2006-08-07T01:25:33 * ThomasWaldmann put cache_dir on tmpfs
2006-08-07T04:13:52 * starshine sees hereslef highlighted on backscroll...
2006-08-07T04:14:07 <starshine> hrmm, I was away in the desert out of computing range all weekend
2006-08-07T04:14:15 <starshine> sorry I didn't see it sooner.
2006-08-07T04:16:53 <starshine> moooooooin first on todo list after unpacking the car :)
2006-08-07T06:53:37 *** irc.freenode.net sets mode: +o ChanServ
2006-08-07T10:41:21 <xorAxAx> Kepplar: that is storage
2006-08-07T10:41:29 <xorAxAx> what?
2006-08-07T10:41:31 <Kepplar> xorAxAx: its been designed like this from day 1
2006-08-07T10:42:15 <Kepplar> I have no idea what plans are for handling multiname spaces
2006-08-07T10:42:54 <Kepplar> storage will support it in some form (engine for each namespace) but its a far larger design consideration than just for storage to take.
2006-08-07T10:43:52 <xorAxAx> i am not sure about the latter assertion
2006-08-07T10:44:28 <Kepplar> Well, I just think this has such a big effect everyone should be working/deciding on it
2006-08-07T10:44:55 <xorAxAx> normally, the last decision is made by the person implementing it
2006-08-07T10:45:54 <Kepplar> I signed up to create a storage engine within these twelve weeks, not to re design's moin's internal architecture that happens to use storage :P. The latter task will take till Christmas
2006-08-07T10:46:25 <ffesti> moin
2006-08-07T10:46:28 <Kepplar> Hi
2006-08-07T10:46:46 <xorAxAx> hi ffesti
2006-08-07T11:02:46 <fpletz> moin
2006-08-07T11:07:54 <Kepplar> moin
2006-08-07T11:08:37 <ThomasWaldmann> ok, lets begin (sorry, had customer on the phone)
2006-08-07T11:08:52 <ThomasWaldmann> Kepplar: your time
2006-08-07T11:08:56 <Kepplar> Cool
2006-08-07T11:09:28 <ThomasWaldmann> (moin everybody, btw :)
2006-08-07T11:09:33 <Kepplar> Ok, so I've now been on full scale intergration, its going well and Page should be done minus any bugs I'm currently ironing out.
2006-08-07T11:09:50 <Kepplar> Next stage is to get PE, PGE and user done this week.
2006-08-07T11:10:02 <ThomasWaldmann> "should be done" means it work mostly in practice?
2006-08-07T11:10:25 <Kepplar> Well it's integrated, I can't see its all working until I got the tests working (ironed out bugs)
2006-08-07T11:10:41 <ThomasWaldmann> ok, so it is NOT done
2006-08-07T11:10:48 <Kepplar> PE, PGE and user
2006-08-07T11:11:00 <Kepplar> and ACL (although this isnt that big)
2006-08-07T11:11:17 <ThomasWaldmann> page working means you can see some wiki pages running with your code
2006-08-07T11:11:22 <Kepplar> PE/PGE only have minor changes to work with Page superclass
2006-08-07T11:11:35 <Kepplar> as i have said twice already not until I get the bugs out :)
2006-08-07T11:12:01 <Kepplar> but yes technically it should be working
2006-08-07T11:12:27 <Kepplar> The main problem at the moment is how to handle the so called "mimetype-metadata" cache
2006-08-07T11:12:33 <Kepplar> for legacy system
2006-08-07T11:12:56 <Kepplar> as you expressed you didnt want it implemented
2006-08-07T11:13:07 <Kepplar> Which leaves a massive problem since I've already coded it in
2006-08-07T11:13:20 <ThomasWaldmann> why do you need that "cache"? I got the impression that this is more than a cache, but authoritative data.
2006-08-07T11:13:45 <Kepplar> well yes, but it's effectivly a cache of the mimetype metadata
2006-08-07T11:14:22 <Kepplar> It is authoritative data though - used to work out where items are locationed (for example is it a page or a user)
2006-08-07T11:14:32 <Kepplar> located*
2006-08-07T11:15:11 <Kepplar> now there are two options
2006-08-07T11:15:21 <ThomasWaldmann> the current code KNOWS when it wants a user or when it wants some "page", so why do you need that?
