MoinMoin: MoinMoinChat/Logs/moin-dev/2009-05-11

2009-05-11T00:48:21  *** grzywacz has quit IRC
2009-05-11T00:58:42  <CIA-18> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 5742:ac8d2837479d 2.0-storage/MoinMoin/ (10 files in 3 dirs): theme.render_content can use old theme code or templating (cfg.templating=True)
2009-05-11T01:24:59  <ronny> ThomasWaldmann: seems like mpm is pissed off at the migration guides for moin
2009-05-11T01:25:12  <ronny> (and thus defers upgrading the hg wiki)
2009-05-11T01:26:01  <ronny> ThomasWaldmann: and they are considering to migrate to move to something "that doesn't break every release"
2009-05-11T01:53:30  <ThomasWaldmann> i have offered mpm about 3 times to help
2009-05-11T01:57:15  <ThomasWaldmann> and the site was broken for the last year or two because nobody with access cared
2009-05-11T01:58:03  <ThomasWaldmann> (cosmetically broken, minor theme issue)
2009-05-11T01:59:55  <ThomasWaldmann> ronny:
2009-05-11T02:07:40  <ronny> ThomasWaldmann: is basic standpoint is - if i have to maintain it on update, its to bad to keep
2009-05-11T02:08:38  <ThomasWaldmann> moin is software that is still in development and not yet dead
2009-05-11T02:09:29  <ronny> ThomasWaldmann: the perceived issue is, that someone has to manage migrations
2009-05-11T02:09:51  <ronny> and tbh, i hate those, i prefer dont care pacakges that just work, and do their magic
2009-05-11T02:10:44  <ThomasWaldmann> use sendmail :P
2009-05-11T02:11:41  <ronny> ThomasWaldmann: well, hg is a good example, i never ever needed to migrate a repo, while bzr on the other hand had dozens of those, monotone too
2009-05-11T02:12:02  <ronny> well, and svn too
2009-05-11T02:12:44  <ThomasWaldmann> if you have a good design from the beginning, that's nice
2009-05-11T02:13:11  <ronny> and git is another good example, they had various new pack formats, and even switched the object type headers in mid-run
2009-05-11T02:14:08  <ronny> ThomasWaldmann: the basic key is to keep the maintaince needs as low as possible, it can work i think
2009-05-11T02:14:13  <ronny> its just harder :/
2009-05-11T02:14:36  <ThomasWaldmann> i rather want stuff fixed than backwards compatible forever
2009-05-11T02:15:20  <ronny> yeah
2009-05-11T02:15:32  <ThomasWaldmann> and running a basic moin setup like mpm does is really no rocket science
2009-05-11T08:34:47  <dreimark> moin
2009-05-11T08:35:26  <dreimark> ronny: me has offered help too
2009-05-11T08:38:04  <dreimark> and I really can't understanding why calling moin migration data should be a problem on the hg wiki
2009-05-11T08:45:28  <ThomasWaldmann> moin
2009-05-11T08:46:26  <dreimark> ronny: there are only moins plugins used. So upgrading sounds like a 5min job.
2009-05-11T08:47:56  <dreimark> and it can only be longer because of thinking longer on file/pagenames with _.
2009-05-11T09:12:58  *** grzywacz has joined #moin-dev
2009-05-11T09:45:27  <ThomasWaldmann> dreimark: http://moinmo.in/4ct10n/diff/MoinMoinPatch/SetuidFixReplaceOsAccess you are introducing the same kind of code he is trying to get rid of
2009-05-11T10:00:33  *** waldi has quit IRC
2009-05-11T10:01:09  *** waldi has joined #moin-dev
2009-05-11T10:36:36  *** devilsadvocate has quit IRC
2009-05-11T11:14:50  <dreimark> \me wonders if this is correct `os.access(...W_OK...) does not report whether or not the current process is able to write`
2009-05-11T11:19:22  <dreimark> Using access() to check if a user is authorized to e.g. open a file before actually doing so using open() creates a security hole, because the user might exploit the short time interval between checking and opening the file to manipulate it.
