#foswiki 2010-10-16,Sat

↑back Search ←Prev date Next date→ Show only urls(Click on time to select a line by its url)

WhoWhatWhen
BobC_UKI've encountered a problenm with the User and Login. I can log in using my sort name, but not my WikiName, and when I sign in in my shortname, the link at the top left bar "Hello Bob?" implies I am not known. My Wikiname is in the USers topic though. THis also gives a problem in adding a user to a group: the user adds with its short name, but not WikiNAme, and the short name does not have
a link to the Main.Username
[00:06]
pharveyBobC_UK: what loginmanager are you using? and usermapper? [00:13]
BobC_UKFoswiki Login and TopicUSerMapping
I am setting up 1.1.0 in a new install to then copy my 0.9 over. I used the normal registration routine to create me, and then used SuperAdmin to add me to the AdminGroup. But when I log in I have issues with shortname/wikiname.
[00:16]
pharveyYou mean TemplateLogin? [00:23]
BobC_UKYes. However, I found the problem. The AllowLoginName was blank in configure, but somehow it allowed me to enter a LoginName and get in. Turning this to allow login gets me my link back. [00:25]
pharveythis sounds familiar
pharvey looks for a task
We tried very hard to avoid bugs introduced by the new features added toward the end of the dev cycle. I guess we just accepted that all the 'old' new features added earlier in the dev cycle surely must be stable, simply because they had been in trunk for longer :)
pharvey will campaign for feature-freeze every 6 months :)
so... end of march should be when we freeze trunk again
[00:25]
BobC_UK:)
I now have another problem.... I've got a "Foswiki has received a suspicious change request from your browser.
There is a risk that someone may be trying to use your identity to change content on Foswiki without your knowledge.
If you are unsure what this means, read: why you are being asked to confirm." bit no confirm button.
I've got round it by using the "Edit Preferences" extension option to save.
[00:36]
gac410Hi BobC_UK: Sorry, I've been away from keyboard. TBH I'm not sure what to do about StringifierContrib
The shell issue is I suspect related to your hosted installation - probably something funky with the shell environment.
[00:43]
BobC_UKBut it's OK in 0.9 [00:45]
gac410I really didn't want to hear that. :-D [00:46]
BobC_UK:-O [00:46]
gac410I guess with the shell hang, what I'd do is edit tools/extender.pl and start adding print STDERR "got here \n"; statements to see where things hang up. I'd start with before/after the $session = new Foswiki($user); statement. [00:51]
For the StringifierContrib hang in the ui, it will be really difficult to debug since you can't access the apache error logs. We really need someone to recreate this with an accessible server. Not sure what to suggest without logs. [01:00]
BobC_UKOK< I then renamed the StringifierContrib back to "pm" so the install would pick it up. I've just got the timer whirring away. In a terminal entering perl -d tools/extension_installer asdf produces this output: http://img715.imageshack.us/img715/3916/snagprogram0043.png [01:01]
gac410At that point it is stopped at a breakpoint on the first executable statement. Enter "c" to continue [01:02]
BobC_UKgot here
got here
Do you want to use locally found installer scripts and archives to install asdf and any dependencies.
If you reply n, then fresh copies will be downloaded from this repository.? [y/n]
[01:02]
gac410wow - that seems to be running normally now.
enter y and keep going.
[01:03]
BobC_UKRemember I installed the Readline package [01:03]
gac410Ah. So maybe ctrl-c out of that and try it again without the -d debug and see if things are working.
You installed Term::Readline::Perl ?
[01:03]
BobC_UKShould I try to uninstall the StringifierContrib using this command? [01:04]
gac410well - do you want to uninstall it? Other than breaking configure, we believe it is fully functional. [01:04]
BobC_UKI wonder if I uninstall and reinstall with the pacakage in place, it may install OK? [01:05]
gac410it's probably installed okay. The issue is unrelated to correct installation. [01:05]
BobC_UK"You installed Term::Readline::Perl ?" < Yes [01:05]
gac410Okay - not knowing much about perl readline requirments. ... I wonder if we need to make that a dependency for Foswiki.
Or if it is just something strange in your hosted env.
[01:06]
BobC_UKDunno, it's all magic to me anyway!
AS I have all my extensions installed and enabled, I may just leave this "installer extension" a miss for now and concentrate on configuring this some more.
[01:07]
gac410yeah - I'd worry about forward progress first. [01:08]
BobC_UKAgreed.
Thanks for your help. Appreciated.
[01:08]
gac410I believe we could have a brute force fix in Dependency::studyInstallation() ... If $module = StringiferContrib, then side-step the error. [01:09]
BobC_UKMmm, that could be useful. Is that in the extender.pl? [01:09]
gac410No - extender.pl is only from the shell. lib/Foswiki/Configure/Dependency.pm is used to load and determine the version of any installed module. [01:10]
BobC_UKOK. let me get that up. [01:10]
gac410I believe that is what is causing the Web UI hang in configure. [01:10]
BobC_UKWhere and what should I put into dependency.pm? [01:11]
gac410(I don't have the exact code off-hand [01:11]
BobC_UKOK [01:12]
gac410I'll need to poke at it a bit. Not sure what it's called with. [01:12]
BobC_UKNo worries. [01:17]
gac410Okay - you can try this to see if it bypasses the hang. In lib/Foswiki/Configure/Dependency,pm in sub studyInstallation
sub studyInstallation {
my $this = shift;
my $load_errors = '';
if ($this->{name} =~ m/StringifierContrib/) {
return 0;
}
Add the last 3 lines there - If Dependency is called for StringifierContrib, it just ignores the module.
If that restores configure findExtensions from hanging, at least we learn some more about the issue.
[01:18]
BobC_UKYup, didn't hang [01:21]
gac410Good - okay but is also now won't show up as installed in the FindMoreExtensions scrren - but that's not the end of the world.
I have to deal with a couple of other things. bbiab
[01:22]
BobC_UKNP [01:24]
.... (idle for 17mn)
gac410Okay - really ugly brute force - hardcoded path - yes, I said ugly ;-)
if ($this->{name} =~ m/StringifierContrib/) {
my $path = "Foswiki/Contrib/StringifierContrib.pm";
return 1 if ( $this->_recover_versions($path) );
return 0;
}
BobC_UK: This should also work without hanging. To do it right, $path should be set based upon a search for the module. But this will get you by I suspect.
[01:41]
.... (idle for 17mn)
***pharvey has quit IRC (Quit: ChatZilla 0.9.86 [Firefox 3.6.10/20100915180533]) [01:59]
BobC_UK@ George... Sorry for the delay.... You're right, coding works and shows Stringifier installed. [01:59]
gac410I'll paste the code into the Task and hopefully we'll find someone with a accessible test system where they can recreate it to really solve the issue. [02:00]
BobC_UKK. Off to bed.... 3:00 am! [02:01]
gac410g'night [02:01]
BobC_UKgn [02:01]
***BobC_UK has left [02:01]
..... (idle for 24mn)
sir_tyrion has quit IRC (Ping timeout: 240 seconds)
Colas has quit IRC (Ping timeout: 240 seconds)
sir_tyrion has joined #foswiki
Colas has joined #foswiki
[02:25]
ArthurClemens has quit IRC (Quit: Leaving...) [02:35]
............. (idle for 1h4mn)
SvenDowideitheya gac410 are you getting somewhere with your group-page? [03:39]
gac410no - i've been sidetracked with life.
Other than chasing the Stringifiercontrib issues for BobC_UK - i've not done anything else.
[03:39]
SvenDowideitlife is good :) [03:40]
gac410The bulk of my fix is okay, but the group cache and recursive expansion have issues. [03:40]
SvenDowideitSvenDowideit hasn't had much success getting through my list either [03:41]
***pharvey has joined #foswiki [03:41]
SvenDowideitone girl sleeping, the other screaming for hours [03:41]
gac410Babar, or pharvey - don't recall which now - gave it a review and said it seemed to be on the right track except for a recursive loop [03:41]
SvenDowideitI'm sadly presuming this is my new reality
oh god :(
http://foswiki.org/Tasks/Item9831#foo_1
i wonder why Lavr didn't notice this one
[03:42]
gac410yeah - I saw that one too. [03:44]
SvenDowideitSvenDowideit runs and hides in a dark hole [03:45]
gac410probably ought to put that one into the KnownIssues [03:45]
***GithubBot has joined #foswiki [03:45]
GithubBotfoswiki: master SvenDowideit * e529571 (9 files in 7 dirs): Item9750: rewrite regex conversion to mongodb queries to take advantage of the new mongodb OR and simplify ... - http://bit.ly/cPBSTv [03:45]
***GithubBot has left [03:45]
gac410SvenDowideit: Foswiki::ListIterator - it only does arrays, not hashes, right? [03:46]
SvenDowideitthats what i thinkyes [03:46]
gac410Also looking/thinking about TUM & cache. We need an afterSaveHandler to refresh the cache if someone edits the metadata of a group manually. [03:47]
SvenDowideiti thought that the 'cache' was not per session, but only per request [03:48]
gac410don't svn up, cuz I reverted, unless you cherry pick [03:48]
SvenDowideitbut i'm not sure at all anmore [03:48]
pharveyI saw it [03:48]
SvenDowideitits all still in the repo, so :) [03:48]
gac410SvenDowideit: - But what about persistent perl - [03:48]
SvenDowideitsee, that's what sun finish() is about
finish is (i think) supposed to be called at the end of every request
and if thats so, then the 'group cache' is a shortlived cache - for exactly this kind of reason
[03:49]
gac410Oh. okay - I have no idea - though it might be a nice perf. boost if the group stuff was cached. I read too much into the concept of the cache. [03:50]
SvenDowideitif you can think of a way to reset caches on seperate processes on seperate servers
then you can consider cacheing for longer
but we don't yet have that kind of cleverness
[03:51]
gac410ugh - never mind. [03:51]
SvenDowideitand yes, there are people that use nfs etc and seperate render servers
mad imo, but still
[03:51]
pharveySvenDowideit: sorry to ask a question I'm sure your E-mail addresses.. does WebDAVPlugin require mod_perl? [03:54]
SvenDowideityes
not for foswiki
but for the separate module that implements DAV
[03:55]
pharveyoh. So I can use fcgid for foswiki itself [03:56]
SvenDowideityup [03:56]
pharveynot that I'll be getting to webdav until late next week
pharvey pressing mongo related buttons
[03:57]
SvenDowideitgrin
have you played with wikidraw yet?
i'm sure there are bugs, but :/
[03:57]
pharveyyes
it's perfectly fine for v1
though I did manage to break it by adding an img element whose image file didn't exist
I was able to edit again and delete the offending element
[03:58]
SvenDowideitaddatask :)
ok, updated http://foswiki.org/Development/MongoDBPlugin
[03:59]
pharvey(it wasn't rendering from view) [03:59]
SvenDowideitah, y, the non-creation issue
SvenDowideit has to go feed girls
[04:00]
pharveySvenDowideit: my upstart script is actually wrong there - the pidfile isn't actually used and upstart gets confused about which job it's restarting (if you issue restart/stop commands). Anyway I'll update that part too.
gac410: mod_perl wasn't working in NACG - I updated again, this time it seems to work except for any script that does redirects... including login :( complicated debug session revealed that I know nothing about anything and I don't really understand why on earth foswiki is doing a redirect *inside* the initial request where we did the POST to login, resulting in a subrequest that isn't...