2006-08-07T11:15:38 <Kepplar> Storage doesn't
2006-08-07T11:15:49 <ThomasWaldmann> it can be told
2006-08-07T11:15:54 <Kepplar> because its the same generic items
2006-08-07T11:16:46 <Kepplar> it makes no sense what-so-ever to have it in the API as types can change per revision and one is not expected to know
2006-08-07T11:17:08 <Kepplar> maybe not for legacy - but that is flatfiles problem to solve not the storage
2006-08-07T11:17:23 <ThomasWaldmann> what happens if your "cache" gets deleted or defective?
2006-08-07T11:18:10 <Kepplar> ThomasWaldmann: If we provide a script to search through all of data and pull out the mimetype then it can just be run
2006-08-07T11:18:40 <Kepplar> effectively regenerating it
2006-08-07T11:18:48 <ThomasWaldmann> that means the wiki will be broken until operator intervention
2006-08-07T11:19:10 <Kepplar> to be honest in six months time one will be mad if they are using legacy
2006-08-07T11:19:37 <Kepplar> Plus its no different from wikidata being corrupted
2006-08-07T11:20:03 <ThomasWaldmann> sure it is. a cache is expected to be deletable.
2006-08-07T11:20:21 * ThomasWaldmann has cache on tmpfs currently
2006-08-07T11:20:26 <Kepplar> the other option is to look at the traceback on every query into storage from flatfile and find out whats calling it, but its an intermediate hack as Page is now deprecated and funtionality in the future will migrate
2006-08-07T11:20:40 <Kepplar> ThomasWaldmann: cache is the wrong word then fine
2006-08-07T11:20:57 <Kepplar> its a central lookup folder
2006-08-07T11:21:44 <ThomasWaldmann> how about some kw arg for that, wouldn't that be much easier?
2006-08-07T11:22:01 <Kepplar> thats a massive hack on the API
2006-08-07T11:22:39 <Kepplar> and Storage's API should not be comprimised because of legacy flatfile's incompatbility
2006-08-07T11:22:55 <xorAxAx> this is not a legacy-problem
2006-08-07T11:23:03 <xorAxAx> it is a fundamental problem
2006-08-07T11:23:07 <Kepplar> xorAxAx: how so?
2006-08-07T11:23:20 <xorAxAx> how should the engine decide if you want to have the homepage or the user data?
2006-08-07T11:23:25 <xorAxAx> for a given name
2006-08-07T11:23:35 <Kepplar> xorAxAx: because there is only one object for a given name.
2006-08-07T11:23:53 <xorAxAx> then how do you want to store user data in a farm?
2006-08-07T11:24:42 <Kepplar> From day one this is how its been designed, bringing this problem up now is not helpful. This is fundermentally how Storage is designed
2006-08-07T11:25:19 <Kepplar> as far as I'm concerned this is a problem on how you use storage.
2006-08-07T11:26:01 <Kepplar> Options are hack it so you have page/subpage notation
2006-08-07T11:26:20 <ThomasWaldmann> we can have (or generate) a different name for user profile and home page, this is not a problem
2006-08-07T11:26:25 <Kepplar> I'm not sure what else, TW has been aware that everything has a single namespace
2006-08-07T11:26:49 <ThomasWaldmann> having duplication is some "cache" maybe is
2006-08-07T11:26:55 <xorAxAx> yeah, Users/ is completly possible
2006-08-07T11:27:01 <Kepplar> ThomasWaldmann: that works, or have the user object with page as an attachment?
2006-08-07T11:27:11 <xorAxAx> and on Users/, the user-legacy-storage-class is mounted
2006-08-07T11:27:36 <Kepplar> Remember that non-legacy might have proper hierarchal storage
2006-08-07T11:28:03 <Kepplar> the other option
2006-08-07T11:28:18 <Kepplar> is to create a separate engine for user, just as underlay and data are separate
2006-08-07T11:28:26 <xorAxAx> hierachical storage as a GUI concept or as an internal thing?
