#foswiki 2011-11-08,Tue

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

WhoWhatWhen
***pharvey has joined #foswiki
terceiro has quit IRC (Remote host closed the connection)
MartinCleaver has quit IRC (Quit: MartinCleaver)
[00:38]
.... (idle for 18mn)
dj_segfault has joined #foswiki [00:59]
...................................... (idle for 3h7mn)
pharvey has quit IRC (Remote host closed the connection) [04:06]
......... (idle for 41mn)
GithubBot has joined #foswiki [04:47]
GithubBot[foswiki] foswiki pushed 1 new commit to master: http://git.io/pmgiVQ
[foswiki/master] Item10724: Make MCESKIN and MCESKIN_VARIANT preference settings - PaulHarvey
[04:47]
***GithubBot has left [04:47]
FoswikiBothttp://foswiki.org/Tasks/Item10724 [ Item10724: Update to TinyMCE 3.4.5 ] [04:47]
***GithubBot has joined #foswiki [05:00]
GithubBot[foswiki] foswiki pushed 1 new commit to master: http://git.io/IB_OCw
[foswiki/master] Item10724: Update to TinyMCE 3.4.7 - PaulHarvey
[05:00]
***GithubBot has left
pharvey has joined #foswiki
pharvey has quit IRC (Changing host)
pharvey has joined #foswiki
[05:00]
...... (idle for 26mn)
Lynnwood has quit IRC (Quit: Lynnwood) [05:27]
...... (idle for 27mn)
gac410 has left
dj_segfault has quit IRC (Quit: Leaving.)
[05:54]
......... (idle for 43mn)
leik has quit IRC (Quit: Leaving) [06:40]
GithubBot has joined #foswiki [06:46]
GithubBot[foswiki] foswiki pushed 1 new commit to master: http://git.io/fgMLEw
[foswiki/master] Item10724: Bump release version - PaulHarvey
[06:46]
***GithubBot has left [06:46]
FoswikiBothttp://foswiki.org/Tasks/Item10724 [ Item10724: Update to TinyMCE 3.4.5 ] [06:46]
***pharvey has quit IRC (Quit: ChatZilla 0.9.87 [Iceweasel 7.0.1/20111015075007]) [06:54]
GithubBot has joined #foswiki [07:00]
GithubBot[foswiki] foswiki pushed 1 new commit to master: http://git.io/7ngr7A
[foswiki/master] Item2174: Testing release - PaulHarvey
[07:00]
***GithubBot has left [07:00]
FoswikiBothttp://foswiki.org/Tasks/Item2174 [ Item2174: WysiwygPlugin removes line breaks ] [07:00]
***CDot has joined #foswiki [07:12]
.... (idle for 17mn)
packoman_ has joined #foswiki [07:29]
denisr has joined #foswiki [07:41]
............... (idle for 1h13mn)
dnavarro has joined #foswiki
MartinRowe has joined #foswiki
[08:54]
pharvey has joined #foswiki [09:04]
CDotpharvey: do you recall doing anything "special" in WyswiygPlugin or TMCEPlugin to support charset conversions?
