27 <ThomasWaldmann> or, you implement it
2006-08-06T19:10:36 <Kepplar> er remove it
2006-08-06T19:11:02 <Kepplar> I assume rootpage stuff is only in request/__init__
2006-08-06T19:11:03 <ThomasWaldmann> and this is well part of storage
2006-08-06T19:11:13 <Kepplar> and not in anything like CGI etc, because make test doesnt do them all?
2006-08-06T19:11:36 <ThomasWaldmann> grep rulez
2006-08-06T19:11:38 <Kepplar> ThomasWaldmann: i mean't its coming way outside storage the call
2006-08-06T19:11:54 <Kepplar> ThomasWaldmann: not my point, point is how am I suppose to test that :P
2006-08-06T19:12:07 <ThomasWaldmann> test what?
2006-08-06T19:13:07 <Kepplar> everything except CLI
2006-08-06T19:13:43 <ThomasWaldmann> storage is not different for different servers
2006-08-06T19:13:43 <Kepplar> ooo!
2006-08-06T19:13:50 <Kepplar> i think its working..well worked WAY more
2006-08-06T19:16:03 <Kepplar> Hmmm
2006-08-06T19:16:06 <Kepplar> this bugs much bigger
2006-08-06T19:16:18 <Kepplar> Traceback (most recent call last):
2006-08-06T19:16:18 <Kepplar> File "/home/kepplar/dev/moinmoin/storage/MoinMoin/_tests/test_formatter.py", line 29, in
2006-08-06T19:16:21 <Kepplar> testSyntaxReference
2006-08-06T19:16:23 <Kepplar> self.formatPage("SyntaxReference", formatter)
2006-08-06T19:16:26 <Kepplar> File "/home/kepplar/dev/moinmoin/storage/MoinMoin/_tests/test_formatter.py", line 38, in
2006-08-06T19:16:29 <Kepplar> formatPage
2006-08-06T19:16:31 <Kepplar> page = Page(self.request, pagename, formatter=formatter)
2006-08-06T19:16:42 <Kepplar> that then runs down to the same error.. flatfile object is ending up as the item_name
2006-08-06T19:16:51 <Kepplar> im going to manually put traces in
2006-08-06T19:16:58 <Kepplar> see if its on the last hop as last time
2006-08-06T19:17:06 <ThomasWaldmann> fpletz: maybe take ANY of those Moooooin gfx until we have better ones. Better than using Gooooooogle.
2006-08-06T19:18:16 <fpletz> ThomasWaldmann: ok
2006-08-06T19:18:49 <ThomasWaldmann> and put them into modern theme
2006-08-06T19:19:47 <ThomasWaldmann> use some common prefix for them and .png
2006-08-06T19:28:07 <Kepplar> ok
2006-08-06T19:28:10 <Kepplar> i've pushed all changes
2006-08-06T19:29:23 <Kepplar> I'd be very very grateful if someone can grab it, make test it (pipe to less) and then look at first error, scroll up to where the error was triggered and see the manual variable prints I put in
2006-08-06T19:29:27 <Kepplar> notice that they are different
2006-08-06T19:29:46 <Kepplar> then look at the code at the last 2 traceback locations, flatfile int_name_to_fs_name and create_type_location
2006-08-06T19:32:40 <Kepplar> then you should see why I'm the edge of dispair ;)
2006-08-06T19:33:35 <xorAxAx> Kepplar: the test code doesnt work
2006-08-06T19:33:46 <Kepplar> well no shit :P
2006-08-06T19:33:47 <xorAxAx> it is bailing out on a name error
2006-08-06T19:34:00 <Kepplar> its not even starting there?
2006-08-06T19:34:06 <xorAxAx> in formats/__init__.py
2006-08-06T19:34:09 <xorAxAx> no, please fix that :)
2006-08-06T19:34:24 <Kepplar> I would if i had the error
2006-08-06T19:34:52 <xorAxAx> `hg status`
2006-08-06T19:35:02 <Kepplar> I don't see how theres and error when its all commented out :P
2006-08-06T19:35:03 <xorAxAx> any uncommitted local changes?