2006-08-07T11:28:34 <xorAxAx> Kepplar: i just said that
2006-08-07T11:28:35 <Kepplar> xorAxAx: internal
2006-08-07T11:28:47 <ThomasWaldmann> Kepplar: how about changing user.py code to request objects named "User/<id>" and then have your code choose right backend?
2006-08-07T11:29:08 <ThomasWaldmann> then you dont need your cache
2006-08-07T11:29:26 <Kepplar> ThomasWaldmann: again only a temporary hack as anything other than legacy should handle it correctly.
2006-08-07T11:29:37 <ThomasWaldmann> you just have different storage backends mounted at different places, like fstab
2006-08-07T11:29:59 <Kepplar> Ok
2006-08-07T11:30:06 <xorAxAx> Kepplar: that is a problem that is not related to legacy things
2006-08-07T11:30:15 <Kepplar> for SOC will proceed like that - aftewards we should discuss with further
2006-08-07T11:30:21 <xorAxAx> thats important to understand here
2006-08-07T11:30:35 <xorAxAx> i.e. a "modern" adapter would have the same problem
2006-08-07T11:30:38 <Kepplar> xorAxAx: is it a big problem not to have a shared namespace, can't pages and users be one and the same?
2006-08-07T11:30:47 <xorAxAx> Kepplar: yes, but the same problems remain
2006-08-07T11:31:17 <Kepplar> oh also, attachments
2006-08-07T11:31:25 <Kepplar> should we access attachments as if they are a subpage?
2006-08-07T11:31:27 <xorAxAx> i.e. if you store it as metadata, you could simply check for the legacy backend, if a user with that name exists and cache that
2006-08-07T11:31:57 <Kepplar> not following
2006-08-07T11:32:49 <xorAxAx> if the normal legacy backend gets a request for page A and is queried for metadata, it would check the existence of legacy user file for name A and return that data if it exists
2006-08-07T11:33:16 <xorAxAx> thats the implementation strategy if you want to attach the metadata of the user objekt to the wikihomepage
2006-08-07T11:33:28 <xorAxAx> which is not that nice, it was one of the disturbing things in twiki
2006-08-07T11:34:06 <Kepplar> Hmm
2006-08-07T11:34:14 <xorAxAx> so this is a semantics thing
2006-08-07T11:34:15 <Kepplar> debatable
2006-08-07T11:34:33 <Kepplar> Maybe mysite.com/MyUserName *should* return the user object?
2006-08-07T11:34:40 <Kepplar> again debatable :)
2006-08-07T11:34:47 <ThomasWaldmann> ok, i guess we have to continue that later.
2006-08-07T11:34:53 <Kepplar> ok
2006-08-07T11:34:54 <xorAxAx> Kepplar: thats pointless
2006-08-07T11:34:56 <Kepplar> cheers
2006-08-07T11:35:01 <ThomasWaldmann> fpletz: your time
2006-08-07T11:35:05 <Kepplar> xorAxAx: discuss this after soc ^^
2006-08-07T11:36:01 <fpletz> LAST: overall googlization: stats bar; a new green line for every search result with size, revision, date of last modification, ...; paging with graphics and design by google; index domain of a page (standard, underlay)
2006-08-07T11:36:19 <fpletz> NEXT: as search results page basically complete (without "advanced" features, grouping for example), advanced search ui with options for every kind of search a user can think of ;), performance tweaks for paging
2006-08-07T11:36:23 <fpletz> BLOCKERS: -
2006-08-07T11:36:26 <fpletz> as a sidenote: found a nice appartment this friday, so I will work every day now ;)
2006-08-07T11:37:25 <ThomasWaldmann> ah, hopefully this finally works...
2006-08-07T11:37:44 <ThomasWaldmann> did you integrate some M o i n images?
2006-08-07T11:37:54 <fpletz> yes
2006-08-07T11:38:23 <fpletz> and wrote en email to google if they approve.. just to be sure ;)
2006-08-07T11:39:18 <ThomasWaldmann> maybe also work on those "similar pages" (when not too much effort) or remove the link until it can work
2006-08-07T11:40:18 <fpletz> ok, I will implement that after other features are complete.. for example this history search would have a higher priority
2006-08-07T11:40:43 <fpletz> but some basic stuff is still missing
2006-08-07T11:43:45 <xorAxAx> should i continue?