...handled by the rewrite rules and then you get a totally blank page and no log messages
I *expect* the initial POST to be one request, then the browser is given a response that is a redirect back to the originating page
[04:01]
gac410ugh. okay - thanks. I won't have time to deal with it for a while - mod perl was the one I didn't test - between having fastcgi and fcgid and mod_perl all installed on my laptop ... [04:05]
pharveyWhat I *got* was what _appeared_ to be foswiki trying to return the content of the originating page in the response to the initial POST request - IE. there is no 302 redirect [04:05]
gac410strange. [04:05]
pharveyno that's okay, I intend to help fix it, just giving you an update :)
my interpretation is of course impossible, I know that's not how foswiki works, but that's what the rewrite log was suggesting to me
[04:05]
gac410mod_-rewrite is truly a quagmire at times. [04:06]
pharveyI'm looking forward to the day where I have the time to do an nginx or lighttpd ConfigGenerator - certainly lighttpd seems to have a much less maddening config arrangement especially for URL rewrites [04:09]
***pharvey has quit IRC (Remote host closed the connection)
Lynnwood has quit IRC (Quit: Lynnwood)
pharvey has joined #foswiki
[04:22]
wbniv_ has joined #foswiki [04:28]
pharveyFoswiki cache is truly remarkable. 660ms down to 70ms [04:36]
.... (idle for 18mn)
***wbniv_ has quit IRC (Ping timeout: 272 seconds)
gac410 has left
[04:54]
....... (idle for 32mn)
wbniv_ has joined #foswiki [05:30]
pharveySvenDowideit: I have a minor fix to help MongoDBPlugin work on trunk. Shall I paste a diff on task, or may I clobber your code? :) [05:33]
SvenDowideitclobber away
you're just moving the getlist of webs call i imagine :)
[05:35]
pharveyI thought I'd make a Foswiki::Plugins::MongoDBPlugin::_getListOfWebs which is a wrapper and calls the right one, depending on Plugins::VERSION (we need to bump it on trunk for this to work) [05:40]
........ (idle for 38mn)
***CDot has joined #foswiki [06:18]
pharveyCDot: is 2.0.0 our next release?
there will be no 1.2?
[06:18]
CDotpharvey: that's what I've been assuming
I think it's time for a bang
[06:19]
pharveyIn Item:9814 you did some good work, but we need a Plugins::VERSION bump on trunk so that MongoDBPlugin can use the appropriate getListOfWebs() [06:20]
foswikibothttp://foswiki.org/Tasks/Item9814 [06:20]
pharvey(DBIStorePlugin too I assume)
so this innocent little task prompted me to create a skeleton ReleaseNotes topic, among other things... and bump from 2.1 -> 2.2
CDot: what are your thoughts on the features that will make the 2.0 release?
I am highly convinced that we need feature freeze more often, to make more frequent releases. We should probably freeze again end of March 2011 (that would have given trunk 6 months between feature freezes)
with only 6 month dev cycle, trunk would hopefully only need to be frozen for a matter of weeks, rather than months
[06:20]
.... (idle for 15mn)
CDot: Foswiki 1.0 Plugins::VERSION = 2.0, 1.1 Plugins::VERSION = 2.1, so I will make 2.0 Plugins::VERSION = 3.0
... or is that unnecessary :)
[06:38]
CDotI subscribe to the frequent-release policy; this is the main reason for the drive to continuous integration
however that approach is not supported wholeheartedly in the community
personallY I avoid using the Foswiki::Plugins version wherever possible
instead, I use if (defined &Foswiki::Func::FunctionINeed)
A DB store is top of my list for 2.0, yes
but, given the rate of development, releases every 6 months *are* frequent. The velocity just doesn't justify more frequent releases
if you are contemplating making a radical change to MongoDBPlugin, consider using the new VC Listener API
[06:38]
pharveythis is trunk only? [06:44]
CDotyes
re: Item9814: getListOfWebs already returns the list of webs in the "natural" order, doesn't it?
the problem I was having with the ordering stemmed from the web= parameter, which was being used as an ordering
CDot would like to know what changes to getListOfWebs are proposed
CDot would be very happy to see that API deprecated, and replaced with a better web iteration/listing API
[06:44]
pharveyI understand; my only concern is that MongoDBPlugin stopped working on trunk, because the method has been moved, which is perfectly fine - I'll use if (defined.... instead of Plugins::VERSION and let someone else do the ReleaseNotes02x00, Plugins::VERSION bumb [06:45]
CDotwhich method was moved? [06:46]
pharveySearch::InfoCache::getListOfWebs [06:46]
CDotthat is *not* a published API
Foswiki::Func::getListOfWebs is the published API
extensions that call unpublished APIs need to find their own compatibility solutions, I'm afraid. otherwise we would have no flex in the internal APIs at all.
[06:46]
pharveyBruteForce doesn't use ::Func [06:48]
CDotBruteForce is a core module, not an extension
when I moved Search::InfoCache::getListOfWebs, I updated BruteForce
[06:48]
pharveyfair enough [06:49]
CDotanother feature i think is key for 2.0 is search results ordering and grouping - basically, finishing the work that Sven started
order="majorkey,minorkey,otherkey,..." groupby="majorkey,minorkey,...."
[06:55]
pharveythis would be nice :) [06:58]
***GithubBot has joined #foswiki [07:00]
GithubBotfoswiki: master PaulHarvey * e1efa90 (3 files in 3 dirs): Item9750: fix for trunk: non-public API _getListOfWebs() moved since 1.1 ... - http://bit.ly/ay2ph0 [07:00]
***GithubBot has left [07:00]
uebera|| has quit IRC (Changing host)
uebera|| has joined #foswiki
pratchett.freenode.net sets mode: +o uebera||
uebera|| has quit IRC (Excess Flood)
uebera|| has joined #foswiki
[07:08]
CDotpharvey: http://foswiki.org/Development/FoswikiTwoZero [07:12]
***AndreU has joined #foswiki [07:12]
SvenDowideitmmm
ok, if you're convinced that you're going to do that major a work in the next 6 months
then I'd like to request that we do away with the insane mess that is queryalgo and searcha lgo
and rather have 'processing' engines that register an ability to process certain Query::Nodes
and added to that, that we find a way to register scopes
ie, scope=tag, scope=log (that thing we're talking to lynnwood about) etc
personally, I prefer the idea of doing a 1.2 in which we focus on UI
as discussed before
especially in the context of 6 months from now
but neh - its not like i'll be able to find that much time, bejond the mongodb work
pharvey, if you can formally confirm that in the mongo work, i don't have to support 1.1, I'll move to trunk next time i work on it
as for m2, the listener API is significantly better
[07:15]
pharveySvenDowideit: will do [07:20]
SvenDowideitthough of course, by m4, its irrelevant, as it'll be a full store
really, the listener/afterSave stuff is a kludge because I want the search to show the benefit first, but its a chicken/egg tissue
[07:20]
***leik has quit IRC (Ping timeout: 276 seconds) [07:23]
SvenDowideitCDot, can you amend http://foswiki.org/Development/FoswikiTwoZero to say 'you realistically think you can do __in the next 3 months__
as 6 month releases means code freeze in 4 monts time..
[07:23]
***MTempest has joined #foswiki
MTempest has quit IRC (Changing host)
MTempest has joined #foswiki
[07:24]
SvenDowideitactually, that topic should be redone to only contain a list of existant and accepted feature proposals that each developer has commited to
and developers should be required to remvoe their commitment / change the tobereleasedni flield to match their realistic estimate
otherwise we just get a list of dreams
not that dreaming is bad, just that we do lots of dreaming here already
[07:25]
pharveySvenDowideit: added two things against my name, neither of which have accepted feature proposals. [07:26]
SvenDowideitSvenDowideit thinks - oops [07:26]
pharvey:)
SvenDowideit: could we seriously work on http://trunk.foswiki.org/Development/QueryAcrossTopicRevisions into '2.0'. I made some design decisions last week which erm, assume this functionality will magically pop into existence at some point next year :}
well, let me re-phrase that
[07:26]
SvenDowideitI had an idea that i've not thought through..
scope=rev
ie, scope=topic == latest topic only
(for search
wrt TOM, ok, we talked about that before
[07:29]
pharveyextra scope param on SEARCH and QUERY macros? Well, we have scope param already on SEARCH [07:31]
SvenDowideitnot an extra param, extra values for it
atm, scope is shite and pointless
[07:31]
pharveyyeah. [07:31]
SvenDowideitexcept for regex context
i needed scope=tag about 6 years ago
and i may be making scope=log soon
[07:31]
pharveycool. well, no need to stress about it immediately, but if we can get this into the next foswiki release I would be immensely happy
it would pay off more than GridFS attachments even .... :)
[07:32]
SvenDowideitexactly [07:34]
***EugenMayer has joined #foswiki [07:35]
pharveyThe foswiki side of getting MongoDBPlugin working was terrifyingly simple btw [07:36]
SvenDowideiti hope so :) its a meme i've used for a very long time
bbl
[07:36]
pharveypharvey wonders why his mongod shards are always eating ~5% CPU when idle [07:37]
SvenDowideitthey need to talk to each other [07:37]
***leik has joined #foswiki [07:38]
SvenDowideitand it wakes up to sync to disk for 'i don't really grok why' [07:38]
pharveypharvey pokes JQGridPlugin to make it work without DBCache :D [07:38]
***SvenDowideit sets mode: +oooo AndreU CDot Colas donbarry__
SvenDowideit sets mode: +oooo EugenMayer gmc MTempest pharvey
SvenDowideit sets mode: +oo usepidgin wbniv_
[07:39]
CDothmmm, scope= might be interesting
CDot likes scope="attachments"
[07:41]
SvenDowideitthat too [07:41]
CDotCDot also likes scope=<query expression> [07:41]
SvenDowideiti wanted that, and even implemented it for kino [07:41]
CDotscope="parent='MyTopic'" [07:41]
MTempestattachment filenames or attachment contents? [07:41]
SvenDowideitCDot, no [07:42]
CDottee hee [07:42]
SvenDowideitscope == use an alternate store backend [07:42]
CDotscope="life, the universe, and everything"
scope="wikipedia"
[07:42]
SvenDowideitSvenDowideit gives up and tries to go back to getting 5mins of something done [07:42]
MTempestscope="http://svn.foswiki.org" [07:43]
SvenDowideitagain, nononono and no [07:43]
MTempestscope="revision-comments" [07:44]
SvenDowideitno
scope=history
[07:44]
CDotscope="Sven's brain" [07:44]
SvenDowideitand then you query the comment [07:44]
CDotwhy note simply query search with a query that limits the scope? [07:44]
EugenMayer'orning [07:44]
SvenDowideitCDot, make one that allows devs to add new types of scope
as in new types of store
as in
[07:45]
CDotSEARCH{"history=~'thing'" ... [07:45]
pharveypharvey still doesn't know where a scope directive goes - into the query string somehow? [07:45]
SvenDowideitsearch the log files
or search the tags database
[07:45]
CDotSEARCH{"log=~'...'" type="query" [07:45]
SvenDowideitatm, scope basically changes the info database that is queried
there are only 2
topic name - as in `ls`
and topic contents
[07:45]
CDotcan't say I'm desperately happy with the idea of an "info database"
CDot needs selling on this idea
[07:46]
SvenDowideitrealistically, tags, log files, and other extenal datasets we want to integrate with [07:46]
CDotcan't see why you wouldn't add those to the query schema [07:46]
SvenDowideitare currently required to make a new macro
really?