2006-08-06T19:35:10 <Kepplar> no
2006-08-06T19:35:12 <Kepplar> i just pushed
2006-08-06T19:35:15 <xorAxAx> then i cant help you
2006-08-06T19:35:37 <Kepplar> what does cat formats/__init__.py show you?
2006-08-06T19:36:01 <Kepplar> everything should be commented (as according to the repo the changes did go through) except for an import line
2006-08-06T19:36:10 <xorAxAx> noting commented out
2006-08-06T19:36:18 <xorAxAx> 6 b0rked lines of code :)
2006-08-06T19:36:25 <Kepplar> xorAxAx: then your pull didnt work :)
2006-08-06T19:36:39 <xorAxAx> Kepplar: no, it was just delayed
2006-08-06T19:36:49 <xorAxAx> i am pulling from my mirror
2006-08-06T19:36:52 <Kepplar> ah
2006-08-06T19:36:54 * xorAxAx tries again
2006-08-06T19:36:56 <Kepplar> auto mirror?
2006-08-06T19:37:00 <xorAxAx> yes
2006-08-06T19:37:02 <Kepplar> neat
2006-08-06T19:37:06 <Kepplar> hg supports that?
2006-08-06T19:37:12 <xorAxAx> its a cron job
2006-08-06T19:37:16 <Kepplar> hehe
2006-08-06T19:38:23 <xorAxAx> ugh, 8 frames in the storage code
2006-08-06T19:38:32 <Kepplar> 8 frames?
2006-08-06T19:38:36 <xorAxAx> stack frames
2006-08-06T19:38:40 <Kepplar> yes
2006-08-06T19:38:46 <Kepplar> you only need to look at last two
2006-08-06T19:38:49 <Kepplar> i've checked the rest
2006-08-06T19:38:57 <Kepplar> its all working until between the last two
2006-08-06T19:39:29 <Kepplar> as you can see from when the errors actually triggered the same variable is outputted to the CLI from both of the functions
2006-08-06T19:39:36 <Kepplar> one works, one is flatfile itself *shrugs*
2006-08-06T19:41:34 <xorAxAx> the docstring of create_type_location doesnt specify types
2006-08-06T19:41:46 <Kepplar> well i know that :P
2006-08-06T19:42:14 <Kepplar> it should be a string/unicode, which it is
2006-08-06T19:42:39 <ThomasWaldmann> AttributeError: 'str' object has no attribute 'rsplit'
2006-08-06T19:42:53 <Kepplar> ThomasWaldmann: ah i haven't put in xorAxAx's hack
2006-08-06T19:43:01 <xorAxAx> Kepplar: its easy
2006-08-06T19:43:11 <xorAxAx> you forgot self in the method signature of int_name_to_fs_name
2006-08-06T19:43:12 <Kepplar> ThomasWaldmann: so *right* this second it only support 2.4
2006-08-06T19:43:31 <Kepplar> AHHH
2006-08-06T19:43:38 <Kepplar> i didnt see it at all ><
2006-08-06T19:43:54 <Kepplar> xorAxAx: many thanks :)
2006-08-06T19:44:19 <Kepplar> that bits working now :)
2006-08-06T19:44:55 <ThomasWaldmann> Kepplar: maybe you should test on 2.3
2006-08-06T20:05:25 <Kepplar> ThomasWaldmann: will do
2006-08-06T20:09:13 <xorAxAx> fpletz: hmm
2006-08-06T20:09:37 <xorAxAx> fpletz: searching for "the" needs quite long, only 4 seconds spent in xapian - are you generating all results even though you just need to display 10?