2006-08-07T11:45:12 <fpletz> afaic, yes
2006-08-07T11:45:51 <xorAxAx> LAST: Load the IWID and the meta dict lazily. Reworked RemotePage/SyncPage, fixed option handling again, refined semantics of options, introduced direction option, replaced "localMatch"/"remoteMatch" by "pageMatch". Store mtime for InterWiki list updates and detect changes based on it. Added support for localPrefix and remotePrefix. Started writing the merge loop.
2006-08-07T11:45:56 <xorAxAx> NEXT: Finish the page list stuff in SyncPages and get merging working for the generated page lists.
2006-08-07T11:45:59 <xorAxAx> BLOCKERS: -
2006-08-07T11:48:14 <ThomasWaldmann> when is sync expected to basically work somehow?
2006-08-07T11:48:23 <ThomasWaldmann> "something visible"
2006-08-07T11:48:38 <xorAxAx> it should be doing its first steps on wednesday
2006-08-07T11:48:57 <xorAxAx> the first planned visible feature is syncing of pages that are in both wikis
2006-08-07T11:49:26 <xorAxAx> after that, i will complete the sync code to handle the other cases
2006-08-07T11:49:32 <ThomasWaldmann> will there be some UI except an action menu entry "start sync"?
2006-08-07T11:49:38 <xorAxAx> no
2006-08-07T11:49:52 <xorAxAx> i am not sure if it is necessary, what do you think?
2006-08-07T11:50:02 <xorAxAx> currently, you just have the parameters on the page
2006-08-07T11:50:06 <xorAxAx> like a job description
2006-08-07T11:50:22 <xorAxAx> then you can create a new job by using a special SyncTemplate
2006-08-07T11:50:32 <ThomasWaldmann> job page example url?
2006-08-07T11:50:34 <xorAxAx> which would contain some short explanations
2006-08-07T11:50:46 <xorAxAx> hmm, i have a page locally, i can put it into the sync wiki ...
2006-08-07T11:51:19 <ThomasWaldmann> did you already begin some docs about it?
2006-08-07T11:51:42 <xorAxAx> http://sync.wikiwikiweb.de/JobExample
2006-08-07T11:51:54 <xorAxAx> the docs are on the wikisync page
2006-08-07T11:52:00 <xorAxAx> with all available options etc.
2006-08-07T11:52:11 <xorAxAx> i.e. i will just copy parts of it and add some other explanations
2006-08-07T11:52:30 <ThomasWaldmann> to some Help* page on master wiki
2006-08-07T11:53:14 <xorAxAx> yes
2006-08-07T11:53:32 <ThomasWaldmann> what's the problem with pages not in both wikis?
2006-08-07T11:53:54 <xorAxAx> i currently cannot determine if they were deleted
2006-08-07T11:54:07 <xorAxAx> which would make special handling necessary
2006-08-07T11:54:28 <xorAxAx> it just sees that they are gone
2006-08-07T11:54:53 <ThomasWaldmann> ah, ok
2006-08-07T11:55:15 <xorAxAx> but thats solvable
2006-08-07T11:56:52 <ThomasWaldmann> ok, looking forward to see it working :)
2006-08-07T11:56:54 <xorAxAx> btw, i wrote my last exam for these months on thursday and so i am very free currently
2006-08-07T11:57:13 <xorAxAx> yeah, me too
2006-08-07T11:57:19 <xorAxAx> i would consolidate all my wikis :)
2006-08-07T11:58:10 <ThomasWaldmann> what happens if transfer gets interrupted while transferring a page?
2006-08-07T11:59:27 <xorAxAx> its completly atomical - i.e. if the remote change doesnt work (interrupted connection), then locally nothing is written either
2006-08-07T11:59:36 <xorAxAx> i am just missing the locking locking
2006-08-07T11:59:39 <xorAxAx> s/locking/local/
2006-08-07T12:00:00 <ThomasWaldmann> is the page data and the log transferred in one go?
