MoinMoin: MoinMoinChat/Logs/moin-dev/2007-09-18

2007-09-18T02:08:03  <ThomasWaldmann> moin 1.6' http headers have duplicate stuff in them: cache-control, expires
2007-09-18T03:32:09  <ThomasWaldmann> hmm cia is sloooow
2007-09-18T11:20:58  <dreimark> moin
2007-09-18T12:26:48  <ThomasWaldmann> moin dreimark
2007-09-18T12:28:29  <dreimark> ThomasWaldmann: I do want to move fix {{{ stuff to 1.7 e.g. refactor nesting of {{{
2007-09-18T12:29:35  <ThomasWaldmann> move?
2007-09-18T12:31:10  <dreimark> MoinMoinTodo/Release_1.6.0 -> MoinMoinTodo/Release_1.7
2007-09-18T12:31:19  <dreimark> the last point
2007-09-18T12:34:10  <ThomasWaldmann> so what do you want to do with the code I copied from 1.7 to 1.6 recently? revert the {{{ changes in 1.6?
2007-09-18T12:36:38  <ThomasWaldmann> (I copied about the complete parser/formatter stuff from 1.7 to 1.6)
2007-09-18T12:40:27  <dreimark> no reverting, otherwise SyntaxReference is broken.
2007-09-18T12:41:34  <ThomasWaldmann> btw, iirc there is one problem left: it eats the space after {{{}}}
2007-09-18T12:43:14  <dreimark> that should be a small change (hopefully). I will look at this later.
2007-09-18T12:44:36  <dreimark> Do get it perfectly like creole the parser needs more changes and we should left that for 1.7
2007-09-18T12:48:09  <ThomasWaldmann> well, I am not too happy with the way creole does it (nor with the way we do it now)
2007-09-18T12:49:33  <ThomasWaldmann> but as I dont have a better solution, I just won't work on it :)
2007-09-18T12:51:44  <ThomasWaldmann> johill: see my latest change for session cookies in 1.6, comments?
2007-09-18T12:52:01  <ThomasWaldmann> did you read the debugging session with nir?
2007-09-18T12:52:02  <johill> no, haven't taken any look at anything
2007-09-18T12:52:16  <johill> I saw that you mentioned me but not more
2007-09-18T12:52:37  <johill> can you send me an email with a log or something so I can take a look tomorrow afternoon? I have an oral exam tomorrow morning
2007-09-18T12:53:23  <ThomasWaldmann> well, just look at the code. the problem was cause by a broken safari on OS X.
2007-09-18T12:53:35  <johill> ok
2007-09-18T12:53:38  <johill> looking
2007-09-18T12:53:46  <ThomasWaldmann> at the end, we found out that maybe renaming the expires= part would've been enough.
2007-09-18T12:54:10  <johill> weird
2007-09-18T12:54:12  <ThomasWaldmann> obviously safari confuses this with the other expires=
2007-09-18T12:54:21  <johill> why would it look into the cookie data?
2007-09-18T12:54:33  <ThomasWaldmann> bug :)
2007-09-18T12:55:07  <johill> hmm. are you sure that '.auth_username.encode('utf-8')' can never contain a colon?
2007-09-18T12:55:33  <johill> that was mostly the reason for using base64 so I didn't have to worry about colons in the strings :)
2007-09-18T12:56:41  <johill> I don't see a reason right now why a colon shouldn't be allowed in auth_username
2007-09-18T12:56:56  <ThomasWaldmann> ACLs :)
2007-09-18T12:57:18  <johill> but auth_username?
2007-09-18T12:57:45  <ThomasWaldmann> hm?
2007-09-18T12:57:56  <johill> auth_username can be totally different from username
2007-09-18T12:58:22  <johill> that's the whole point of having it
2007-09-18T12:59:07  <ThomasWaldmann> hmm. ok, maybe i just revert that changeset and rename expires=
2007-09-18T13:00:02  <ThomasWaldmann> (we first suspected it is caused by = in general, this is why I encoded the whole thing)
2007-09-18T13:00:16  <johill> I don't know what causes the bug but if that helps I think I'd prefer that since auth_username ought to be allowable to be set to anything imho
2007-09-18T13:00:30  <johill> auth_username is basically the internal identifier that the auth method uses
2007-09-18T13:01:06  <ThomasWaldmann> btw, it is no problem that we still have the uid in there?
