MoinMoin: MoinMoinChat/Logs/moin-dev/2007-06-09

2007-06-09T02:09:59  <zenhase> hmm
2007-06-09T02:10:17  <zenhase> i am hunting py:match-bugs
2007-06-09T02:14:13  <zenhase> hmm, perhaps trying to parse the input as soup-html does not really work nice
2007-06-09T05:41:39  <zenhase> *yawn*
2007-06-09T05:50:45  <zenhase> capturing the output works fine now i think, will fix up remaining (missing) parts of the themebase tomorrow and make a modern.html which will be included by the _base.html template
2007-06-09T05:51:27  <zenhase> and also do the promised wiki-page
2007-06-09T12:21:59  <ThomasWaldmann> moin
2007-06-09T12:22:08  <lanius> hi thomas
2007-06-09T12:23:46  <ThomasWaldmann> lanius: did you see my new fs storage proposal on the wiki?
2007-06-09T12:56:55  <lanius> on the wiki?
2007-06-09T13:01:27  <ThomasWaldmann> yes
2007-06-09T13:02:20  <ThomasWaldmann> http://moinmo.in/MoinMoinTodo/Release_2.0/OpenProblems?action=diff&rev2=21&rev1=18
2007-06-09T13:05:11  <lanius> i think separate metdata revisioning is a good idea because then we also eliminate the problem of transactional corruptions because each metadata change creates a new metarev
2007-06-09T13:05:14  <lanius> xorAxAx: what do you think
2007-06-09T13:07:35  <CIA-27> moin: Thomas Waldmann <tw AT waldmann-edv DOT de> * 2051:069a52eafe09 1.7-accessibility/ (8 files in 2 dirs): added insertion and deletion formatter methods, fixed small <th> issues, rearrange similar formatter method near each other
2007-06-09T13:11:38  <grzywacz> morning
2007-06-09T13:11:50  <grzywacz> xorAxAx: did you take a look at that last commit?
2007-06-09T13:12:07  <grzywacz> If so, please comment on it briefly. ;) I'll be back in a few hours. :)
2007-06-09T13:31:36  <xorAxAx> lanius: yes, its ok
2007-06-09T13:31:49  <xorAxAx> grzywacz: which one?
2007-06-09T13:32:48  <xorAxAx> grzywacz: you didnt push anyhting after my last comments
2007-06-09T14:22:41  <CIA-27> moin: Heinrich Wendel <h_wendel@cojobo.net> * 2063:f3dabaa1c918 1.7-storage-hwendel/MoinMoin/ (6 files in 3 dirs): implement DataBackend
2007-06-09T14:32:47  <ThomasWaldmann> lanius: +            if len(revs) > 0:           >    if revs:
2007-06-09T14:33:55  <lanius> k
2007-06-09T14:35:04  <zenhase> oh, the moinmo.in-domain is in place now?
2007-06-09T14:36:00  <ThomasWaldmann> lanius: \ No newline at end of file
2007-06-09T14:36:19  <ThomasWaldmann> zenhase: good morning :)
2007-06-09T14:36:54  <zenhase> ThomasWaldmann: actually i am already awake for 2hours :>
2007-06-09T14:37:01  <lanius> ThomasWaldmann: can't see which one
2007-06-09T14:37:03  <zenhase> but thanks and good morning to you too
2007-06-09T14:37:34  <ThomasWaldmann> lanius: read your diffs
2007-06-09T14:37:45  <zenhase> now i will have to aquire a birthday present for a friend
2007-06-09T14:37:48  <zenhase> bbl
2007-06-09T14:37:52  <ThomasWaldmann> zenhase: it is online since 2 weeks or so
2007-06-09T14:38:22  <zenhase> hrhr ... i just remember the conversation if to reg it or not
2007-06-09T14:38:37  <zenhase> didn't know it went online in an instant :>
2007-06-09T14:38:54  * ThomasWaldmann decided rather fast the gag is worth 19eur/a :)
2007-06-09T14:39:08  <zenhase> not only the gag :>
2007-06-09T14:39:22  <zenhase> it's a domain you can tell people in RL and they remember it ;>
2007-06-09T14:39:40  <ThomasWaldmann> yeah
2007-06-09T14:39:53  <xorAxAx> except if their native language isnt german :)
2007-06-09T14:40:01  <zenhase> (it's funny how search-engine resistant some people are, you always have to tell them urls)
2007-06-09T14:40:03  <xorAxAx> then you have to spell at least
2007-06-09T14:40:47  <ThomasWaldmann> lanius: +            revs = [rev for rev in revs if rev[-3:] != "tmp"]    > ... not rev.endswith(".tmp") ... (the dot is correct?)