[07:46]
pharveyso each entity in these external datasets defines its own property.map schema? [07:47]
CDotreally [07:47]
SvenDowideitmmm, i thought that having all data become TOM data, by allowing extensibility [07:47]
CDotsplitting up a query to address different databii is not that scary [07:47]
SvenDowideitand then being able to query it
would help nicely
but, tis a point
[07:47]
CDotthat's what I'm saying [07:47]
SvenDowideitits a halfway house to the prefered thing [07:48]
CDotextend the query schema, not the SEARCH macro [07:48]
SvenDowideitok, so deprecate scope as rubbish
y
ok, sold
the need is still there
and goes back to my first sentense
[07:48]
CDotCDot is worried that was too easy, and checks his chair for a whoopee cushion [07:48]
SvenDowideitdamn, third
<SvenDowideit> and rather have 'processing' engines that register an ability to process certain Query::Nodes
[07:48]
CDotyes, now *that* is something I like [07:49]
SvenDowideitso taht we can query from whatever source, and several sources at once [07:49]
CDotI was heading that way with the recent hoisting changes [07:49]
SvenDowideitthen some regex search's would be served by solr
y
[07:49]
pharveycan you mock up an example query string that queries a 'logs' database? [07:49]
SvenDowideitbut =~ is a mess
as HoistRE doesn't hoist it
[07:49]
CDotyeah [07:50]
SvenDowideitone moe
i have one lynnwood watned
[07:50]
CDotas well as =~ the query lingo needs a way to express the other search types - keyword, literal [07:50]
***EugenMayer has quit IRC (Quit: Leaving.) [07:51]
SvenDowideitlog[action='save' AND '12/12/2005'<date<'12/12/2008']
er, does it?
[07:51]
***AndreU has quit IRC (Quit: AndreU) [07:52]
SvenDowideiti thought that was pretty much what ~ does [07:52]
CDotwell, after rocessing through the searchAlgorithm it does [07:52]
SvenDowideitcourse, that also requires that we have pluggable Formatting bits [07:52]
CDotbut there is some pre-processing required [07:52]
SvenDowideitwhich I'm also working towards in the FORMAT codebase
ok
[07:52]
CDotwhich could be delegated to the query algo, eliminating the search algo completely [07:52]
SvenDowideitwell - that's the concept i've been working towards, but expressed more concretely, now that we have more concrete work in that direction
yes
[07:53]
CDotwe are starting to get a picture of a "proper" search algo now [07:53]
SvenDowideitya
SvenDowideit will write up a feature req later :)
[07:53]
CDotstill not found a nice way to degrade gracefully, tho
I find falling back to evaluate rather...... sledghammery
[07:54]
SvenDowideitits not
its actually very very useful
as we're able to test and confirm results third parties give us
[07:54]
CDotit would be usefuller if the search algo could return results at a finer granulariuty
rather than always "topic"
[07:54]
SvenDowideitthat was going to be 2.0 [07:55]
CDotk [07:55]
SvenDowideiti wanted search result sets to be able to return any TOM addressable [07:55]
CDothttp://foswiki.org/Development/FoswikiTwoZero [07:55]
pharveyso a normal vanilla QUERY is really working on a 'topiocs' db - topics[webpart='Sandbox' AND SomeField='JoeBloggs'] [07:55]
SvenDowideitbut if 2.0 is the next release, then buckleys
pharvey, very neat point
[07:55]
CDotpharvey: yes, good point [07:57]
SvenDowideitSvenDowideit isabout to go feed the family spag-bog, but will be back in ~3 hours to try to pull this together with whatever other fragments of 'intention' i can drag up [07:57]
CDotisn't the DB "webs" tho?
or, shouldn't it be webs?
[07:57]
pharveyCDot: probably, I haven't thought it through [07:58]
SvenDowideitSvenDowideit never considered a web as a real container [07:58]
CDotCDot is aware of the mess the web= parameter makes of it [07:58]
SvenDowideitand users probably think of them as databases of topics [07:58]
CDotweb *is* a container, and should be treated as such. Webs have history too! [07:58]
pharveythey do? [07:58]
CDotwebs should have all the rights of topics [07:58]
SvenDowideityes, i agree in a formalsense [07:58]
CDotsure they do. We just throw it all away. [07:58]
SvenDowideitbut in a user perspective, not so much
i'm not sure its worth confucing users for the sense of purity you're espousing :p
[07:58]
CDotwebs *have* meta-data - e.g. the WEBLOGO and shit like that
and I want to search it
[07:59]
SvenDowideitah, ok, so the excuse is that it makes the preference tree more sane [07:59]
pharveyRight; WebPreferences = web meta [07:59]
CDote.g. "find me all webs that contain the issue tracker wiki app" [07:59]
SvenDowideity
whereas atm, we think of 'find me all topics of type blah that indicate there might be a tracker app in them
[07:59]
pharveyso to query all QueryAcrossTopicRevisions we might have a revisions[webpart='Sandbox' AND author='JoeBloggs'] [08:00]
CDottopics don't necessarily know. It's *web* metadata
pharvey: yes
we were discussing a revision syntax some time ago
Blahweb/BlahTopic?rev=blah
[08:00]
pharveyyes, that's the only sticky part left then [08:01]
CDotBlahWeb/BlahTopic?rev=* [08:01]
pharveyI really don't like ?rev=blah because you can no longer concatenate the emitted value blindly onto a URL (the concatenation might have multiple ?)
but maybe that's ok, I dunno
[08:01]
CDotCDot isn;t serious about that syntax, just illustrating in terms we all know [08:02]
pharveyk [08:02]
CDotwe are running out of ascii :-( [08:02]
pharveyhehe [08:02]
SvenDowideiti like the idea that we add scoping to a prefix [08:03]
CDotnot used # yet [08:03]
SvenDowideitthat way we don't need more chars
rev[youre='asdf'] query
[08:03]
CDotSvenDowideit: illustrate?
oic
[08:03]
SvenDowideittopic[youre='asdf'] being the long form of the [08:03]
CDotrev of what? [08:03]
SvenDowideityoure='asdf'
that query would be look through every rev of every topic
[08:03]
pharveybrowsers mess up URLs containing # [08:04]
SvenDowideitbecause its sooooooo broad [08:04]
CDotso, revisions[parent='Mum'] returns an array of meta? [08:04]
SvenDowideityes
array of TOM addresses
or today, array of webtopic names
[08:04]
CDotthus revisions[parent='Mum'][age=40] should return all 40 year olds with a mother? [08:04]
SvenDowideitthat way log[action='save'] [08:04]
pharveyhmmm, new problem... can I unique-ify the list so that I only get the first or last rev where parent='Mum', instead of every rev of every topic? :) [08:05]
SvenDowideityes, we need first, last and unique operators
but thats pretty seperate
[08:05]
pharveyCDot: I thought it would be revisions[parent='Mum' AND age='40']
agreed
[08:05]
CDotfirst(revisions[parent='Mum']) works for me [08:05]
SvenDowideitdepends
CDot, queried for all reves that have a mum that is currently 40
whereas you queried for all revs that had a mum that was 40 in that rev
[08:05]
CDotpharvey: yes, it could be - but a query might be composed from sub-queries [08:06]
SvenDowideitdifferent query
at which point, you will want a visualiser
[08:06]
CDotCDot wants result sets to act as lists in queries [08:06]
pharveynice [08:06]
SvenDowideitCDot, i think so too [08:06]
pharveypharvey esplodes [08:07]
SvenDowideitquite a long way to go for it tho
SvenDowideit has to go, back in 3 hours
laters
[08:07]
CDot%SEARCH{results="saveme" ...}% %SEARCH{type="query" "saveme[constraint]" ... [08:07]
pharveyseeya [08:07]
CDotSvenDowideit: ciao
we had this type of result set in FormQueryPlugin for ~10 years now
[08:07]
pharveyCDot, that would be nifty :) Regarding first(), would that give us a single hit, or the first rev of each topic that is in the resultset [08:08]
CDotpharvey: think like a programmer [08:08]
pharveyCDot: my current job makes that diffuclt
difficult*
[08:09]
CDotso first([[A]]) == [A] [08:09]
pharveyI guess I'm not sure what the resultset of a revisions[] query looks like. Is this an array of arrays? [ ['Web.Topic--4', 'Web.Topic--8'], ['Web.Topic99--3', 'Web.Topic99--63'] ]
or do we get a flat list? ['Web.Topic--4', 'Web.Topic--8', 'Web.Topic99--3', 'Web.Topic99--63'] ]
if the latter, then first() can only return 'Web.Topic--4'
if the former, than first() might be able to return ['Web.Topic--4', 'Web.Topic99--3']
[08:11]
CDotpharvey: internally, it's a list of Meta objects, each of which is tied to a specific rev
CDot isn't revealing how it's visually represented in HTML, because he doesn't know
[08:13]
pharveyso a flat list?
hmmm
[08:14]
CDotI would say so, yes
there is no concept of "topic" as a container for "revisions"
that was blown away when I was told it was too complicated for mere mortals
[08:14]
pharveyd'oh :)
if each $topicObject in the result set was a subset of the matching topic revs from the query, there might be some hope.
ie. $topicObject contained multiple revs
[08:15]
CDothope of what? a topic/rev is an addressable entity
a "range of revs" is not
revisions[rev>3 AND rev<5]
revisions[info.rev>3 AND info.rev<5] I mean
[08:16]
pharveyI was imagining that the query would build a brand new $topicObject which only existed for the result-set. It would be loaded only with the topic revs that match the query. [08:18]
CDota $topicObject is tied *to a single rev*
or is uncommitted, in which case it refers to *all* revs
[08:18]
pharveyI know this
I was existing in a parallel universe where that wasn't the case :)
[08:18]
CDota subset of revs is a result set from a query, not a topic object [08:18]
pharveythinking*
except now each member of the resultset are peers with all the others
[08:18]
CDotyes; because you can only returns sets of addressable objects
or (thinking) subqueries
so a result set could be an iterator over a set of result sets (ouch, my head)
[08:19]
pharveyyou can return subqueries; ok - so revisions[name='Web.Topic99' AND (rev='3' OR rev='63')] [08:20]
***EugenMayer has joined #foswiki [08:20]
CDotyes. To fully qualify the query: 'Blahweb.BlahTopic'/revisions[name='Web.Topic99' AND (rev='3' OR rev='63')] [08:22]
pharveythe problem: user is likely to want to know, for example, all the topics he/she every contributed to, but construct a SEARCH which only applies the format="" once for each topic (and *not* once for each matching topic-rev) [08:22]
CDotCDot reels at the ugliness of that
pharvey: then they should nested-search
[08:22]
pharveyI don't understand how that solves the problem [08:23]
CDotthis is what FormQueryPlugin is so good at
ho hum
[08:23]
pharveythe problem is that the initial query returns too many hits. A nested search just makes it worse?... [08:24]
CDotno; the initial query finds "topics I contributed to" and the nested search finds "revisions of those topics I contributed to"
not sure how the first query works if you contributed in the past, however
unique()
[08:25]
pharveyright, that is what I was getting at
unique() would work
[08:25]
CDotwell, I'll think about it. Right now, I need to prep the bike for today's ride. The sun is shining! [08:26]
pharveyhave fun :) [08:26]
***ktwilight has quit IRC (Quit: dead) [08:27]
ktwilight has joined #foswiki
AndreU has joined #foswiki
[08:35]
sreher has joined #foswiki [08:46]
sreher has quit IRC (Quit: sreher) [08:52]
sreher has joined #foswiki [09:02]
HolstenerLiesel has joined #foswiki [09:08]
usepidgin has quit IRC (Quit: Leaving.) [09:13]
CDot has quit IRC (Ping timeout: 264 seconds) [09:22]
CDot has joined #foswiki [09:36]
AndreU has quit IRC (Quit: AndreU) [09:50]
AndreU has joined #foswiki [09:55]
CDot has quit IRC (Quit: Leaving.)