CDot is poking around the edges of requests/responses, trying to work out what the "best" behaviour would be
[09:07]
pharveyCDot: it's all a bit wonky at the moment. I've actually been poking that code today. [09:08]
CDot"all a bit wonky" == SNAFU? [09:08]
pharveyfar from "best" :) [09:08]
CDotgotcha. Well, I'm not testing, just asking so I can find code to go and read
cos ATM I have Foswiki::Request working in unicode, and response too, but the bit in the middle is still a bit.... wonky
[09:09]
FoswikiBothttp://trunk.foswiki.org/System/PerlDoc?module=Foswiki::Request [ (Foswiki login) PerlDoc ] [09:10]
pharveyCDot: as long as the bits in the middle are working with unicode strings, it should be ok? [09:11]
CDotI don't think they can. Mainly because of plugins. [09:11]
pharveySo we're doomed forever? :) [09:11]
CDotthere is just so much code that assumes byte strings - or rather, it assumes there will be no problem with ops such as "open file and print to it"
if you use unicode, the old "wide character" keeps biting
[09:12]
pharveythese would be plugins opening files to working? [09:13]
***leik has joined #foswiki [09:13]
CDotyeah; but not just file ops. Other ops - quite unexpected ones - also trigger wide char problems [09:13]
pharveymay I ask which? [09:14]
CDotmany. Just as an example, md4_hex
md5
gives a "wide character in subroutine entry"
[09:15]
pharveyYes, I remember this one [09:16]
CDotalso in Encode::encode [09:16]
pharveybut that's by design, isn't it? The md5 function doesn't understand characters [09:16]
CDotMIME::Base64 as well
right, I know, but at the moment that's no problem because we never feed it characters
because all strings are byte arrays, not character strings
as soon as we use unicode - i.e. character strings - the game changes
and callers have to "know" to convert
and sometimes it is very, very, very obscure what is going on - even by perl's terms
^sometimes^usually
[09:16]
pharveyYes, I know. But the current situation is also unsustainable... it just has the added bonus of being wrong [09:18]
CDotso we have a choice; keep all strings as utf-8 encoded byte strings, and lose the ability to use locale-based collation etc
or convert to unicode, and deal with it
CDot notes that conversion to/from unicode is not fast in <5.12
[09:18]
pharveyWe're really going to suffer, if we don't standardize on unicode character strings in core. Double/triple de/encoding problems will linger
It's easy for a plugin to test if a string has utf8 flag on
I don't want to see any more pass-the-"let's Encode::decode this one more time"-parcel :)
[09:21]
CDottrue. But detecting when a plugin has passed a string that is utf8 encoded but does *not* have the utf8-bit set is not possible
that can happen when a plugin - for example - ready utf8-encoded data from a binmode'd file
^ready^reads
[09:23]
pharveywe have $NO_PREFS_IN_TOPIC... can we do a $IS_UNICODE_AWARE
I guess that still doesn't help
[09:24]
CDotouch. Lotsa "does it speak unicode yes/no" code in the core :-(
if we go to unicode, we go to unicode, and accept that plugins will flame out.
[09:25]
pharveyOk. [09:25]
CDotI'm struggling a bit with what CGI - and therefore Foswiki::Request - does with unicode chars in requests. Whether all fields are decoded, left with %88 type chars, etc [09:26]
pharveyJust avoid CGI.pm 3.43 :) [09:27]
CDot'cos while it seems to make a reasonable job of converting an incompling utf8 request to unicode (if you $request->charset('utf-8') that is)
it makes a real arse of going the other way (wide character under fingernails error)
I thought 3.43 was the one that fixed all these problems?
[09:27]
pharveyno. 3.43 has a double-encode problem
on debian/ubuntu you need to install libcgi-pm-perl to upgrade it
[09:29]
CDotayayay [09:29]
pharveypharvey thought we had a configure warning for 3.43 [09:29]
CDotphew, I have 3.49 [09:30]
pharveyanyway, it's funny, because 3.43 ships with ubuntu 10.04 & also lenny [09:30]
CDotI need to start writing some "explore around charsets in requests" test scripts
cos I'm unsure of what does/does not work
[09:30]
pharveyCDot: I seem to recall that we don't binmode STDOUT to take "wide" chars, is that why CGI.pm fails with those warnings? [09:32]
CDotnah, it doesn't fail. The failure I'm fighting with is somewhere in the depths of the store; despite binmoding and utf8::downgrading the data, i still get a wide character when printing
CDot is getting a bloody forehead, and needs to do something different for a while.
[09:33]
pharveyis that a problem in PlainFile logger? I think I began to fix that
pharvey understands :)
[09:33]
CDotno, in Foswiki::Store::VC::RcsLiteHandler [09:33]
FoswikiBothttp://trunk.foswiki.org/System/PerlDoc?module=Foswiki::Store::VC::RcsLiteHandler [ (Foswiki login) PerlDoc ] [09:33]
pharveyahh [09:33]
CDotit's frustrating that ASSERT(!utf8::is_utf8($str)) is not enough to detect a wide char in $str :-( [09:34]
pharveywe have to utf8::upgrade all strings as early as possible - because if they only contain "boring" chars, it'll have utf8 flag off
pharvey re-reads what CDot said...