2007-06-09T14:41:52  <zenhase> should be faster with .endswith
2007-06-09T14:42:18  <ThomasWaldmann> and please merge that map/lambda into that construction, without using map/lambda but list comprehension
2007-06-09T14:42:21  <zenhase> the first one will trigger strcpy in the c-code below i suppose
2007-06-09T14:42:46  * ThomasWaldmann unplugs
2007-06-09T14:48:31  <ThomasWaldmann> lanius: fs_moin16.py:55 is missing a TODO I guess?
2007-06-09T14:53:52  <ThomasWaldmann> lanius: the StorageErrors you raise sometimes can obfuscate the real reason of failure
2007-06-09T14:54:27  <xorAxAx> lanius: dont forget to provide the cause, see how config errors are handled
2007-06-09T15:00:10  <ThomasWaldmann> fs_moin16:127 tells Item ... does not exist. It should tell something about the type of item and operation and the cause of failure.
2007-06-09T15:00:46  <ThomasWaldmann> e.g. UserItem.remove_item failed: Permission denied
2007-06-09T15:01:19  <ThomasWaldmann> e.g. UserItem.remove_item failed: Permission denied (/var/moin/data/user/1234.56.78)
2007-06-09T15:03:25  <ThomasWaldmann> lanius: still alive or hit by lightning? :)
2007-06-09T15:03:42  <lanius> hit by lightning ;) no just cooking
2007-06-09T15:04:26  <ThomasWaldmann> (here is a rather intense thunderstorm going on)
2007-06-09T15:04:47  <lanius> it's coming to here soon
2007-06-09T15:08:04  <ThomasWaldmann> fs_moin16:149    149            if line[0] == '#':    > if line.startswith('#'):     and you maybe also want to catch empty lines before
2007-06-09T15:14:25  <ThomasWaldmann> 283            elif started == True:    >   elif started:
2007-06-09T15:14:52  <lanius> or even started is True ?
2007-06-09T15:19:32  <ThomasWaldmann> that would only make sense if you use True and 1 and ... and you only want to check for True and other stuff shall behave different
2007-06-09T15:19:55  <ThomasWaldmann> (which would be rather weird :)
2007-06-09T15:20:44  <ThomasWaldmann> os.path.join(path, name, "revisions", get_rev_string(revno))  < that should go into some function
2007-06-09T15:21:21  <ThomasWaldmann> (see old code)
2007-06-09T15:23:51  * ThomasWaldmann guesses PageData implementation will make people migrate rather fast to the still-to-be-implemented fs_moin17 backend :D
2007-06-09T15:25:03  <lanius> that was the intention, wasn't it?
2007-06-09T15:25:18  <lanius> just kidding, why?
2007-06-09T15:27:44  <ThomasWaldmann> because it creates and removes .tmp files even if you dont write data
2007-06-09T15:29:09  <ThomasWaldmann> and it will cause trouble doing that, btw
2007-06-09T15:33:32  <ThomasWaldmann> maybe make read_file and write_file properties
2007-06-09T15:33:50  <ThomasWaldmann> update self.written and self.read
2007-06-09T15:34:19  <ThomasWaldmann> close them only if they were used
2007-06-09T15:37:32  <ThomasWaldmann> lanius: look at MoinMoin.util.filesys.rename
2007-06-09T15:37:48  <ThomasWaldmann> and at caching.py
2007-06-09T15:45:50  <lanius> do we really need the rename method in filesys.py?