BobC_UK has joined #foswiki
[10:00]
BobC_UKMorning all. I have one issue when editing a topic: I get this message "Unable to install TinyMCE; <META name='TINYMCEPLUGIN_INIT' is missing". On saving I get "Foswiki has received a suspicious change request from your browser. ......" but there is no "OK" button to confirm. I am unable to edit my preferences at the moment. [10:05]
pharveyBobC_UK: in System.InstalledPlugins, what is the TinyMCEPlugin version?
bbl, but make sure you've got (28 Sep 2010, $Rev: 9405 (2010-09-28) $) - also check that you haven't set a TINYMCEPLUGIN_INIT variable anywhere; you might try reading the HTML source and telling us if you can find a <meta name="TINYMCEPLUGIN_INIT" content="...."/> tag in the HTML <head>
lastly, if it was on a site that previously held foswiki 1.0.x, try clearing browser cache as the 1.1.x version has new js
[10:11]
BobC_UKThanks. This is a clean install of 1.1.0 before I migrate over 0.9.0 I am just configuring a bare bones set up at the moment. I am Foswiki-1.1.0, Mon, 04 Oct 2010, build 9498, Plugin API version 2.1
There is <meta name="foswiki.TINYMCEPLUGIN_INIT_ENCODED" content="%0a%22mode%22%2......... /><!--TINYMCEPLUGIN_INIT_ENCODED-->
[10:16]
***ArthurClemens has joined #foswiki
pharvey has quit IRC (Ping timeout: 265 seconds)
[10:21]
.... (idle for 19mn)
BobC_UKWould the absence of a topic "TINYMCEPLUGIN_INIT_TOPIC" in the latest 1.1.0 install result in the <META name='TINYMCEPLUGIN_INIT' is missing" message? [10:43]
***uebera|| has quit IRC (Excess Flood) [10:47]
BabarBabar reviewed gac410 patch, but hasn;t finished yet :) [10:47]
***uebera|| has joined #foswiki [10:48]
SvenDowideitguys - does anyone have the time to upgrade foswiki.org? [10:48]
BabarSvenDowideit: I thought we said we don't want to do that :) [10:49]
SvenDowideitSvenDowideit wasn't nearby when 'we' - i presume there was a reason? [10:49]
Babarbecause of the bug I'm working on with gac410, the one Raul found [10:49]
SvenDowideitoh, the group editing bug? [10:49]
***uebera|| has quit IRC (Changing host)
uebera|| has joined #foswiki
[10:50]
BabarItem:98019 and Item:9821 [10:50]
SvenDowideitok i guess - not that it should affect us [10:50]
foswikibothttp://foswiki.org/Tasks/Item98019
http://foswiki.org/Tasks/Item9821
[10:50]
Babarnot editing
the ACLs
[10:50]
SvenDowideitoh,:/ [10:50]
Babareven though I couldn't reproduce, and neither could George [10:50]
SvenDowideitdid you mana ge to repro it?
cos i had something similar 6 months ago
SvenDowideit pokes self in eye
[10:50]
Babar^T [10:52]
SvenDowideitthat is odd, but grr [10:52]
***Babar sets mode: +oooo ArthurClemens AndreU EugenMayer ktwilight
Babar sets mode: +ooo leik sreher uebera||
[10:52]
.... (idle for 15mn)
BobC_UKHi, I have a problem correctly initialising TinyMCE Plugin. I cannot get rid of "<META name='TINYMCEPLUGIN_INIT' is missing" every time I edit a topic. [11:07]
***pharvey has joined #foswiki [11:07]
BobC_UK@pharvey- are you able to offer some additional help to sort the timymceplugin isseu I have? Thx [11:14]
pharveyBobC_UK: strange. Can you say if the <meta name="TINYMCEPLUGIN.... appears before or after the <script ..... src="...../tiny_mce.js"/> ?
ctrl+f in the html source (find) where tiny_mce.js is mentioned. It should be followed by foswiki_tiny.js and foswiki.js - these <script> tags should be all together with other .js files right before </head>
[11:15]
BobC_UKMm, I read a bit more about TinyMCEPlugin and I have created a new topi in System called TinyMCEInit and copied the %STARTSECTION{"TINYMCEPLUGIN_INIT"}% .... to it, and changed the DefaultPreference to point TINYMCEPLUGIN_INIT_TOPIC at the new topic. So the TINYMCEPLUGIN message is no longer in the Header. However, I still get the pop up messgae about the MEta. [11:18]
pharveywhat TINYMCEPLUGIN message is that?
(or was that?)
[11:19]
BobC_UKUnable to install TinyMCE; <META name='TINYMCEPLUGIN_INIT' is missing [11:19]
pharveybut you said it was 'no longer in the Header' [11:19]
BobC_UKCorrect [11:19]
pharveywhat message is no longer in the header? [11:20]
***Babar sets mode: +o pharvey [11:20]
pharveypharvey horribly confused - this stuff should just work :} [11:20]
Babaryou installed and enabled the plugin? [11:20]
pharveyit's TinyMCE - it should just work [11:21]
Babaryeah, it ought to be installed by default
have you checked the System/InstalledPlugins? Maybe theplugin reports some initialisation error?
[11:21]
pharveyhe's getting the popup which means it's trying to load but fails to get the <meta tag holding the init settings [11:21]
Babar(would be a first, but who knows?) [11:21]
pharveysounds like javascript/client-side problem [11:21]
BobC_UKI searched for TINYMCEPLUGIN_INIT in the source and its not there. [11:21]
pharveyoh, cool :)
that *is* interesting
could you please un-set the TINYMCEPLUGIN_INIT_TOPIC?
[11:22]
BabarBobC_UK: you sorted out your StringifierContrib issue btw? [11:22]
pharveyif you set it to anything
set it to System.TinyMCEPlugin
[11:22]
BobC_UKYes, Gac did a hard coed work around for now
Do you want it set to TinyyMCEPlugin or commented out?
[11:23]
pharveycomment it out for now [11:23]
Babarbtw, have you read the upgrade notes? [11:23]
pharveyhe's doing a fresh install AFAIK [11:24]
BobC_UKYeah, fresh install into another web [11:24]
Babarso the system web is fresh?
or you renamed the system web to like System01x01?
[11:25]
BobC_UKIt was......... but now I think I may have copied the Pub and Data "System" directories over...... [11:26]
pharveyok
let's do a sanity check
[11:26]
BobC_UKI've commented out the line as asked... [11:26]
Babardo you have the anchor: System/UpgradeGuide#TinyMCEPlugin_TINYMCEPLUGIN_INIT
to know if you have the 1.1 or 1.0 version :)
[11:26]
pharveyCreate a topic, write %TINYMCEPLUGIN_INIT% and %TINYMCEPLUGIN_INIT_TOPIC% and see if they expand to anything
pharvey pretty sure you don't have them set, but should only take a couple of seconds...
[11:26]
BabarBabar meant System/TinyMCEPlugin#UpgradeFrom01x00 but both should do [11:27]
BobC_UKSystem/UpgradeGuide#TinyMCEPlugin_TINYMCEPLUGIN_INIT resolves OK and is present
http://www.paredox.com/foswiki2/bin/view/System/UpgradeGuide#TinyMCEPlugin_61TINYMCEPLUGIN_INIT_61_changes
oops. System/UpgradeGuide#TinyMCEPlugin_61TINYMCEPLUGIN_INIT_61_changes
[11:27]
pharveyI'm afraid that System.TinyMCEPlugin is a very tedious, boring, long-winded piece of documentation now :( I will have to clean it up at some point [11:29]
BobC_UKLet me do... " Create a topic, write %TINYMCEPLUGIN_INIT% and %TINYMCEPLUGIN_INIT_TOPIC% and see if they expand to anything" [11:29]
pharveygood [11:29]
BabarBobC_UK: is present as in brings you to the very bottom of the page, where there is: TinyMCEPlugin's configuration scheme has been reorganised. If you had a customised TINYMCEPLUGIN_INIT preference variable on Foswiki 1.0.x, it will need to be removed, re-built, replaced with a dedicated topic section or replaced with newly introduced preference variables for Foswiki 1.1.x. See upgrade advice for more information.
but pharvey's test is more acurate :)
[11:31]
BobC_UKI never installed 1.0 I was setting up 1.1.0 to migrate 0.9 dta over [11:32]
HolstenerLieselhi! we used to restrict viewing acces to group topics. with the new optimized group topics that doesn' seem to work. anyone can see all optimized group topics and the names of everyone in them unless I restrict access to the entire main web in the webpreferences. should I file this as a bug or is it already there, probably related to item:9809 ?
Item:9809
[11:32]
foswikibothttp://foswiki.org/Tasks/Item9809 [11:32]
BabarHolstenerLiesel: please update this bug with your remark
so we can add a unit test for it too
[11:32]
HolstenerLieselok [11:33]
BobC_UKPharvey - sorry for dealy, my second issue prevents me saving, so having to use txtx editer at a terminal! [11:37]
pharveyouch :( [11:37]
BobC_UKStill doesn't work. I get message about security and no "OK" button to confirm
Anyway, %TINYMCEPLUGIN_INIT% %TINYMCEPLUGIN_INIT_TOPIC% does not expand, they appear like that on the screen
[11:39]
Babarouch... your installation seems to be really fscked up :) [11:39]
BobC_UKI get the same whether I use Firefox or Chrome (hate IE) [11:40]
pharveyBobC_UK: is it possible for you to pastebin the HTML on an edit screen that gives you the 'unable to find <META' message? [11:40]
BobC_UKMM, I heard about pastebin but never used it. Do I need to register or something?> [11:41]
pharveyI really want to check the order that the <script> tags appear
no
http://pastebin.com/
[11:41]
Babaryou just copy the output on this website, click submit, it will give you a link, which you paste here [11:42]
BobC_UKIt's .org... but do I just paste inot the first box? [11:42]
BabarBabar wonders who invented pastebins, but this guy ought to be rich and famous now :) [11:42]
BobC_UKhttp://pastebin.org/234107 [11:43]
pharveyyou're using BrowserBooster!
disable that thing!