2007-09-18T13:01:15  <johill> no
2007-09-18T13:01:18  <johill> the uid is useless now
2007-09-18T13:01:32  <johill> well, at least in 1.
2007-09-18T13:01:32  <johill> well, at least in 1.I should not see the unecrypted packet going out at all? Because if I ping
2007-09-18T13:01:35  <johill> A:eth1 from HB then I don't see the unencrypted echo-reply packet (which has
2007-09-18T13:01:38  <johill> the same source-address as the fragmentation needed) but only the outgoing
2007-09-18T13:01:39  <johill> huh sorry
2007-09-18T13:01:41  <johill> at least in 1.7
2007-09-18T13:01:54  <johill> hmm
2007-09-18T13:02:04  <johill> not entirely sure. does 1.6 have the xmlrpc session thing?
2007-09-18T13:02:29  <ThomasWaldmann> dont remember
2007-09-18T13:03:05  <johill> ok, it's not good in 1.6
2007-09-18T13:03:11  <johill> xmlrpc auth token is the u.id there
2007-09-18T13:05:45  <johill> that means we should look up the uid instead of having it in the cookie lest the cookie is intercepted by somebody
2007-09-18T13:06:03  <ThomasWaldmann> uuh
2007-09-18T13:06:42  <ThomasWaldmann> can you look into that stuff? you know your code better. :)
2007-09-18T13:07:52  <johill> not today, but yeah
2007-09-18T13:08:10  <johill> otoh, we only give the cookie to the user
2007-09-18T13:08:15  <johill> and they can already find out their uid
2007-09-18T13:08:28  <johill> and we currently do that in 1.5 too, iirc
2007-09-18T13:08:50  <ThomasWaldmann> yeah, but that is a known security problem
2007-09-18T13:09:54  <johill> yeah but it's not really fixable by just making the uid less visible
2007-09-18T13:10:02  <johill> it may mitigate it a bit
2007-09-18T13:10:51  <ThomasWaldmann> no, base64 encoding it doesnt really help :)
2007-09-18T13:11:07  <johill> well 1.7 has no problem
2007-09-18T13:11:11  <ThomasWaldmann> removing it from the cookie would help - if that is doable
2007-09-18T13:11:13  <johill> it doesn't use the u.id any more
2007-09-18T13:11:24  <johill> so in 1.7 you can publish the u.id as much as you wish
2007-09-18T13:11:46  <johill> maybe we should instead port the bits from xmlrpc from 1.7
2007-09-18T13:11:51  <johill> so that we don't use the u.id there
2007-09-18T13:12:29  <ThomasWaldmann> hmm, ok, then we maybe just continue living with it until 1.7
2007-09-18T13:13:19  <johill> I can see how hard it would be to backport the uid use in xmlrpc
2007-09-18T13:13:28  <johill> with the new cookie stuff we no longer rely on it in cookiee auth
2007-09-18T13:15:08  <johill> well we still use it but we sign it so it can't be used by somebody else
2007-09-18T13:15:19  <johill> so we could just do the same for xmlrpc like I did in 1.7 and the uid is useless
2007-09-18T13:17:20  <johill> anyway, I want to have lunch now and then I'm off till after my exam :)
2007-09-18T13:18:18  <ThomasWaldmann> ok, good luck in your exam :)
2007-09-18T13:20:56  <johill> thanks, later
2007-09-18T13:25:13  <xorAxAx> good luck johill
2007-09-18T13:29:43  <dreimark> good luck johill
2007-09-18T18:31:27  <ThomasWaldmann> dreimark: http://master.moinmo.in/HelpOnMacros see under "Writing your own macro" the strange rendering of All `Macro` instances ...