2007-06-09T15:48:47  <lanius> i'm using shutil.move but i don't know if it is atomic
2007-06-09T15:49:05  <ThomasWaldmann> it's a non-atomic posix emulation for win32
2007-06-09T15:49:18  <ThomasWaldmann> python stdlib does not do that
2007-06-09T15:49:53  <lanius> basically it is the same as your move
2007-06-09T15:49:58  <lanius> so for unix it is atomic
2007-06-09T15:50:43  <lanius> what about the dircache problems, do they still exist?
2007-06-09T15:51:23  <ThomasWaldmann> i think so
2007-06-09T15:51:40  <ThomasWaldmann> (as long as windows exists :)
2007-06-09T15:52:32  <lanius> but i can't find any reference to such problems other than linking to moin
2007-06-09T15:53:15  <ThomasWaldmann> we sometimes find new problems in stdlib
2007-06-09T15:53:48  <ThomasWaldmann> or discover strange win32 problems
2007-06-09T15:54:10  <lanius> on my windows system the problem seems to not exist
2007-06-09T15:55:39  <lanius> the implementation seems quite simple...
2007-06-09T15:56:00  <lanius> not much to do wrong in 20 lines of code
2007-06-09T16:00:51  <ThomasWaldmann> lanius: shutil.move will copy the file, even if it is on same filesystem
2007-06-09T16:01:24  <ThomasWaldmann> for bigger files, this is a big difference
2007-06-09T16:02:10  <ThomasWaldmann> and looking at concurrent access, it is much worse than what our rename does on win32
2007-06-09T16:02:16  <lanius> but only on windows
2007-06-09T16:54:52  <ThomasWaldmann> moin neagulm
2007-06-09T16:57:42  <lanius> ThomasWaldmann: the problem is the st_mtime which dircache uses
2007-06-09T16:57:46  <lanius> "Note: The exact meaning and resolution of the st_atime, st_mtime, and st_ctime members depends on the operating system and the file system. For example, on Windows systems using the FAT or FAT32 file systems, st_mtime has 2-second resolution, and st_atime has only 1-day resolution. See your operating system documentation for details."
2007-06-09T16:58:55  <ThomasWaldmann> neagulm: +    oldText = pg.get_raw_body()
2007-06-09T16:58:56  <ThomasWaldmann> +    newText = oldText+" * "+pagename+"\n"
2007-06-09T16:58:56  <ThomasWaldmann> +    pg.saveText(newText, rev)
2007-06-09T16:59:07  <ThomasWaldmann> text = ...
2007-06-09T16:59:39  <ThomasWaldmann> text += " * %s\n" % pagename
2007-06-09T16:59:41  <ThomasWaldmann> ...
2007-06-09T17:00:06  <neagulm> hello ThomasWaldmann
2007-06-09T17:04:56  <grzywacz> xorAxAx: no, but you commented on imports and pep8 only.
2007-06-09T17:06:24  <ThomasWaldmann> lanius: that doesnt explain why it does not work at all
2007-06-09T17:08:02  <ThomasWaldmann> neagulm: +                        elif (pageClass == "Spam" and not isMemberOf(request, self.page_name, "HamPages")):
2007-06-09T17:08:12  <ThomasWaldmann> try to avoid C-style bracketing
2007-06-09T17:09:08  <ThomasWaldmann> neagulm: +        if not cache is None:
2007-06-09T17:09:33  <ThomasWaldmann> if cache can either be None or a cache object, you can simply use if cache: here.
2007-06-09T17:10:52  <neagulm> already fixed while debugging. I have some big headaches because of SB...