[09:35]
CDoty, but tracking down the source of strings that have *not* been utf8::upgraded is a bugger [09:37]
***ThomasPHaeck has joined #foswiki [09:41]
CDotCDot has gone, back l8r [09:43]
***jayen has joined #foswiki [09:47]
jayenHi. What's the best way to make the Main/WebHome page contain a login box? [09:49]
***terceiro has joined #foswiki [09:54]
......... (idle for 42mn)
pharveyjayen: you want a custom login box, or just the same thing TemplateLogin gives you (assuming you're using TemplateLogin) [10:36]
packoman_pharvey: I was trying to filter the fields that are shown in bibtex-entry dialog yesterday
that is trying to modify your code
what's wrong in the following(?):
\"($item = 'title') AND ('%qweb%.%qtopic%'/$item =~ '\w')\"
I thought that should only show the title and only in case it's not empty
instead it does not show anything...
[10:42]
pharveywhere is the code?
which topic
[10:45]
packoman_sorry
Cite-topic
the ajax part that querys the fields and formats them for output.
a sec
[10:45]
pharveyI guess I need to see this code in context
because my guess is you might need to delay the $item token
i.e. $dollaritem
[10:46]
packoman_http://pastebin.com/NzVyUsBg [10:46]
pharveyah [10:46]
packoman_it occurd to me later that it might make more sense and create a variable BIBLIOGRAPHY_SHOWN_FIELDS
so that BIBLIOGRAPHY_SHOWN_FIELDS = title, author, year, etc.
[10:47]
pharveyhrm [10:48]
packoman_and query that from there. that way we can determine the sequence of shown field
s
the code as I pasted it is from you
[10:48]
pharveypackoman_: it works here http://trunk.foswiki.org/Sandbox/TestTopic44740 [10:49]
packoman_except for the () around the IF-condition [10:50]
pharveymaybe it doesn't work on 1.1.3 [10:50]
packoman_yes sorry
that *does* work
[10:50]
pharveyno, works here http://foswiki.org/Sandbox/TestTopic44751 [10:51]
packoman_what's not working is the modification: \"($item = 'title') AND ('%qweb%.%qtopic%'/$item =~ '\w')\" [10:51]
pharveyoh
but if you only wanted the 'title' formfield, why bother iterating the list of all field names? :)
[10:51]
packoman_yeah. that's what I figured afterwards too... [10:52]
pharveypackoman_: the problem is you're not testing $item against the topic in the FORMAT - it's testing against BASETOPIC [10:52]
packoman_hence my statement above:
it occurd to me later that it might make more sense and create a variable BIBLIOGRAPHY_SHOWN_FIELDS
:)
[10:52]
pharveyso it should be [10:53]
packoman_oh. hm. but doesn't:"%QUERY{
"'%qweb%.%qtopic%'/fields.name"
[10:54]
pharvey\"('%qweb%.%qtopic%'/$item = 'title') AND ('%qweb%.%qtopic%'/$item =~ '\w')\" [10:54]
packoman_}%"
set the fieldname that's save in $item ?
[10:54]
pharveyyargh, I'm an idiot
$item will expand to a bareword
[10:54]
packoman_wouldn't that check that the *content* is 'title' [10:54]
pharveyin QuerySearch language, Foswiki tries to treat that as a field name
so we actually need ('$item' = 'title'), not ($item='title')
[10:54]
packoman_oh. I'll be damned :)
just I sec. I'll check
[10:55]
pharveybut yeah, a BIBLIOGRAPHY_FIELDS would be a saner idea :) [10:55]
packoman_ok. I'll probably do that, when I get around to it. [10:56]
pharveylet me make a little change to your pastebin [10:56]
packoman_be my guest :) [10:57]
pharveyhttp://pastebin.com/index/NzVyUsBg
oh damn
that's not it :)
http://pastebin.com/eUh9KeQc
BTW, \"('%qweb%.%qtopic%'/$item =~ '\w')\" doesn't need the parenthesis, but maybe you're already aware of that
[10:59]
packoman_yeah. it in the parenthesis cause I wanted a second condition for the fields
however with your modification that will not be necessary anymore
i.e. using BIBLIOGRAPHYPLUGIN_FIELDS
about that: what will happen BIBLIOGRAPHYPLUGIN_FIELDS contains field names not in the form of a cited topic
ie. book will not have an abstract...
will just omit it or throw an error
?
hrm. using BIBLIOGRAPHYPLUGIN_FIELDS brings me to another point that I have been ignoring up to now:
Till now I had my Cite topic in Main from where I will be doing most of my referencing. Now with BIBLIOGRAPHYPLUGIN_FIELDS I will want it in the Bibliography web so it finds that variable
how do make the Cite topic site-wide available?