2006-08-07T12:00:20 <xorAxAx> the log is not transferred
2006-08-07T12:00:30 <xorAxAx> it just creates one new revision on both sides
2006-08-07T12:00:41 <ThomasWaldmann> no edit comments?
2006-08-07T12:00:44 <xorAxAx> no
2006-08-07T12:00:58 <ThomasWaldmann> sometimes they are useful :)
2006-08-07T12:01:07 <xorAxAx> sure
2006-08-07T12:01:26 <xorAxAx> but then you have to bug the person that did the merge or look into the other wiki :)
2006-08-07T12:01:36 <xorAxAx> because the point is too ensure that there is just one branch per wiki
2006-08-07T12:01:46 <xorAxAx> otherwise it would have many parallel branches per page
2006-08-07T12:01:53 <ThomasWaldmann> and in future, the edit comment will be part of metadata and we will want to transfer metadata
2006-08-07T12:02:09 <xorAxAx> as i said, it is just transferring the diff, not a page
2006-08-07T12:02:30 <xorAxAx> thats one of the main design choices, explained in sentence 2 or 3 on the wikisync page :)
2006-08-07T12:02:32 <ThomasWaldmann> how about attachments?
2006-08-07T12:02:39 <xorAxAx> they are not covered yet
2006-08-07T12:02:52 <xorAxAx> because they will be items ...
2006-08-07T12:03:08 <ThomasWaldmann> ok
2006-08-07T12:03:09 <xorAxAx> so it should work automatically with items in that case
2006-08-07T12:03:26 <ThomasWaldmann> they also will need metadata transferred (e.g. mimetype)
2006-08-07T12:03:34 <xorAxAx> yes
2006-08-07T12:04:13 <xorAxAx> that would have to extended later on i fear
2006-08-07T12:04:20 <xorAxAx> but it shouldnt be too hard
2006-08-07T12:06:20 <fpletz> xorAxAx: do you update the xapian index when syncing pages? (i.e. PageEditor.saveText does it, if you use it)
2006-08-07T12:06:47 <xorAxAx> fpletz: yes, i will use public methods
2006-08-07T12:07:03 <xorAxAx> i want to have e-mail notifications etc. as well
2006-08-07T12:07:05 <fpletz> ok, great
2006-08-07T12:09:52 <ThomasWaldmann> ok, thanks for your reports. happy hacking. :)
2006-08-07T12:11:54 <fpletz> bbl, lunch
2006-08-07T12:31:39 <Kepplar> moin
2006-08-07T12:50:45 <Kepplar> ThomasWaldmann: we can't have it in separate engines, all of data must be in the same engine - unless I rewrite everything (which is not an option as its a big hack anyway)
2006-08-07T12:51:10 <Kepplar> The only option then is to prefix everything
2006-08-07T12:54:35 <Kepplar> I would prefer though if I knew a direction we are going to go in in regards to this in the future to minimise changes later
2006-08-07T12:54:51 <Kepplar> In my eyes at least the storage engines together should all store things of the same namespace
2006-08-07T12:56:01 <Kepplar> Namespace should be another object/class alltogether
2006-08-07T12:56:29 <Kepplar> engines are members of namespaces
2006-08-07T12:56:38 <Kepplar> can be in more than one prehaps
2006-08-07T12:57:00 <Kepplar> and moin-land (non-storage code) will access a namespace for a wiki name
2006-08-07T12:58:43 <Kepplar> and there should be a universal storage indicator I think too
2006-08-07T12:58:51 <Kepplar> Namespace/WikiName
2006-08-07T12:59:05 <Kepplar> (I smell a Rubik =) )
2006-08-07T13:10:06 <Kepplar> Hmm