:)
[11:44]
BobC_UKOK.
Still same message. do I need to repost the HTML
[11:45]
pharveyyes please - BrowserBooster polluted the other one too much, didn't see any script tags [11:46]
BobC_UKk
http://pastebin.org/234357
Seesm to be truncating it for some reason
Shall I log out , close the bowser and start agan?
[11:46]
pharveyBobC_UK: was that an edit page? [11:50]
BobC_UKYes
No
!!
[11:50]
pharvey:) [11:50]
BobC_UKI press edit, get the message then go back. Should I remain on the edit screen? [11:51]
pharveyyes please; I need the HTML from the edit page [11:51]
BobC_UKK
http://pastebin.org/234393
Remember that %TINYMCEPLUGIN_INIT% %TINYMCEPLUGIN_INIT_TOPIC% do not expand in a topic.
[11:51]
pharveyyes, the plugin will use internal defaults [11:53]
BobC_UKk [11:53]
pharvey(if they are not set)
ok, I am not seeing anything obviously wrong :/
hrm
[11:53]
BobC_UKAH, hang on , I may have it in the Main preferences.... let me look
No, I don't
[11:54]
pharvey'This document was successfully checked as XHTML 1.0 Transitional!' so it seems to be well-formed [11:56]
BobC_UKlet me log out, close browser, clear cache, etc [11:56]
pharveyare there any javascript errors? In firefox you can see them using Tools -> Error console, or ctrl+shift+j
pharvey tries to get the page loaded into his own browser :D
[11:56]
BobC_UKYes.... http://img221.imageshack.us/img221/2158/snagprogram0044.png
I think I may have overwritten the new pattern.js with the old one when I copied the Pub directory?
[11:58]
pharveyBobC_UK: I just loaded the *exact html* you pasted and TinyMCE opens fine!!!!
you must have errors with the browser fetching javascript
do you have firebug
?
[11:59]
BobC_UKYes [12:00]
pharveycan you enable the net panel? [12:00]
BobC_UKDone [12:00]
pharveynow re-load the page
(edit screen)
[12:00]
BobC_UKedit?
k
k message in on screen atm
[12:01]
***khnz has quit IRC (Quit: No Ping reply in 180 seconds.) [12:01]
pharveyI am expecting that there might be some errors on the net panel
in red
[12:01]
***khnz has joined #foswiki [12:01]
BobC_UKIn the "All" tab? There are no red bits
Shall I respond to the "OK" and load the edit screen itself?
[12:02]
pharveyyes [12:03]
BobC_UKAll looks good. No red bits [12:04]
pharveywow
I am totally stumped.
There is nothing wrong with the rendered HTML
and all of the JS files are apparently delivered successfully to your web browser
[12:04]
BobC_UKWhat about those errors in Pattern.js?
All JS packages look fine 8 requestsd, 912ms
[12:04]
pharveyonly 8 requested? some must be in your browser cache I guess
the errors, now that's weird
[12:06]
BobC_UKLEt me log out, clear cahce and restart browser.... [12:06]
pharveywhen I open up your HTML in my own browser, with the URLs re-written to my test server, it works [12:06]
BobC_UKSounds like a cache issue [12:07]
pharveyor corrupted .js on the server [12:07]
BobC_UKI coukd try IE as I haven't used that to access the pages
MM, I get the message in IE as well
[12:07]
pharveybob [12:09]
BobC_UKYup? [12:09]
pharveythis is impossible
the error message you describe is no longer in the 1.1 TinyMCEPlugin
[12:09]
BobC_UKPerhasp something to do with the Twisty remember function? [12:09]
pharveywhat is the exact message? [12:10]
BobC_UKhttp://img837.imageshack.us/img837/5981/snagprogram0045.png [12:10]
pharveythat is the 1.0 TinyMCE
if we were running 1.1 JS it should read:
the current source code reads:
Unable to install TinyMCE; 'TINYMCEPLUGIN_INIT_ENCODED' preference missing
[12:11]
BobC_UKWhat could I have copied across from Data and Pub directories that could have affected taht? [12:11]
pharveydo you have TWikiCompatibilityPlugin enabled? [12:12]
BobC_UKYes [12:12]
pharveypharvey should have noticed the message wasn't correct about an hour ago >:) [12:12]
BobC_UKShall I Uninstall and reinstall TinyMCE? [12:12]
pharveyno
I think you should try disabling TWikiCompat
[12:12]
BobC_UKk [12:12]
pharveymaybe the old TWiki web is getting used instead of current System
if this doesn't work, then yes, try re-installing TinyMCE
[12:13]
BobC_UKNope, same message [12:13]
pharveyok, re-install TinyMCE [12:13]
BobC_UKK will do... [12:14]
pharveywe should have tried that about an hour ago too ;-) [12:15]
BobC_UKYup, that appears to have worked. [12:15]
pharveyexcellent
as for twisty and pattern
perhaps you need to re-install those too
[12:15]
BobC_UKYeah! [12:16]
pharveyand who knows what else, just re-install them all :)
actually
this might explain your problem with the validation prompt
if you don't have correct pub/System/JavascriptFiles/*.js versions
[12:16]
SvenDowideitpharvey, http://foswiki.org/Development/GrandUnifiedSearchAPITwoZero [12:16]
BobC_UKWhen I read the notes about moving from 0.9 to 1.1 it said configure a new install then copy across the Pub and Date directories. Perhaps I should have left off the Main and System direcotries
OK I will reinstall all modules.......
[12:17]
pharveyyes :( [12:17]
SvenDowideitthat was horriffically productive i think - shame the girls spent the last hours screaming and not sleeping, back later [12:17]
pharveyBobC_UK: I don't think you can get JavascriptFiles as an extensions... may be better off finding a 1.1 and copying the pub/System and data/System files from that
SvenDowideit: thanks, reading now
[12:18]
BobC_UKYeah I can just extract form the tar and then FTP across [12:18]
pharveysounds good
did we really tell people to copy over the data/ and pub/ dirs in the upgrade advice?
[12:19]
BobC_UKDamn, I just selefcted all the extensions for a rei-install and got this message "Undefined subroutine &Foswiki::Configure::Util::rewriteShebang called at /home2/paredoxc/public_html/foswiki2/lib/Foswiki/Configure/Package.pm line 412.
"
[12:20]
Babarthat's not 1.1 [12:20]
BobC_UKOne by one the? :( [12:21]
Babarthat change has been done by IngoK after 1.1
did you install from Release01x01?
[12:21]
BobC_UKOK, let me get 1.1 system pub and data filess across . I must have screwed it up when I copied the complete Data and Pub directories
I just installed the latets package into a new directory on the webserver
[12:21]
pharveyfrom foswiki.org/Extensions or svn? [12:22]
BabarWhen you have a working installation you can simply copy all your webs from the 1.0.X to the 1.1.0 installation. Simply copy the directories in the Foswiki data and pub directories except the webs that are distributed with Foswiki. For the Main web you will probably want to copy all the files from the 1.0.X to 1.1.0 without overwriting the files distributed with 1.1.0. You may need to manually merge [12:23]
BobC_UKThat's where I Went wrng. I copied the lot.
http://sourceforge.net/projects/foswiki/files/foswiki/1.1.0/Foswiki-1.1.0.tgz
[12:23]
HolstenerLieselI've tested and updated Item:9809. my point about not being able to restrict access to optimized group topics might be unrelated. maybe this is an issue with the new design of those group topics, I don't know. [12:23]
foswikibothttp://foswiki.org/Tasks/Item9809. [12:23]
pharveythat upgrade advice should be explicit: "don't copy the damn System web " :) [12:23]
BobC_UKYEs. I didn't realsie [12:23]
pharveyyou won't be the last person who does this [12:24]
BobC_UKSo shall I just un zip the System and Main in Data and Pub and get those across?
I presume my current prefs are OK?
[12:24]
pharveydata/System/* and pub/System/* should be sufficient
yes, I assume so
[12:25]
BobC_UKK
bbiab
[12:25]
pharveySvenDowideit: awesome amount of work in that proposal, would be utterly cool to get done >:)
*that proposal represents an awesome amount of work
pharvey ponders how nested result sets might work, then goes back to stacking bikes
[12:28]
SvenDowideityes, it is a huge blob of work, but at least now we have gotten enough bits lined up that we can even envision this level of detail
exciting really
best bit for me, is that almost all of the lifting can be done in mongodb :)
[12:30]
pharveyI hope we can eventually do some sort of Cache::FastMmap cache that can ship with standard foswiki core - that would do some of what DBCacheContrib does; and perhaps an alternative to RcsWrap/RcsLite stores that could offer QueryAcrossTopicRevisions
in the spirit of shipping a Foswiki with minimal dependencies, no database admin
[12:32]
SvenDowideitit'll be pretty easy actually
yup
result sets, i i realised as i was moving computers that clearly, those need to be overridable classes
boy, what fun
thats a really big kettle of complexification
my main concern is really about muxing sets across boundaries
ie, moving 10000 topic addresses from mongo to solr
but thats like worrying about regex search being slow - when you really need to do it, its clearly worth it
[12:33]
pharveyI'm a little concerned we'll try to save the whole world and give ourselves too much work. But I assume you and Crawford know what you're doing ;-)
ooh
backlinks[source='Web.Topic']
backlinks[target='Web.Topic']
[12:38]
SvenDowideity [12:39]
pharveynice :) [12:39]
SvenDowideitwell, the thing is that cdot has been bugging me for years to refin and docco my dream
and you guys managed to define by example quite a bit of it
so clearly, the codebase in trunk is far enough along for me to be concrete
implementation is a staged animal - based more on who has what itch tho
SvenDowideit is starting to wonder about evaluating Query::Nodes on -non-TOM data
[12:40]
pharveycan we ship it by april? ;-) [12:42]
SvenDowideityes, but it will cost
mostly in childminding fees and coffee
[12:43]
***Colas has quit IRC (Remote host closed the connection) [12:43]
BobC_UKWell, restored the Data/System and Pub/System from the archive but sill get the "Undefined subroutine &Foswiki::Configure::Util::rewriteShebang called at /home2/paredoxc/public_html/foswiki2/lib/Foswiki/Configure/Package.pm line 412.
" message if I try to reinstall. It uninstalls and installs fine, but reinstall doesn't work.