2006-08-06T20:10:42 <fpletz> xorAxAx: yes, I need some way to improve this because the sorting is currently done with parts of moinSearch which needs all pages to be loaded (and the acl stuff, too)
2006-08-06T20:10:51 <xorAxAx> fpletz: ugh
2006-08-06T20:11:19 <xorAxAx> the multiconfig is pretty expensive ...
2006-08-06T20:14:03 <fpletz> I'll surely find some way to do this stuff with xapian ;)
2006-08-06T20:16:44 <Kepplar> so out of interest, whats so good about xapian?
2006-08-06T20:19:02 <Kepplar> does str(unicodeobject) result in a String?
2006-08-06T20:19:23 <Kepplar> (i know i know, i just need a quick test)
2006-08-06T20:19:42 <ThomasWaldmann> either in a str or in a exception :)
2006-08-06T20:19:50 <Kepplar> lol!
2006-08-06T20:20:04 <xorAxAx> so never do it
2006-08-06T20:20:05 <Kepplar> good news, ive got it almost accesesing the page
2006-08-06T20:20:10 <Kepplar> it just can't open the str location :)
2006-08-06T20:20:14 <Kepplar> as its a unicode
2006-08-06T20:21:39 <Kepplar> self.cfg.data_underlay_dir < perchance is this unicode?
2006-08-06T20:22:29 <xorAxAx> unicode directory names are not supported
2006-08-06T20:22:37 <xorAxAx> you set that bit in your cfg file
2006-08-06T20:22:51 <Kepplar> but the variable is stored as a unicode?
2006-08-06T20:23:13 <Kepplar> because the itemname should be being translated to ascii
2006-08-06T20:23:26 <Kepplar> which means this is the only piece of the uri thats unicode
2006-08-06T20:23:31 <Kepplar> filepath :P
2006-08-06T20:23:45 <Kepplar> well actually, since its outside storage, "storage location" :P
2006-08-06T21:00:47 <Kepplar> Woo, ok, we're getting somewhere
2006-08-06T21:00:57 <Kepplar> now we I have a decision on how to proceed
2006-08-06T21:01:28 <Kepplar> The mimetype-metadata cache directory
2006-08-06T21:01:45 <Kepplar> how shall we go abouts implementing it?
2006-08-06T21:02:01 <Kepplar> this is the directory with a tiny file for each WikiName saying what mimetype it is
2006-08-06T21:02:09 <Kepplar> its litterally a copy of the mimetype metadata
2006-08-06T21:02:14 <Kepplar> its used to find out where to search
2006-08-06T21:02:40 <Kepplar> aka the mechanism I've put in find where it is before knowing what it is :)
2006-08-06T21:04:20 <Kepplar> (btw I think Page is very much almost working =)
2006-08-06T21:05:19 <ThomasWaldmann> xorAxAx: import multiconfig takes much time
2006-08-06T21:05:42 <Kepplar> plus mimetype is mandatory metadata
2006-08-06T21:06:32 <Kepplar> (and I can't automatically reverse it from itemtype to mimetype as it doesnt work like that and WILL cause issues further down the line)
2006-08-06T21:09:12 <Kepplar> (personally i think we should release a covertion script which just lifts the mimetype out of every wiki entry and puts it into the cache
2006-08-06T21:09:16 <Kepplar> as its far less problematic
2006-08-06T21:11:03 <Kepplar> raise storage_siapi.LocationNonExistant(str_location)
2006-08-06T21:11:03 <Kepplar> LocationNonExistant: ./testwiki/underlay/cache/mimetypes/SyntaxReference
2006-08-06T21:11:05 <Kepplar> Um
2006-08-06T21:11:12 <Kepplar> I just had a poke around the test wiki
2006-08-06T21:11:19 <Kepplar> SyntaxReference doesnt exist
2006-08-06T21:11:22 <Kepplar> why is this being called?