2007-06-09T17:12:21  <ThomasWaldmann> neagulm: +                        url = '%s?action=togglespamstatus' % wikiutil.quoteWikinameURL(self.page_name)
2007-06-09T17:12:40  <ThomasWaldmann> that won't work for many cases, see the Page class for better ways
2007-06-09T17:14:27  <ThomasWaldmann> neagulm: -            mt = wikiutil.MimeType(filename=filename)
2007-06-09T17:14:27  <ThomasWaldmann> +            mt = wikiutil.MimeType(filename = filename)
2007-06-09T17:14:53  <ThomasWaldmann> function calls are one of the cases where no spaces are used around =
2007-06-09T17:17:36  <ThomasWaldmann> neagulm: instead of creating a string representation before and eval() after using the cache object, you can simply specify the use_pickle flag for the cache object and it will do it for you
2007-06-09T17:20:57  <ThomasWaldmann> did somebody use pycallgraph on moin source yet?
2007-06-09T17:34:20  <CIA-27> moin: Thomas Waldmann <tw AT waldmann-edv DOT de> * 2049:94af8c2afeb0 1.7/MoinMoin/logfile/__init__.py: make logfile damage happen less often
2007-06-09T17:43:45  <CIA-27> moin: Heinrich Wendel <h_wendel@cojobo.net> * 2064:0f6f100f52f7 1.7-storage-hwendel/MoinMoin/storage/ (8 files in 3 dirs): cleanup error handling; implement handling of current revision at the correct level; use MoinMoin.util.filesys
2007-06-09T17:50:16  <CIA-27> moin: Marian Neagul marian@info.uvt.ro * 2022:ae5ff368fe07 1.7-classify-mneagul/ (6 files in 5 dirs): Caching update and fix the action problem.
2007-06-09T17:55:48  <CIA-27> moin: Thomas Waldmann <tw AT waldmann-edv DOT de> * 2065:967d529b31e9 1.7-storage-hwendel/ (MoinMoin/logfile/__init__.py wiki/data/edit-log): workaround for initially non-existing edit-log, remove edit-log from repo
2007-06-09T17:55:49  <xorAxAx> grzywacz: besides that, it was ok
2007-06-09T17:55:50  <CIA-27> moin: Reimar Bauer <rb.proj AT googlemail DOT com> * 2066:b3ddb4504ec6 1.7-storage-hwendel/MoinMoin/parser/text_moin_wiki.py: text_moin_wiki: bug fix of {{{ aaa {{{\n }}} bbb }}}
2007-06-09T17:55:50  <CIA-27> moin: Reimar Bauer <rb.proj AT googlemail DOT com> * 2067:1bdbf54a33ff 1.7-storage-hwendel/MoinMoin/parser/text_moin_wiki.py: text_moin_wiki:bug fix of NoURLsAfterBlock
2007-06-09T17:55:52  <CIA-27> moin: Alexander Schremmer <alex AT alexanderweb DOT de> * 2068:e5526f9c2113 1.7-storage-hwendel/MoinMoin/_tests/ (_test_template.py test_security.py): Fixed test_template, streamlined reimars last test.
2007-06-09T17:55:54  <CIA-27> moin: Alexander Schremmer <alex AT alexanderweb DOT de> * 2069:8ac1efc21728 1.7-storage-hwendel/ (10 files in 8 dirs): Merge with main.
2007-06-09T17:55:57  <CIA-27> moin: Thomas Waldmann <tw AT waldmann-edv DOT de> * 2070:94af8c2afeb0 1.7-storage-hwendel/MoinMoin/logfile/__init__.py: make logfile damage happen less often
2007-06-09T17:56:00  <CIA-27> moin: Heinrich Wendel <h_wendel@cojobo.net> * 2071:8de877c92e2e 1.7-storage-hwendel/ (5 files in 4 dirs): merge with 1.7 head
2007-06-09T18:09:52  <ThomasWaldmann> lanius: PLEASE READ YOUR DIFFS BEFORE COMMITTING
2007-06-09T18:10:15  <lanius> why, it's a branch
2007-06-09T18:10:39  <ThomasWaldmann> hg diff | less
2007-06-09T18:11:20  <ThomasWaldmann> http://hg.thinkmo.de/moin/1.7-storage-hwendel/rev/0f6f100f52f7 or see here for last commit
2007-06-09T18:12:32  <lanius> looks good, what is the problem?