[12:44]
***NeilGoodUk has joined #foswiki
NeilGoodUk has left
[12:44]
SvenDowideitoh, it'd have to be an inhome nanny
;p
april being very very agressive
hopefully someone with a critical eye can spot some redundancies in what i propose
[12:45]
pharveywe don't have to deliver everything by april - I had assumed QueryAcrossTopicRevisions would be later in the project
maybe I should have said " *what* can we ship by april"
[12:47]
SvenDowideiteasy - this is open source, so the answer is always 'whatever you implement'
:p
[12:48]
pharveynot if Lavr is release manager ;-)
pharvey not sure what an evaluation engine looks like
[12:48]
SvenDowideitpharvey me neither right now :/
its kind of an extraction of the SearchAlgo, into something that the evaluate() method on a node can push to
or, pull, or something
[12:49]
pharveyI vaguely get that
what plugins could we enable with this?
[12:49]
SvenDowideitin the 12months fomr now plan, if this existed, then the mongodb evaluation engine would be able to evaluate all those [12:50]
BobC_UKI think I'm going to do a clean install and start again. [12:50]
SvenDowideitkino for eg, would be good at words, and not much else [12:50]
pharveyI mean with respect to registering new syntax (pluggable syntax)? [12:51]
SvenDowideiti'm not positive that a Node should 'evaluate' itself, but meh [12:51]
pharveyBobC_UK: Were you told at some point to replace your lib/Foswiki/Configure/Package.pm file? [12:51]
SvenDowideitDBCachePlugin would be able to use the SEARCH and QUERY macro's, rather than its own
DatabasePlugin can do inline SQL
um, a fun one - query the internet
[12:52]
pharveyI see [12:52]
BobC_UKDunno, sounds familiar. Should i copy that to be sure? [12:52]
pharveyyes, sparql would be crazy nifty :D [12:52]
SvenDowideitSEARCH{"google[site:ibm.com thinkpad x61] [12:52]
pharveyBobC_UK: if you updated that file, then whoever told you to do it may have forgotten to tell you to copy the new lib/Foswiki/Configure/Util.pm [12:53]
BobC_UKI changed the lib/Foswiki/Configure/Dependency,pm to get around the stringifier issue. [12:53]
SvenDowideitthen take each result and use format= on it [12:53]
pharveyBobC_UK: http://svn.foswiki.org/trunk/core/lib/Foswiki/Configure/Util.pm
SvenDowideit: I am under the impression that we don't have detailed enough spec, hypothetical use-cases for that... the example you imply is operating on a resultset that contains non-topic elements
unless the query engine, as a part of delivering the result set, creates those topics...
... in fact this is almost what I'm doing with ExternalDataPlugin.. hmmmm
[12:54]
BobC_UKCopied the new Util nad reinstall gives the smae rewriteshbang error [12:57]
pharveyBobC_UK: you copied over the old Util.pm? (ie in the same place)? [12:57]
SvenDowideitbasically, for that query to work, someone would need to define their own: parser, evaluator, resultset and formatter [12:58]
BobC_UKI copied the link you gave to the same directory on the web and overwrote what was there [12:58]
SvenDowideitall of which would be very trivial, because they're pretty much handing off directly to google [12:58]
pharveyah of course, if you can define your own resultsets... hmm [12:59]
SvenDowideitand then parsing the replied XML, and $whatevering() [12:59]
pharveycould resultsets be complex graphs? Or just nested? [12:59]
SvenDowideitsorry, i added that to the topic after you read it
they can be whatever eventually evaluated into an ordered set
as SEARCH/FORMAT expect a list
[12:59]
pharveyBobC_UK: it sounds like re-installing might be the best way forward :( [13:00]
BobC_UKYep. :( [13:00]
SvenDowideitbut, as you know, hashes eventually can be farced into a list
there is, in addition, nothing to stop you from making a new macro that can use extended graph properties of some resultsets
[13:00]
BobC_UKSo when I copy over my 0.9.0 preferences I just copy the Main, and additional sites in both the Data and Pub Directories?
That wil also do the users and groups?
[13:01]
pharveyBobC_UK: everything in data/ and pub/ except System
yes
[13:01]
BobC_UKCool. I think I'll restart.... after lunch! [13:01]
pharvey(well, I think we have slightly different WebSearch, WebIndex, etc. topics but you can easily recover the Foswiki 1.1 versions into the Main web after you've migrated your data)
pharvey can't remember if Foswiki 1.0 used INCLUDEs in the Web* topics or not
[13:02]
BobC_UKAh, so just unzip the Main back into the copied stuff? [13:02]
pharveyBobC_UK: yes [13:03]
BobC_UKCool x 2. Lunch! Thanks for your patience and help [13:03]
pharveynp, sorry it's been such an arduous journey [13:03]
BobC_UKI find that on arduous journeys one tends to learn more! [13:03]
pharveyindeed
:)
[13:04]
BobC_UK:-)
l8rs
[13:04]
***BobC_UK has quit IRC ()
Lynnwood has joined #foswiki
Lynnwood has quit IRC (Changing host)
Lynnwood has joined #foswiki
[13:05]
...... (idle for 26mn)
HolstenerLieseldamn, I forgot to add a few <nop>s to Item:9833, thus locking myself out. any admins in the house nice enough to correct this? [13:31]
foswikibothttp://foswiki.org/Tasks/Item9833 [13:31]
***favioflamingo has joined #foswiki
leik has quit IRC (Ping timeout: 245 seconds)
[13:37]
BabarSvenDowideit is editing Item9833.
so I can't help you :)
[13:45]
***herkulez has joined #foswiki [13:45]
SvenDowideitSvenDowideit is finding that horrid non-feature annoying [13:46]
BabarI could force-update and merge, but... I'd rather wait :) [13:47]
herkulezHello. Ich bekomme ständig eine Email "Tasks are waiting for feedback from you" was soll ich tun? [13:47]
SvenDowideiti mean the one hwere anyline containing %META... is stolen and hidden as though it were real meta [13:47]
Babarherkulez: check the task, and provide feedback, if possible. If not, then simply remove yourself from the list
SvenDowideit: ah right...
[13:47]
herkulezhow to remove? [13:48]
HolstenerLieselSvenDowideit: thanks! man, that was acutally pretty stupid. i think i'll stop doing anything at all today, things might get broken. [13:50]
SvenDowideitHolstenerLiesel na, don't
its a really dumb mis-feature, that i recal having a hand in a very long time ago - because back then, i used to misuse it alot
long time - maybe 2001
alot of really poor hack decisions stem back to then
[13:52]
***leik has joined #foswiki [13:55]
HolstenerLieselI'm going to have to correct this a little bit, now just just to make sure: putting
<nop>%META:PREFERENCE{name="ALLOWTOPICVIEW" title="ALLOWTOPICVIEW" type="Set" value="SomeGroup"}% in there won't do this again? I just tested it alright on my 1.1.0 installation.
[13:55]
SvenDowideityup [13:56]
HolstenerLieselok then [13:56]
SvenDowideitits only if you have a line that looks like a valid META [13:56]
Babarherkulez: at the very bottom of the page you should have a field called WaitingFor with your wiki name in it. Edit the page, remove yourself from the field
herkulez: but if you could tell us which bug item, that would help :)
[14:00]
herkulezBabar: http://foswiki.org/Tasks/Item1319 I'm NicolasWrobel. I'll remove myself since I don't know what kind of feedback you want... [14:01]
BabarAh no, then better put it into Confirmed or WorkingOn [14:04]
***leik has quit IRC (Ping timeout: 265 seconds)
HolstenerLiesel has quit IRC (Quit: Leaving.)
[14:05]
Babarhum, BeingWorkedOn
it's true the WaitingForFeedback is wrong.it's just waiting for someone, hopefully you, to do it :)
[14:08]
***gac410 has joined #foswiki [14:12]
herkulezBabar: doesn't it work anymore with the current release? [14:13]
***BobC_UK has joined #foswiki [14:22]
............ (idle for 57mn)
BobC_UKI think I've spotted an incorrect instruction in System/PatternSkinColorSettings in Step 1 [15:19]
***favioflamingo has quit IRC (Ping timeout: 250 seconds) [15:32]
Babarherkulez: sorry? [15:36]
BobC_UKCould someone please help me test my understanding of steps 1 and 3 in System/PatternSkinColorSettings [15:37]
BabarFoswiki:System/PatternSkinColorSettings
Babar is lazy
[15:39]
foswikibothttp://foswiki.org/System/PatternSkinColorSettings [15:39]
BobC_UKYes, step 1 states ...... USERCOLORSURL = %PUBURLPATH%/%SYSTEMWEB%/PatternSkin/my_colors.css
Step 3 implies the attachement is saved to PatternSkinTheme, not PatternSkin
[15:40]
Babardoes it?
%STARTATTACH{"my_colors.css" web="%SYSTEMWEB%" topic="PatternSkin"}%
[15:40]
BobC_UKyes it saves it to the Them directory [15:41]
Babarlooks like PatternSkin to me
maybe this changed in 1.1?
right, it changed on trunk
so you're right, that's a typo
please open a bug request, I'll commit a fix :)
[15:41]
BobC_UKYeah, couldn't figure out why CSS wasn't being picked up. [15:43]
Babarthis has been changed in 1.1 [15:43]
BobC_UKOK. [15:43]
***AndreU has quit IRC (Quit: AndreU) [15:43]
BabarFoswikirev:7102 and Item:8840 [15:46]
foswikibothttp://trac.foswiki.org/changeset/7102
http://foswiki.org/Tasks/Item8840
[15:46]
BabarArthur simply forgot to update Step 1 :)
oh, this bug isn't closed
so I can commit on it
but it might be good opening a bug report thought
[15:46]
gac410Sounds like Lavr missed this for the release notes - if it has been partially committed into 1.1.0, then we probably need a new bug.
Babar, your recursion loop with my fixes to Item:9810 - any particular group layout trigger it? I don't seem to be able to recreate it.
[15:48]
foswikibothttp://foswiki.org/Tasks/Item9810 [15:49]
Babarit has been partially committed, but it looks like it wsn't closed
BobC_UK: could you please create a bug report?
Babar will use the old one for trunk, but would like a dedicated one for the release branch. Seems cleaner that way
[15:50]
***AndreU has joined #foswiki [15:50]
gac410Babar, never mind - recreate loop by crossing groups AGroup includes BGroup which includes AGroup [15:51]
Babargac410: it was the unit tests
running the full suite, with no LSC, gave this output
[15:52]
gac410Oh - okay, I was only running the individual unit tests - they take so darn long. [15:52]
Babaryeah, agreed
but often there are odd correlations between the tests
[15:52]
gac410In any event, my fix clearly has a bit of a bogosity factor [15:53]
***Babar sets mode: +o gac410 [15:53]
gac410Okay - that was a dumb error - recursion fixed. moved a variable declaration inside the loop that needed to be on the outside. [15:56]
Babaroh, I saw that while reading the code, and thought... no... it's too obvious :) [16:08]
gac410Now my next question - in UserMapping.pm sub isInGroup, there is a $scanning variable passed internally to block recursion there.
For TopicUserMapping, I don't believe that recursion at this point is possible, since the groups are flattened and cached. So isInGroup should not see a nested group
However I'm wondering if it is still needed for other mappers. And if there is some other trick other than passing it as an internal parm. Which makes it very difficult to add the "noexpand" parametere
[16:09]
LavrHi All. I am back from South Africa. Long 12 + 2 hours of flying and no sleep. Cannot sleep sitting up in an airplane. [16:13]
gac410Welcome back [16:14]
LavrWe have some nasty bugs on 1.1.0 so my plan is to get 1.1.1 out as fast as possible. But natually we need the urgent bugs fixed.