2006-08-06T21:13:01 <ThomasWaldmann> Kepplar: cache must not be authoritative, thus it needs to work without cache, too
2006-08-06T21:13:16 <Kepplar> thats impossible (i think)
2006-08-06T21:13:34 <Kepplar> all it needs to be able to do is have a one time convertion of existing pages upon upgrade
2006-08-06T21:13:44 <Kepplar> it already adds to the cache upon creation of new pages
2006-08-06T21:14:18 <Kepplar> The reason I think this is impossible is that there would have to be so much changes to how storage is designed it will no longer actually be storage implementation neutral
2006-08-06T21:15:12 <Kepplar> as the legacy Flatfile itself would need to know who is calling it
2006-08-06T21:15:24 <Kepplar> which the API does not support, as its not suppose to know
2006-08-06T21:15:34 <Kepplar> which is the entire point of unifying everything
2006-08-06T21:16:42 <Kepplar> ThomasWaldmann: Do you have any idea why it is trying to access SyntaxReference, which according to my nosing around doesn't exist in underlay at all, (In the test wiki that is)
2006-08-06T21:18:22 <ThomasWaldmann> see underlay
2006-08-06T21:18:31 <ThomasWaldmann> it is there
2006-08-06T21:25:22 <fpletz> http://code-bear.com/code/parsedatetime/
2006-08-06T21:26:59 <Kepplar> kepplar@corus:~/dev/moinmoin/storage/testwiki/underlay/pages$ ls -l | grep SyntaxReference
2006-08-06T21:27:02 <Kepplar> kepplar@corus:~/dev/moinmoin/storage/testwiki/underlay/pages$
2006-08-06T21:27:04 <Kepplar> ThomasWaldmann: nope :/
2006-08-06T21:27:11 <ThomasWaldmann> make test
2006-08-06T21:27:29 <Kepplar> it is unpacked
2006-08-06T21:28:01 <Kepplar> HelpOnEditing/Administration etc etc its all there
2006-08-06T21:28:12 <Kepplar> kepplar@corus:~/dev/moinmoin/storage/testwiki/underlay/pages$ ls -l | grep Sy
2006-08-06T21:28:12 <Kepplar> drwxr-xr-x 3 kepplar kepplar 4096 2006-08-06 20:11 SystemPagesInBulgarianGroup
2006-08-06T21:28:12 <Kepplar> drwxr-xr-x 3 kepplar kepplar 4096 2006-08-06 20:11 SystemPagesInLatvianGroup
2006-08-06T21:28:12 <Kepplar> drwxr-xr-x 3 kepplar kepplar 4096 2006-08-06 20:11 SystemPagesInSpanishGroup
2006-08-06T21:28:12 <Kepplar> drwxr-xr-x 4 kepplar kepplar 4096 2006-08-06 20:11 SystemPagesSetup
2006-08-06T21:28:14 <Kepplar> drwxr-xr-x 3 kepplar kepplar 4096 2006-08-06 20:11 WikiKurs(2f)19_Symbole
2006-08-06T21:28:47 <ThomasWaldmann> see underlay.tar.bz2
2006-08-06T21:29:06 <Kepplar> zgrep work for bz2?
2006-08-06T21:33:28 <Kepplar> thats odd
2006-08-06T21:33:32 <Kepplar> it not fully unpacking
2006-08-06T21:34:01 <Kepplar> or does it do it pages in stages to cut down space?
2006-08-06T21:42:13 <ThomasWaldmann> no
2006-08-06T21:46:13 <Kepplar> No one else has got this problem?
2006-08-06T21:46:43 <ThomasWaldmann> maybe your disk is full?
2006-08-06T21:46:51 <Kepplar> :S
2006-08-06T21:47:06 <Kepplar> /dev/hda1 145G 137G 1.4G 99% /
2006-08-06T21:48:00 <ThomasWaldmann> much too full
2006-08-06T21:48:51 <ThomasWaldmann> you should have less percents used also for performance reasons
2006-08-06T21:50:14 <Kepplar> ThomasWaldmann: hda1 :P
2006-08-06T21:50:40 <Kepplar> and 2 of my disks have copped it, so I only have 1
2006-08-06T21:50:41 <Kepplar> too much?