2007-06-09T18:14:07  <ThomasWaldmann> read it again
2007-06-09T18:14:57  <lanius> other than that it says "\ No newline at end of file" although there is a newline i don't see anything
2007-06-09T18:18:07  <lanius> ?
2007-06-09T18:19:36  <ThomasWaldmann> no, there is no newline at the end
2007-06-09T18:20:22  <lanius> my editor shows one with the exception of current, but why would you need one there
2007-06-09T18:20:43  <ThomasWaldmann> which editor are you using?
2007-06-09T18:20:47  <lanius> eclipse
2007-06-09T18:21:44  <ThomasWaldmann> the file ends with some spaces
2007-06-09T18:21:58  <lanius> seems to be some kind of auto indent feature
2007-06-09T18:23:51  <lanius> so why does it matter at all
2007-06-09T18:24:35  <ThomasWaldmann> it does matter because it annoys one when you review the diffs
2007-06-09T18:24:51  <lanius> so fix the diff tool
2007-06-09T18:25:10  <ThomasWaldmann> and it shouldnt be too hard to have a file ending with \n
2007-06-09T18:25:35  <lanius> seems to be very hard
2007-06-09T18:27:28  <ThomasWaldmann> no, it is not hard. you just do a hg diff | less before you commit. Then you will notice easily all those small things (and be reviewing it yourself, you also notice some other small errors).
2007-06-09T18:27:39  <ThomasWaldmann> s/be/by/
2007-06-09T18:28:07  <lanius> newlines at the end of a file is nothing a developer has to care for
2007-06-09T18:29:05  <ThomasWaldmann> well, it just annoys me when reviewing your stuff. enough said?
2007-06-09T18:29:17  <lanius> yeah, then the tool needs to be fixed
2007-06-09T18:37:27  <grzywacz> xorAxAx: ok, thanks
2007-06-09T18:59:21  * ThomasWaldmann needs a bigger screen :D
2007-06-09T19:15:50  <lanius> doesn't help, trust me
2007-06-09T22:28:47  <CIA-27> moin: Thomas Waldmann <tw AT waldmann-edv DOT de> * 2098:8e463da7f360 1.6/MoinMoin/action/diff.py: diff action: fix UnboundLocalError caused by invalid rev2 argument
2007-06-09T22:38:56  <dreimark> moin
2007-06-09T23:14:36  <ThomasWaldmann> hi dreimark
2007-06-09T23:19:39  <dreimark> hi ThomasWaldmann, call graph is a nice tool, but the image size gets my konqueror version down, I thought such old bugs were already solved
2007-06-09T23:22:02  <ThomasWaldmann> firefox or gimp :)
2007-06-09T23:28:40  <zenhase> oh boy
2007-06-09T23:29:01  <zenhase> pycallgraph is something i was looking for! :>
2007-06-09T23:32:16  <dreimark> ThomasWaldmann: some browsers could have some vulnerability reading large images (some years ago this was used for attacks)
2007-06-09T23:44:21  <ThomasWaldmann> dreimark: if your browser crashes, you may have discovered one :)
2007-06-09T23:46:14  <zenhase> there are .dot-file viewers to, as i can see
2007-06-09T23:46:17  <zenhase> too
2007-06-09T23:46:38  <dreimark> the browser eats all cpu and I have had to kill it :( unfortunately you have no idea if a large image is there
2007-06-09T23:46:55  <zenhase> perhaps they are better for larger graphs
2007-06-09T23:47:11  <dreimark> I will check this with a newer konqueror the next days