The important "family" of bugs are the ones related to the groups and access rights of groups.
I will try and create a test environment with more groups and nested groups so I can help testing the fixes.
This will be tomorrow I start that.
[16:14]
gac410Well Babar, fixing the recursion totally broke my nested groups code. So my code appears to either nest and occasionally loop, or neither. :-( [16:17]
LavrI am not going to add the updated CommentPlugin to 1.1.1. I am not even sure it should go into a 1.1 release after having seen the bugs reported on it. So please do not spend time on that. Better to get all the group related stuff fixed fast.
I know the 1.1.0 has some nasty bugs and that it would have been better without these bugs, but overall the number of bugs reported in areas like Wysiwyg editor and other UI thing are not arriving in great numbers. So once we get the group code and group UI fixed, 1.1.1 will be pretty good.
[16:17]
gac410I wish we could figure out why StringifierContrib totally breaks FindExtensions on some systems, but not everybodys. [16:20]
Item:9311 - I have a horrid patch which causes Dependency.pm to bypass the toxic extension - but hate to commit it rather than understand the issue. [16:25]
foswikibothttp://foswiki.org/Tasks/Item9311 [16:25]
LavrSo you need to have installed StringifierContrib to trigger the issue?! And it only fails on some. Sounds like the contrib relies on some CPAN lib that returns different result depending on version. I have seen that happen before. [16:27]
gac410Its the Module::Pluggable - which dynamically loads all the stringifier plugins. But why it actually hangs ... not a clue. [16:28]
Babaryeah, do we have someone who has the broken version, and reported all versions used? Like perl, Module::Pluggable, etc... ? [16:29]
BobC_UKAnd just to add to your woes, I've done a complete new install from scratch, (learning from my last mistakes!) and ActionTrackerPlugin does not work. With topic from my previosu 0.9 install, when I load the topic I get : Can't locate object method "new" via package "Foswiki::Plugins::ActionTrackerPlugin::Action"
I've reinstalled the plugin but no effect
[16:37]
LavrI saw that there are some compatibility issues with ActionTrackerPlugin. Which is strange because the one that has changed the most in APIs is the one that maintains ActionTrackerPlugin. [16:40]
Hmm. ActionTrackerPlugin seems to be active on my 1.1.0 installation. [16:45]
BobC_UK:blush: found the problem..... I forgot to transfer the settings from my old locallib.cfg to the new install. [16:45]
Lavr%ACTIONSEARCH seems broken on the plugin page though. I do not have much Action examples in my home site. [16:47]
BobC_UKI was just going to say, the ActionSeacrh doesn't work
I use it to populate as ingle topic for a persons actions across all topics in a web
[16:47]
LavrThere is a bug report on ActionTrackerPlugin. I saw it briefly on my ipad while I was on vacation so I could not follow up on it. [16:49]
gac410I'm pretty sure it was also a report that ACTIONSEARCH is not working... iirc [16:50]
BobC_UKhttp://foswiki.org/Tasks/Item9825 [16:50]
Lavr"Looks like the bug is due to the changes in Foswiki::Func::searchInWebContent. It returns an iterator now, while it used to return a hashref." ????!!
Who changed this Func? THis is API. We cannot change API like this!
[16:51]
BobC_UKOnly reported yesterday and a fix a couple of days away. Just have to wait I suppose! [16:52]
LavrIn my view it is the Func API that needs fixing. Noone should change Func API without a feature proposal and I do not remember one for this one.
I wonder how many plugins are broken because if this.
[16:55]
***AndreU has quit IRC (Quit: AndreU)
EugenMayer has quit IRC (Quit: Leaving.)
[16:58]
LavrSvenDowideit you broke the API for searchInWebContent. You broke one of the most important business application extensions in the process. When you do guys learn?
http://trac.foswiki.org/changeset/6744
[17:03]
BobC_UKWhere would I be able to find then change the class="TwistyPlugin", the boxes aren't displaying nicely for me..... http://img685.imageshack.us/img685/4934/snagprogram0046.png [17:04]
***HolstenerLiesel has joined #foswiki [17:07]
BabarLavr: __WARNING: does not return a hash - returns an iterator (else you will crash your server needlessly)__
Please report a Task item if you're using the non-iterator interface)
so, that's odd
[17:10]
LavrVery odd.
I am doing a grep to see where we have calls to this API
it is MANY plugins
[17:11]
Babaryeah, but then you have to check if they expect a hash or an iterator... [17:12]
BobC_UKAlso....(I'm on 1.1.0) When I do a WebSearch and the results include multiple webs, the "Results from %WEB% ...." is not color coded for the color of the web. Has this chnaged or is this connected with the PatternSkin my_colors.css change? It is much easier to undesrtand results when the web colour is presented above the results. [17:12]
LavrUnless updated after Sven's change they can be assumed broken [17:12]
gac410perl question - is there a way to determine that a subroutine has called itself? ie how to determine if this is the first/top call in a recursion? Google has not been my friend today.
Without playing tricks by passing extra parameters?
[17:13]
Babargac410: sure... use caller() [17:13]
gac410Ah - okay thx! [17:13]
Babarlike... my (undef, undef, undef, $sub) = caller 1;
if it's just one call
BobC_UK: that's odd I cannot find this CSS class being defined anywhere in the source
[17:14]
BobC_UKYeah, I was struggling with it
Perhaps I should just add it to my_colors.css as a one off
[17:16]
BabarLavr: so ATP, BioKb, Object, TCP, Tags and ValidateUrls [17:18]
LavrThe list of possible broken plugins are BioKbPlugin, ObjectPlugin, ValidateUrlsPlugin, ActionTrackerPlugin [17:18]
Babarnot that many after all
Babar would add TagsPlugin and TWikiCompatibilityPlugin :)
ah no, Tags redefines it
[17:18]
LavrThe remaining call the searchInWebContent elsewhere than through Func it seems. TCP is hopefully updated but yes we need to check that also.
If plugins called directly in Store then they use none-API. That we cannot prevent or be responsible for. But we are responsible for the Func and Meta API.
[17:19]
BabarTCP is not updated. TCP maps the TWiki searchInWebContent to Foswiki one, which is bound to fail
but tbh, I couldn't care less about TCP :)
(for this very specific function)
The return value is a reference to a hash which maps each matching topic
name to a list of the lines in that topic that matched the search,
as would be returned by 'grep'.
_WARNING: does not return a hash - returns an iterator (else you will crash your server needlessly)__
Sven... be coherent please :)
[17:20]
***Dan_R has joined #foswiki
MTempest has quit IRC (Quit: Leaving)
[17:24]
LavrI do not understand how to use this Func now. In the old one there was a good example. It has been replaced by an example how to call the function but now how to use the result.
Is there any reason not to change the result back to the format it was in 1.0.10 and earlier?
[17:24]
Dan_RIs there a way to prevent foswiki from creating links to local drives? People keep adding content that includes hard links to stuff like S:\HR\Doc.doc and the VP's don't have an S: drive so they scream [17:26]
BabarLavr: yes, because the all point of Sven's redesign was to use resultsets
Dan_R: a cluebat? :)
[17:26]
Dan_Rand the people writing this use IE, so they're links work on their browser and they don't understand why they shouldn't create local file links [17:26]
Babaralso, S:\... shouldn't create links [17:26]
Dan_Rnotice I said hard links...
they're forcing it, and it works
for some
[17:27]
Babarah... yeah... it works, only on IE :) [17:27]
Dan_Rexactly [17:27]
BabarFF finds those insecure [17:27]
Dan_Rwhich is a broken POS, I constantly get issues with authentication because IE screws up the cache [17:27]
Babarhonestly, apart from hacking the Render.pm... [17:27]
Dan_R*cookie
I agree with your first suggestion, but that puts me in jail..
[17:27]
LavrThen we should have added a new Func call and left the old untouched. OR made the function so you can call it multiple ways OR fixed all the plugins that uses the old format AFTER a feature proposal that enabled plugins authors to have their say. We cannot just check in API changes in large checkins where noone notice such a change. [17:28]
Babarthen hack the renderred so that [[:space:]]\[\[S: is rendered as *DO NOT DO THAT* [17:29]
Dan_Rthat sounds so awesome [17:29]
LavrDan_R - making a small plugin yourself that does this is not too hard to do. [17:29]
Dan_RI'm not a coder :( [17:29]
***railmeat has joined #foswiki [17:30]
Dan_RI make vbscript and that's it, not a real coder
but plugins have the ability to modify the rendering engine?
[17:30]
BabarBabar loves the code in Render.pm. Extract: [17:31]
Dan_Ror is it applied after [17:31]
BabarSMELL: why is this available to Func?
5 lines later:
# SMELL - shouldn't it be callable by Foswiki::Func as well?
[17:31]
Dan_R... heh, great memory [17:31]
LavrPeople that check in SMELLs and have not removed them themselves within 1-2 weeks should be spanked. [17:31]
Babaroh Dan_R, I think I have a solution for you
ugly hack, but might work...
Dan_R: can you please show me one example of a hard-coded link?
like the [[...S:/ part, just the beginning
[17:32]
Dan_Ron sec
<a target="_blank" href="O:\Custo
[17:34]
Babarouch.
then no, sorry, my idea won't work for this case :(
[17:35]
Dan_Ryeah, not even a wiki link [17:35]
Babarwould it have been a wiki link with file:/// it might have
(idea was: remove file from the LinkProtocolPattern in configure)
[17:35]
Dan_RI don't know how href="O:\asdf" should even be possible
it's horrid
[17:36]
BabarLavr: so, as we're too late for feature proposal, what about fixing all plugins? :)
you open the bug report, I'll fix the code :)
[17:36]
LavrThe bug to fix all 4 plugins is http://foswiki.org/Tasks/Item9825 [17:38]
Babarand the example in the documentation is complete
my $result = Foswiki::Func::searchInWebContent( "Slimy Toad", $web, \@topics, { casesensitive => 0, files_without_match => 0 } );
[17:38]
LavrBut please note that plugins will still need to work also for 1.0 [17:38]
Babarthat's the call
while ($matches->hasNext) { my $webtopic = $matches->next;
that's the use
[17:38]
LavrSo the code in the plugins need to have different behavior depending on the plugin version. [17:38]
Babarif I try a fix on ATP, will you be able to test it? [17:39]
LavrIt would help noobs like me if you could improve the Func example with the few extra lines of code that shows how to use the result. [17:39]
Babar(as I've never used any of the incriminated plugins, it's hard for me to test) [17:39]
LavrI have an ATP at least running.
So start with ATP.
[17:39]
BabarLavr: as said, there is an example already :)
ok
[17:39]
***AndreU has joined #foswiki [17:39]
gac410Babar - I think I have the recursive issue cleanly fixed - but getting "encouraged" to go finish painting the house - I'll pick it back up tonight, clean it up a bit and commit again. [17:40]
LavrAh. The example indents the code so it looks like it is the call to the function that continues on more lines.