[11:02]
pharveypackoman_: our format has an IF that tests to only emit if there's an $item field that contains word characters
if it doesn't exist it'll emit nothing (except maybe a newline...)
[11:08]
packoman_ok. great. [11:09]
pharveypackoman_: actually, there's another way
we can store the list of fields to be displayed in the TypeForm
add a new field to TypeForm
Fields
and then we can display %FORMAT{"%QUERY{'%qweb%.%qtopic%'/form.name/Fields}%"
[11:09]
packoman_oh. and then. people will be able to modify that?
cause it would be nice if the user could still configure it
[11:11]
pharveyI guess so, if they can be bothered to modify Eg. ArticleForm, BookForm, etc.
I guess I'm just brainstorming, not really thinking it through :)
[11:12]
packoman_hehe. well the nice thing about BIBLIOGRAPHYPLUGIN_FIELDS would be one setting for all forms in one location
if there's nothing speaking against that I think I'd prefer that.
but I am not sure...
[11:12]
pharveyAs the primary user of this application (until I get around to upgrading our site), I trust your decisions :) [11:15]
packoman_hehe. too kind. :)
it just all depends. the idea behind my consideration was about easiness of configurability. but probably only people will use foswiki (along with this) that at least know how to modify a form
or is that a misconception?
[11:15]
pharveyit will be modifying the values in a form, not the form itself
I mean, if you wanted to store the field list in the ArticleForm, BookForm, etc. To have a per-type field list
each of those forms has a TypeForm attached
It's not very common to have a Form attached to a Form
[11:18]
packoman_yeah. so you would set default values of displayFieds formfield. right? [11:20]
pharveyso that might just be confusing for people, I dunno
I think so
I've done this kind of thing in other wiki apps
[11:20]
packoman_hm. honestly I stumbled across the type form thingy when looking at the forms. but up to know I wasn't even aware that it is another form and of what it does
:)
hm. anyway. if I do * set BIBLIOGRAPHYPLUGIN_FIELDS = title, author
inside the Cite-topic itself should your posted code already work?
[11:21]
pharveyno [11:23]
packoman_oh.
why not?
[11:24]
pharveySet statements have scope
the scope of a Set statement is relative to the BASETOPIC, i.e. the requested topic in the URL
[11:25]
packoman_oh.
did not know that
so. just one more question: how do I make the Cite topic available site-wide?
so that I can do Cite::SomeThing from Main and have Cite in the Bibliography web (where all the settings should also end up...)
[11:26]
pharveypackoman_: I'm afraid I haven't thought of a good mechanism to do that, yet. On the bright side, you can create a barebones Cite topic that's just got a %TMPL:INCLUDE{"Foo.Cite"}% in it, which will absorb any template defs from a single place
in other news, I need to change our [[SemanticLinksPlugin_DEFAULTWEB::Foo]] to a * Set SEMANTICLINKSPLUGIN_DEFAULTWEB setting instead
then we can set this from Main.SitePreferences
[11:30]
packoman_ok. that would be nice [11:31]
pharveyactually, I can probably do that right now
as my last action before going to sleep :P
[11:32]
packoman_hehe. I'll let you go
;)
about the %TMPL:INCLUDE{"Foo.Cite"}%
[11:32]
pharveypackoman_: actually, why don't I make DEFAULT_PROPERTYWEB and DEFAULT_VALUEWEB
this would kill two birds with one stone
[11:33]
packoman_what will happen with the AJAX sections that are defined in the Cite topic that would be living in the Bibliography web and included through TMP:INCLUDE [11:33]
pharveyand no need for %TMPL:INCLUDE Cite topics all over the place [11:33]
packoman_?
not sure I understand
DEFAULT_VALUEWEB = bibliography and DEFAULT_PROPERTYWEB = Main in my case?