2007-09-18T18:34:06  <dreimark> arrgh, unwanted linebreak
2007-09-18T18:35:25  <ThomasWaldmann> http://master.moinmo.in/HelpOnMacros/MailTo same there, near the top
2007-09-18T18:40:17  <dreimark> If I try to Edit this page my browser tab gots frozen, info does work,
2007-09-18T18:43:03  <ThomasWaldmann> dreimark: shift-reload
2007-09-18T18:45:25  <dreimark> hmm, does not really work, I will restart the browser after some other debuggings
2007-09-18T18:52:36  <dreimark> browser restart solved it
2007-09-18T18:53:50  <dreimark> ThomasWaldmann: its the <<BR>><<BR>> which caused that line break
2007-09-18T18:54:00  <dreimark> that is removed its gone
2007-09-18T18:58:11  <ThomasWaldmann> dreimark: as a strange side effect you mean? at the place where it displayed a linebreak was no <<BR>>.
2007-09-18T18:58:29  <dreimark> yeah
2007-09-18T18:58:41  <dreimark> and the other is different
2007-09-18T18:59:52  <dreimark> a line break between both BR solves it too
2007-09-18T19:00:09  * ThomasWaldmann .oO(MarkupVoodoo)
2007-09-18T19:02:29  <dreimark> ThomasWaldmann:  see  if '>><<' in line.replace(' ', ''): may be I have broken it
2007-09-18T19:06:36  <dreimark> wow looks like I have fixed a cut and paste bug for tables from OOo to GUI
2007-09-18T19:07:26  <ThomasWaldmann> converter is a good place for bugfixing :)
2007-09-18T19:18:59  <ThomasWaldmann> gna
2007-09-18T19:26:47  <dreimark> ?
2007-09-18T19:27:34  <dreimark> cool that large table works, I try to add a test
2007-09-18T19:30:04  * ThomasWaldmann fixes ku i18n in the filesystem...
2007-09-18T19:32:00  <starshine> :)
2007-09-18T19:37:39  *** irc.freenode.net sets mode: +o ChanServ
2007-09-18T19:39:39  <ThomasWaldmann> looks like we have 5 new languages for the system texts
2007-09-18T19:48:40  <CIA-27> moin: Thomas Waldmann <tw AT waldmann-edv DOT de> default * 2178:505c9b42e0e3 1.6/MoinMoin/i18n/ (9 files): i18n: avoid crashing, log errors in po files, update i18n
2007-09-18T22:35:53  <grzywacz> evening
2007-09-18T22:36:09  <xorAxAx> hi grzywacz
2007-09-18T22:36:17  <grzywacz> hi xorAxAx
2007-09-18T22:36:18  <xorAxAx> how is your temporary flat?
2007-09-18T22:36:59  <grzywacz> One week more, at most. No idea if I'll have to spent another one till the end of September in 2nd temp place. ;-)))
2007-09-18T22:37:21  <xorAxAx> hehe
2007-09-18T22:37:41  <xorAxAx> as long as you dont have to unpack/pack all your belongings
2007-09-18T22:37:51  <grzywacz> I keep them stored in neat boxes.
2007-09-18T22:37:52  <grzywacz> :P
2007-09-18T22:37:59  <xorAxAx> yeah
2007-09-18T22:38:11  <grzywacz> It's annoying, but I can't help it atm.
2007-09-18T22:39:03  <xorAxAx> yeah
2007-09-18T22:58:25  <grzywacz> http://www.ex-parrot.com/pdw/Mail-RFC822-Address.html ;-)
2007-09-18T22:59:47  <xorAxAx> old
2007-09-18T23:01:44  <grzywacz> Sorry. :P
2007-09-18T23:02:10  <grzywacz> I know it's so 2002, but certainly *some* people haven't seen it... yet.
2007-09-18T23:03:27  <xorAxAx> and i assume its still wrong for addresses that can be found in daily use
2007-09-18T23:04:53  <TheSheep> most clients will accept unescaped accented characters, for example
2007-09-18T23:05:28  <TheSheep> like http://ść.pl  :)
2007-09-18T23:17:24  <grzywacz> Or żółw.pl
2007-09-18T23:36:30  <CIA-27> moin: Thomas Waldmann <tw AT waldmann-edv DOT de> default * 2179:5ff478aa6fdb 1.6/ (5 files in 5 dirs): use a image as placeholder for non-existing embedded attachment images