I need to join my lovely wife (SWMBO) and eat now. I will be back late tonight after prime time and for sure be working on Foswiki Sunday.
[17:40]
BabarBon appetit !
gac410: paint it as good as you fix the bug :)
hum, I could optimise the code for ATP, but I'd rather not ... :)
[17:42]
hum... the code looks flawed :) [17:51]
ok, code isn't flawed, doc is :) [17:56]
***EugenMayer has joined #foswiki
terceiro has joined #foswiki
[18:00]
BabarBobC_UK: did you open a bug report for the PatternSkin "typo"? [18:07]
***GithubBot has joined #foswiki [18:11]
GithubBotfoswiki: master OlivierRaginel * 10a934a (1 files in 1 dirs): Item8840: PatternSkin color customisation has been moved to PatternSkinTheme, but Step 1 was still pointing to PatternSkin. Thanks BobC_UK for pointing that out
foswiki: master OlivierRaginel * 6fce904 (2 files in 1 dirs): Item9825: Sven in his wonderful (that is _NOT_ sarcastic) search re-implementation changed Foswiki::Func::searchInWebContent. As it is not possible to keep the old functionnality, all plugins using this Func needs to be updated. Starting with ActionTrackerPlugin
foswiki: master commits e1efa90...6fce904 - http://bit.ly/d0PcpA
[18:11]
***GithubBot has left [18:11]
Babaroh no, what did I just do?
trying to fix, hoping nobody will update from github in the meantime
[18:11]
***GithubBot has joined #foswiki [18:13]
GithubBotfoswiki: Item9825 OlivierRaginel * 10a934a (1 files in 1 dirs): Item8840: PatternSkin color customisation has been moved to PatternSkinTheme, but Step 1 was still pointing to PatternSkin. Thanks BobC_UK for pointing that out
foswiki: Item9825 OlivierRaginel * 6fce904 (2 files in 1 dirs): Item9825: Sven in his wonderful (that is _NOT_ sarcastic) search re-implementation changed Foswiki::Func::searchInWebContent. As it is not possible to keep the old functionnality, all plugins using this Func needs to be updated. Starting with ActionTrackerPlugin
foswiki: Item9825 commits 0000000...6fce904 - http://bit.ly/cle2XU
[18:13]
***GithubBot has left
sreher has quit IRC (Quit: sreher)
[18:13]
BabarLavr: untested: http://github.com/foswiki/foswiki/commit/6fce904c86d22909bd2ef5855a6d271d5e4dfbc5
I chose deliberately to fix both codes differently, so:
1. You can see the differences and different solutions
2. One can easily ditch old support with one implementation, but it would duplicate way too much for the other. So...
and again, I haven't tested it. I'll try to be online again later
[18:14]
***railmeat has left [18:19]
..... (idle for 23mn)
Dan_R has quit IRC (Quit: Don't push the red button!)
railmeat has joined #foswiki
[18:42]
BobC_UK@ Barbar re "BobC_UK: did you open a bug report for the PatternSkin "typo"?" < Yes I did
http://foswiki.org/Tasks/Item9834
[18:58]
BTW, the Freemind extension won't install because it cannot find the package on the server. [19:06]
....... (idle for 31mn)
***railmeat has quit IRC (Quit: railmeat)
herkulez has quit IRC ()
[19:37]
BabarBobC_UK: thanks :)
I'll commit the fix in a minute
[19:50]
***GithubBot has joined #foswiki [19:55]
GithubBotfoswiki: master OlivierRaginel * dd4f42b (1 files in 1 dirs): Item8840: PatternSkin color customisation has been moved to PatternSkinTheme, but Step 1 was still pointing to PatternSkin. Thanks BobC_UK for pointing that out ... - http://bit.ly/akzgRu [19:55]
***GithubBot has left [19:55]
GithubBot has joined #foswiki [20:02]
GithubBotfoswiki: Release01x01 OlivierRaginel * 8f3a25b (1 files in 1 dirs): Item9834: PatternSkin color customisation has been moved to PatternSkinTheme, but Step 1 was still pointing to PatternSkin. Thanks BobC_UK for pointing that out ... - http://bit.ly/dw6XoT [20:02]
***GithubBot has left [20:02]
Babardone ;) [20:02]
BobC_UKIs there any reason why when a search returns multiple finds off different webs (FindElseWhere), that the headers for each web section in the search no longer has the web color as the background? [20:08]
BabarBabar thought we wanted to remove the web color thingy
but... no, I dunno if it's on purpose or not, you'd have to ask Arthur
[20:10]
wbniv_wbniv_ does not [20:11]
BobC_UKSo it's not just me then thinking I've fusked it up? [20:11]
wbniv_dunno; i believe arthur was trying to bring it back in more places / slightly more prominently, but it could have been removed from any spot as much as added to any other
tho i'm not sure how FindElseWhere factors into this... ?
[20:12]
BobC_UKI don't think it does. I wasn't sure if the "color" was associated with it or not. The Foswiki site searches still show the colors as I've grown to know them in 0.9.0, and they are a huge help in quickly finding the document in question. [20:14]
wbniv_ah ok.
try trunk.foswiki.org
does that show what you think it should be, or what you have on your installation?
howdy Babar
[20:16]
BobC_UK@ wbniv.. re trunk.foswiki.org. Yes, that looks like what I have. No colours on the web section. http://img87.imageshack.us/img87/4237/snagprogram0047.png [20:22]
wbniv_i see
disappointing
you could file a task as an enhancement request at trunk.foswiki.org/Tasks O;-)
[20:23]
BobC_UK:) BTW, when I log in with my Foswiki sign on,is it correct that the Truk appears not to know my user name? http://img821.imageshack.us/img821/9772/snagprogram0049.png [20:25]
Babaryes
Main in trunk is not Main in f.o
because Main and System are system webs, so they're not shared
[20:26]
BobC_UKOK, just wondered as I had that issue before with the Config not having the UseLoginName set. [20:26]
***terceiro has quit IRC (Quit: Saindo) [20:26]
BobC_UKBTW, is everyone OK with me asking questions in here? Just want to make sure I'm doing the right thing! [20:27]
Babarhey, that's what we're here for
(more or less, don't forget the fun ))
[20:29]
BobC_UKBobC_UK asks Babar to roll over for a tummy scratch. ;) [20:29]
Babar*purrs* [20:30]
BobC_UKI solved that button issue with Twisty by adding this to the my_colors.css .twistyPlugin { margin:10px 0px 5px 0px; }
http://img826.imageshack.us/img826/6789/snagprogram0050.png
[20:31]
Babarinteresting... could you please open a task and assign it to Arthur, he might want to have a look [20:45]
BobC_UKOK, will do. [20:47]
..... (idle for 22mn)
***usepidgin has joined #foswiki [21:09]
BobC_UKBobC_UK waves goodbye for now. [21:14]
***BobC_UK has quit IRC ()
ktwilight_ has joined #foswiki
ktwilight has quit IRC (Ping timeout: 245 seconds)
[21:15]
......... (idle for 42mn)
SvenDowideitBabar your fix looks exactly what i was just thinking to do
i wonder why your commit isn't showing up on the task
[21:59]
***SvenDowideit sets mode: +oooo AndreU EugenMayer ktwilight_ Lynnwood [22:04]
SvenDowideitmmm, i wonder if it might be better to do
$grep = new Foswiki::ListIterator(keys(%$grep)) if ref($grep) eq 'HASH'
that way we convert to iterator in one fel swoop
[22:11]
SvenDowideit will commit to trunk my proposed docco and change [22:24]
gac410SvenDowideit: I was wondering instead of ListIterator might detect if the list is a hash or array and handle it automatically. [22:26]
SvenDowideitwhat ould it fo?
what... would it do
i wondered myself
but then, should it use keys()
values()
or thunk to list
none of which are good answers?
SvenDowideit is loving git atm, but still struggleing with the slightly smaller than standard keyboard on the x61
[22:26]
***gac410 has quit IRC (Ping timeout: 255 seconds) [22:30]
SvenDowideitits not good hoping that Babar doesn't sleep, and has a look at what i commited :/
SvenDowideit goes to join breakfast
[22:32]
.... (idle for 19mn)
***gac410 has joined #foswiki [22:51]
SvenDowideitgac410, i wondered about that too, but didn't think of a good way to automatically select keys(), values() or @% [22:58]
***HolstenerLiesel has left [22:58]
gac410Yeah - I had blinders on - I had forgotten about keys/values and was only thinking of the narrow use with the users = key of has
has/hash/
I'll run a full unit test suite, then try to change from noexpand to expand, and commit again. So far so good.
Tripped up a bit because one routine is called recursively once removed A calls B calls A -
[22:59]
***EugenMayer has quit IRC (Quit: Leaving.) [23:01]
gac410hmmm HTMLValidationTests::verify_switchboard_function_statistics_default seems to be looping. [23:04]
***BobC_UK has joined #foswiki [23:04]
gac410no - just running forever 100% cpu - it finally moved on [23:05]
BobC_UKHi-de-Hi campers! My Kinosearch is giving me a headache on my new 1.1.0 install. When I run a KinoSearch I get this: Can't locate object method "mapToIconFileName" via package "Foswiki"
I'm sure I have all the Contrib variables set as 0.9 so not sure what this could be.
[23:05]
SvenDowideitBobC_UK, ooo, another of my wonderful bugs
makeatask
[23:08]
gac410Look like another plugin is using an undocumented (and removed) function call. ObjectPlugin also has the same issue. [23:08]
BobC_UKAh, so it's not me then! ;) [23:10]
gac410no [23:10]
BobC_UKBobC_UK goes to make a cup of tea. [23:11]
gac410There is a good reason the only calls a plugin should use are the documented API - Foswiki::Func, Foswiki::Sandbox and Foswiki::Meta. [23:11]
SvenDowideityup, the problem is that there's a limit to the expressiveness of the published API
mind you, the really big tissue is that just when I'd have been testing all the plugins, i've been taken away from foswiki work to look after kids
[23:13]
gac410Thought the solution should be to "add a func" and unit tests. [23:14]
SvenDowideitso, its quite largely my fault
we can't afford to add Func APIs for ever little need
as that ties the core down too much tto
[23:14]
gac410Well, not really, as the Func unit test would fail, and an alternate implementation could be handled in Func instead of in each plugin that needed the access.
But I do agree that the need should be justified and legitimate with no reasonable alternatives.
[23:16]
***terceiro has joined #foswiki [23:20]
AndreU has quit IRC (Quit: AndreU)
AndreU has joined #foswiki
AndreU has quit IRC (Ping timeout: 245 seconds)
[23:28]
..... (idle for 23mn)
BobC_UKAlthough the KinoSearch appears to be temporarily out of action, the tools/Kinoindex and tools/kinoupdate are working. I do get this message on update for one topic which has 3 x pdf's attached:- "OopsException(attention/no_form_def web=>Main topic=>WebHome params=>[My_Topic,TestForm])mywebsignon@myweb.com ". Is this an unconnected problem? Or is this something else? [23:56]

↑back Search ←Prev date Next date→ Show only urls(Click on time to select a line by its url)