2009-05-11T11:23:42  <dreimark> that is not happening here.
2009-05-11T11:36:08  *** devilsadvocate has joined #moin-dev
2009-05-11T12:42:15  <dennda> ThomasWaldmann: this is only about how acls/permissions are checked internally, not talking about the user action "revert" (and the like) I assume?
2009-05-11T12:42:21  <dennda> http://moinmo.in/4ct10n/info/FeatureRequests/AclRefactoring?action=diff&rev2=14&rev1=13
2009-05-11T12:45:15  <ThomasWaldmann> i won't kill the revert button on the history UI or the revert action triggered by it
2009-05-11T12:45:28  <dennda> good
2009-05-11T12:45:30  <ThomasWaldmann> but "revert" as a ACL capability is superfluous
2009-05-11T12:45:34  <dennda> yes
2009-05-11T12:45:36  <dennda> I agree
2009-05-11T12:45:56  <dennda> at least for revert it's clear that it's a combination of read and write
2009-05-11T12:47:40  <ThomasWaldmann> maybe leave some comment on the wiki page, so we all know what we agreed on in the end :)
2009-05-11T12:48:50  <dennda> delete / move / remove* require further thought
2009-05-11T12:48:59  <dennda> at least it's not as obvious
2009-05-11T12:50:52  <ThomasWaldmann> delete will depend a bit on whether we do just flag the item and leave it in place, or move it to Trash
2009-05-11T12:51:24  <ThomasWaldmann> or move it to Trash/$uniquenewname (and keep orig name in metadata)
2009-05-11T12:51:48  <dennda> well
2009-05-11T12:52:15  <dennda> these acls will apply for all items, so they will apply for images / binary data as well
2009-05-11T12:52:29  <dennda> am I correct here?
2009-05-11T12:54:56  <dreimark> yes
2009-05-11T12:55:19  <dennda> ok
2009-05-11T12:55:51  <dennda> A question we didn't discuss last year: What happens if someone uploads 30 gig of porn to some wiki?
2009-05-11T12:58:17  <ThomasWaldmann> you will make a copy of it and then delete it :D
2009-05-11T12:59:52  <ThomasWaldmann> and as I already noted last year, there should be some "nuke item" and "nuke revision" functionality
2009-05-11T13:02:02  <ThomasWaldmann> (at least for backend types who can support this, for others it could be "move to trash and make unaccessible")
2009-05-11T13:27:31  <dreimark> http://www.vimeo.com/3949077
2009-05-11T13:43:05  <ThomasWaldmann> btw, we could call it item_nuke_forever
2009-05-11T13:43:37  <ThomasWaldmann> (or Item.nuke_em_forever)
2009-05-11T14:20:34  *** vpv has quit IRC
2009-05-11T14:20:42  *** vpv has joined #moin-dev
2009-05-11T14:24:38  <dreimark> we need an acl right "no gui"
2009-05-11T14:34:42  <ThomasWaldmann> dreimark: for sure not :)
2009-05-11T16:41:48  <dreimark> can we archieve to move links by renaming pages ?
2009-05-11T16:55:30  *** devilsadvocate has quit IRC
2009-05-11T17:31:00  *** devilsadvocate has joined #moin-dev
2009-05-11T17:42:29  <ThomasWaldmann> dreimark: that problem didn't change in hardness and looking at the amount of other changes I would say "not now"
2009-05-11T19:36:11  *** grzywacz has quit IRC
2009-05-11T19:38:41  *** grzywacz has joined #moin-dev
2009-05-11T21:23:04  <ThomasWaldmann> re
2009-05-11T23:11:02  <CIA-18> Thomas Waldmann <tw AT waldmann-edv DOT de> default * 3866:512aeb66b980 1.7/MoinMoin/auth/http.py: http auth: do auth_type comparisons case-insensitive