2006-08-07T13:10:17 <Kepplar> my main problem is *seperating* pages and users from the same namespace
2006-08-07T13:10:30 <Kepplar> not technically
2006-08-07T13:10:45 <Kepplar> But my belief in the actual project was this was the desired effect
2006-08-07T13:10:51 <Kepplar> the entire project is geared towards this
2006-08-07T13:11:04 <Kepplar> changing it gives me major doubts over the direction of Moin
2006-08-07T13:11:33 <Kepplar> Its all wikidata, it should be datatype indepenedant
2006-08-07T13:11:42 <Kepplar> having a hierarchy based thing like user/username is very web1
2006-08-07T13:11:51 <Kepplar> which is not very wiki
2006-08-07T13:19:37 <Kepplar> Ooooooooo
2006-08-07T13:19:43 <Kepplar> i have a cool idea(not for now)
2006-08-07T13:19:50 <Kepplar> but have item revisions less than 1 work like arrays in python
2006-08-07T13:19:57 <Kepplar> so -1 will get second last revision
2006-08-07T13:19:59 <Kepplar> and so on
2006-08-07T13:23:41 <Kepplar> hmm
2006-08-07T13:23:45 <Kepplar> i dont see a way out
2006-08-07T13:23:49 <Kepplar> so Im stuck
2006-08-07T13:26:26 <xorAxAx> an industry-grade programmer is never stuck
2006-08-07T13:26:32 <xorAxAx> i.e. never gets stuck
2006-08-07T13:26:40 <Kepplar> I'm stuck because i have standards =)
2006-08-07T13:27:17 <Kepplar> If I hack it it will such a huge mess it will take this long to rework everything back how it should be
2006-08-07T13:27:23 <Kepplar> after soc
2006-08-07T13:29:46 <Kepplar> flatfile is simply incompatbile or best a big hack
2006-08-07T13:30:03 <Kepplar> thats of course going on the notion of a single namespace shared by users pages etc
2006-08-07T13:30:46 <Kepplar> If i put hacks for subpages as discussed in the meeting in flatfile and the front end, it will mean that other SI will break
2006-08-07T13:32:44 <Kepplar> unless we armour it internally with the "rootpage" in a subpage pair
2006-08-07T13:32:55 <Kepplar> so User/Name is created in users
2006-08-07T13:32:57 <Kepplar> sent down
2006-08-07T13:33:02 <Kepplar> armour removed
2006-08-07T13:33:05 <Kepplar> so its just Name
2006-08-07T13:33:11 <Kepplar> and each SI can handle it in its own way
2006-08-07T13:35:09 <Kepplar> hmm
2006-08-07T13:35:20 <Kepplar> only acceptable temp way is to get it to amour it IF it going to a flatfile SI
2006-08-07T13:35:39 <Kepplar> and get SI flatfile to unarmour it
2006-08-07T13:35:57 <Kepplar> yea
2006-08-07T13:36:00 <Kepplar> i think thats acceptable
2006-08-07T13:36:36 <Kepplar> but there will be a point where we'd need to pull this out and decide to how migrate namespaces
2006-08-07T13:36:44 <Kepplar> maybe that'll be what I work on after finishing this
2006-08-07T13:41:16 <Kepplar> xorAxAx: btw I'm not saying that user-data and user-homepage should be in a shared object - i just think that all namespaces should be generically constructed and are capable of holding multiple types of information
2006-08-07T13:41:36 <xorAxAx> yes, but then you have the addressing problem
2006-08-07T13:41:43 <Kepplar> xorAxAx: should there be a namespace object as I mentioned above I have no problem in the concept of separating them
2006-08-07T13:54:30 <Kepplar> btw do people realise MetaDicts will be going?