2006-08-06T21:50:43 <Kepplar> hmm
2006-08-06T21:53:19 <Kepplar> ThomasWaldmann: well I dont think so seeing I just extracted it somewhere else on the disk :P
2006-08-06T22:02:55 <ThomasWaldmann> Kepplar: 1. python stdlib imports 2. MoinMoin imports
2006-08-06T22:03:10 <Kepplar> sorry?
2006-08-06T22:06:52 <ThomasWaldmann> is something unclear about that?
2006-08-06T22:07:16 <Kepplar> I don't know what you mean with that line ;)
2006-08-06T22:07:37 <ThomasWaldmann> when implementing modules, you do imports in that sequence
2006-08-06T22:08:02 <xorAxAx> Kepplar: there is a known bug in tarfile.py
2006-08-06T22:08:11 <xorAxAx> the module that should extract pages
2006-08-06T22:08:36 <xorAxAx> i am not sure if it is related
2006-08-06T22:08:45 <Kepplar> Hmm
2006-08-06T22:09:00 <Kepplar> got the bug name on hand (so i can check the wiki)?
2006-08-06T22:09:29 <xorAxAx> it is in the python bugtracker of course
2006-08-06T22:09:48 <xorAxAx> let me fetch the bug no.
2006-08-06T22:09:58 <ThomasWaldmann> Kepplar: don't name modules storage*.py if they are in storage/ anyway
2006-08-06T22:10:16 <Kepplar> ThomasWaldmann: I'll refactor it once i get everythign working, i think thats the most important :P
2006-08-06T22:10:24 <ThomasWaldmann> and *api.py is maybe also not telling much
2006-08-06T22:10:56 <ThomasWaldmann> but expanding that "si" (which is completely none-saying) would be interesting, maybe
2006-08-06T22:11:20 <ThomasWaldmann> put it on your todo
2006-08-06T22:12:24 <xorAxAx> Kepplar: http://sourceforge.net/tracker/index.php?func=detail&aid=1471427&group_id=5470&atid=105470
2006-08-06T22:12:59 <Kepplar> cheers
2006-08-06T22:13:01 <Kepplar> oops
2006-08-06T22:13:07 <Kepplar> er
2006-08-06T22:13:12 <Kepplar> irsii is really playing up
2006-08-06T22:13:33 <xorAxAx> ?
2006-08-06T22:14:33 <Kepplar> I had a line above where I typed oops from another PM window on another server with the irc nick "Alex"
2006-08-06T22:14:48 <Kepplar> lol :)
2006-08-06T22:41:33 <Kepplar> I'm going to remove maketestwiki.py from the Makefile
2006-08-06T22:41:44 <Kepplar> hmm
2006-08-06T22:41:49 <Kepplar> infact I might write a bash script
2006-08-06T23:12:21 <ThomasWaldmann> fpletz: Xapian-Suche Aktiviert, not active
2006-08-06T23:12:33 <ThomasWaldmann> that is slightly confusing :)
2006-08-06T23:12:48 <ThomasWaldmann> (SystemInfo)
2006-08-06T23:18:44 <ThomasWaldmann> xorAxAx: is a Set as fast as a dict for a "x in set" test?
2006-08-06T23:19:00 <xorAxAx> ThomasWaldmann: yes
2006-08-06T23:19:05 <xorAxAx> a Set is a dict wrapper
2006-08-06T23:19:11 <xorAxAx> a set is a little bit faster
2006-08-06T23:19:13 <ThomasWaldmann> ok MoinMoin: MoinMoinChat/Logs/moin-dev/2006-08-06 (last edited 2007-10-29 19:07:00 by localhost)