2006-08-07T16:03:36 <Kepplar> Hack should be in place now
2006-08-07T19:17:42 <xorAxAx> if someone sees editing conflicts that are not merged well by moinmoin, please save the text that you submitted (go back on seeing the edit conflict with your browser, ^C and then go forward and submit) and give me the url and your text
2006-08-07T19:18:10 <xorAxAx> this is unrelated to wikisync, but maybe i can enhance the merging code in moin a bit
2006-08-07T20:24:04 <Kepplar> return self.access_storage(str_location, (lambda fp_revision_file: self.parse_metadata(fp_revision_file, metadata_field)), "r")
2006-08-07T20:24:08 <Kepplar> File "/home/kepplar/dev/moinmoin/storage/MoinMoin/storage/formats/flatfile.py", line 339, in parse_metadata
2006-08-07T20:24:11 <Kepplar> line = fp_metadata.readline()
2006-08-07T20:24:13 <Kepplar> KeyboardInterrupt
2006-08-07T20:24:16 <Kepplar> make: *** [test] Error 1
2006-08-07T20:24:18 <Kepplar> make test is hanging here
2006-08-07T20:26:18 <Kepplar> also i cant seem to commit
2006-08-07T20:26:26 <Kepplar> its saying there are no changes
2006-08-07T20:26:35 <Kepplar> status is blank so is diff
2006-08-07T20:26:37 <ThomasWaldmann> moin
2006-08-07T20:26:47 <Kepplar> moin
2006-08-07T20:30:21 <ThomasWaldmann> from MoinMoin.storage import storage, flatfile
2006-08-07T20:30:21 <ThomasWaldmann> ImportError: cannot import name flatfile
2006-08-07T20:32:01 <ThomasWaldmann> your code is not pep8
2006-08-07T20:32:40 <Kepplar> what modules that in?
2006-08-07T20:35:05 <Kepplar> as flatfile's not even there anymore
2006-08-07T20:35:15 <Kepplar> any idea why i cannot commit/
2006-08-07T20:35:17 <Kepplar> ?
2006-08-07T20:35:52 <ThomasWaldmann> make test
2006-08-07T20:36:38 <xorAxAx> Kepplar: try `hg id`, `hg verify`, `hg heads` and check the current state
2006-08-07T20:37:35 * ThomasWaldmann put a pep8 checker in pre-commit. it rejects code that uses == None as has too few blanks.
2006-08-07T20:44:40 <xorAxAx> LOL
2006-08-07T20:45:11 <xorAxAx> isnt pre-commit local?
2006-08-07T20:45:22 <xorAxAx> prenxtchangegroup might be the right one
2006-08-07T20:46:14 <xorAxAx> but i am not sure
2006-08-07T20:46:27 <xorAxAx> Kepplar: and what do you get for those commands?
2006-08-07T20:53:45 <Kepplar> ThomasWaldmann: your running a very old version then :)
2006-08-07T20:54:22 <ThomasWaldmann> no changes found
2006-08-07T20:54:27 <Kepplar> hg id
2006-08-07T20:54:28 <Kepplar> 85237c6edba2 tip
2006-08-07T20:55:11 <Kepplar> kepplar@corus:~/dev/moinmoin/storage$ hg verify
2006-08-07T20:55:11 <Kepplar> repository uses revlog format 0
2006-08-07T20:55:11 <Kepplar> checking changesets
2006-08-07T20:55:11 <Kepplar> checking manifests
2006-08-07T20:55:11 <Kepplar> crosschecking files in changesets and manifests
2006-08-07T20:55:14 <Kepplar> checking files
2006-08-07T20:55:16 <Kepplar> 2389 files, 1076 changesets, 7166 total revisions
2006-08-07T20:55:40 <Kepplar> still nothing
2006-08-07T20:55:46 <xorAxAx> Kepplar: and why do you think that you have to commit anything?
2006-08-07T20:55:54 <xorAxAx> are you sure that you modified files?
2006-08-07T20:56:02 <Kepplar> *heavily*
2006-08-07T20:56:13 <Kepplar> there are a fair few changes!
2006-08-07T20:56:19 <Kepplar> including the hack around cache (which works)
2006-08-07T20:57:41 <xorAxAx> you didnt push your last commits
2006-08-07T20:57:49 <Kepplar> kepplar@corus:~/dev/moinmoin/storage$ hg push
2006-08-07T20:57:50 <Kepplar> pushing to ssh://hg@hg.thinkmo.de/moin/1.6-storage-a
2006-08-07T20:57:50 <Kepplar> searching for changes
2006-08-07T20:57:50 <Kepplar> abort: unsynced remote changes!
2006-08-07T20:57:50 <Kepplar> (did you forget to sync? use push -f to force)
2006-08-07T20:58:01 <Kepplar> ah
2006-08-07T20:58:04 <Kepplar> TW's update
2006-08-07T20:58:57 <Kepplar> ok it worked :)
2006-08-07T21:00:13 <Kepplar> so.. how expensive is readline()?
2006-08-07T21:00:35 <Kepplar> as I'm readline()-ing SyntaxReference and its hanging
2006-08-07T21:01:28 <ThomasWaldmann> Kepplar: if you insist on reinventing the wheel, at least use same format as wikiutil.MetaDict
2006-08-07T21:01:42 <Kepplar> format of?
2006-08-07T21:04:26 <xorAxAx> Kepplar: a deadlock has the price "infinity"
2006-08-07T21:05:43 <Kepplar> lol well yes
2006-08-07T21:09:27 <Kepplar> ah
2006-08-07T21:09:32 <Kepplar> my indentings off
2006-08-07T21:09:40 <Kepplar> its in a while True ><
2006-08-07T21:11:17 <Kepplar> Bah that method was bad, I must have been drinking when i wrote that
2006-08-07T21:18:46 <Kepplar> ThomasWaldmann: i just got rid of all of ==None and the like just for you :P
2006-08-07T21:19:12 <xorAxAx> Kepplar: did you set your name as well?
2006-08-07T21:19:22 <Kepplar> xorAxAx: that keeps breaking :/
2006-08-07T21:19:33 <Kepplar> as I'm having to clone often
2006-08-07T21:19:40 <xorAxAx> ?
2006-08-07T21:19:52 <Kepplar> i remove the directory and reclone
2006-08-07T21:19:56 <Kepplar> due to mess ups on local
2006-08-07T21:20:00 <xorAxAx> the name has to be set in the global hgrc
2006-08-07T21:20:14 <xorAxAx> which is docs and settings\*\mercurial.ini or ~/.hgrc
2006-08-07T21:20:14 <Kepplar> ~/.hgrc?
2006-08-07T21:20:27 <xorAxAx> furthermore, you should not mess with your incoming branch
2006-08-07T21:20:36 <xorAxAx> i.e. only clone locally
2006-08-07T21:23:35 <Kepplar> Ok I'm confused
2006-08-07T21:23:43 <Kepplar> metadata seems to be implimented one way
2006-08-07T21:23:54 <Kepplar> and the docs at MetaDataInMoin point to different syntax
2006-08-07T21:24:36 <Kepplar> hmmm
2006-08-07T21:25:26 <Kepplar> ##master-page:Unknown-Page
2006-08-07T21:25:26 <Kepplar> ##master-date:Unknown-Date
2006-08-07T21:25:26 <Kepplar> #acl MoinPagesEditorGroup:read,write,delete,revert All:read
2006-08-07T21:25:26 <Kepplar> #format wiki
2006-08-07T21:25:26 <Kepplar> #language en
2006-08-07T21:25:28 <Kepplar> or not
2006-08-07T21:25:35 <Kepplar> first two lines are comments?
2006-08-07T21:27:06 <xorAxAx> yes
2006-08-07T21:27:46 <Kepplar> for some reason the regex (that i lifted from Page) is picking it up
2006-08-07T22:41:02 * xorAxAx added recursion support in Request.Clock
2006-08-07T22:56:13 <xorAxAx> this is the new line:
2006-08-07T22:56:14 <xorAxAx> Page.execute = 0.142s base__init__ = 0.125s compile_huge_and_ugly = 0.260s getACL = 0.102s getPageCount = 0.000s getPageList = 0.067s i18n_init = 0.002s loadLanguage = 0.004s load_multi_cfg = 0.117s run = 0.781s send_page = 0.585s send_page_content = 0.492s send_page_content|1 = 0.027s send_page|1 = 0.031s total = 0.906s
2006-08-07T22:56:58 <xorAxAx> you see that send_page and send_page_content call themselves
2006-08-07T23:03:14 <xorAxAx> which was the randomquote macro in this case
2006-08-07T23:05:12 <xorAxAx> fpletz: now feel free to add timers where you want, even in places that get called recursively
MoinMoin: MoinMoinChat/Logs/moin-dev/2006-08-07 (last edited 2021-02-16 00:17:49 by Bind)