[11:33]
pharveyyes
Just remember: dont' use %TOPIC% or %WEB% from inside a %TMPL:DEF unless you really mean it - it doesn't expand into the topic which contains the DEF
so if you're building an AJAX URL you'll need to hard-code the topic name
[11:35]
packoman_is there a way to make Cite truely site-wide available? cause I would likely have other webs that will do Cites
?
I have this:
<a id="citation" href="#" class="jqSimpleModal {url:'%SCRIPTURLPATH{"view"}%/%WEB%/Cite3?skin=text;section=script;qtopic=$value(topic);qweb=$value(web);qbaseweb=%BASEWEB%;qbasetopic=%BASETOPIC%;qseq=$property(seq);qciteID=citeIDvalue'}">
[11:35]
pharveyyes, I think we just discussed it - DEFAULT_PROPERTYWEB ;-) [11:36]
packoman_oh. hehe. I am an idiot :D [11:36]
pharveyan idiot? highly doubtful [11:37]
packoman_well in german we'd say trottel [11:37]
pharveyYour usage of %WEB% is probably not doing what you hope it does, in that code you just pasted [11:37]
packoman_:) [11:37]
pharveybtw. How many property topics have you defined?
Just Cite, Cite3?
[11:38]
packoman_actually I only have one: Cite [11:38]
pharveypharvey should take this opportunity to change from $value(web) to %value{web}% notation [11:38]
packoman_Cite3 was just another version of it [11:38]
pharveyok [11:38]
packoman_but later I would likly want some other property topics [11:39]
pharveyOk.
pharvey begins breaking SemanticLinksPlugin some more
[11:39]
packoman_ok. that would be cool. I was a little distracted just now. so we would have a web containing values (in my case cite-topics) and one for properties (in my case the cite-topic). I understand correctly?
per-property-properties such as BIBLIOGRAPHYPLUGIN_FIELDS in the cite-case would live in the DEFAULT_PROPERTYWEB
sounds good
[11:44]
hm. one more thing:
you wrote "Your usage of %WEB% is probably not doing what you hope it does, in that code you just pasted"
what am I hoping it will do? :)
I just wanted a link to my Cite topic
to which I then pass all those options
is there a better/more robust way?
[11:50]
pharveypackoman_: I'm afraid Foswiki:System.SkinTemplates system isn't entirely intuitive... you see, templates are assembled very early in the rendering pipeline; and by the time %WEB% and %TOPIC% are expanded, the context in which they came from (your Cite3 topic) is lost, and are instead expanded in the context of the BASETOPIC (usually) [11:54]
FoswikiBothttp://foswiki.org/System.SkinTemplates [ SkinTemplates ] [11:54]
pharveyYou've hard-coded the name of your topic, you typed Cite3 instead of %TOPIC%
so you need to do the same for your %WEB%
[11:54]
packoman_you know something... I just realized that I need to change that as I have renamed the topic to Cite. And I was wondering, why it's behaving so weirdly
ok. I will do that
[11:55]
pharveyso I guess I'm saying that %WEB% will probably expand to the web in which the Cite3 is *used*, rather than the web in which Cite3 *lives* [11:55]
packoman_ok. now I got it
thanks for the info
[11:56]
pharveyyeah, that took me a lot of banging my head on desk when I was first learning templates :)
pharvey should probably write an FAQ about that
[11:56]
..... (idle for 20mn)
SvenDowideityes, BASEWEB and WEB are terrible in that way
SvenDowideit dissappears in a puff of smoke again :/
[12:17]
***dj_segfault has joined #foswiki [12:19]
SvenDowideitSvenDowideit pokes pharvey In der Ey [12:23]
pharveyow! [12:23]
SvenDowideit(suburb / street just around the corner from me)
SvenDowideit feels better for some reason :)
'some elefants are just plain nasty' simpsons...
[12:23]
packoman_pharvey: %TMPL:INCLUDE{Main.Cite}% is the correct syntax? [12:24]
pharveypackoman_: I'm just about to commit, you won't need that hack :P [12:25]
packoman_it's not working :(
oh. ok. cool. :)
[12:25]
pharveyI'm stuck on a boring detail [12:26]
SvenDowideitCDot: next Thur, can you do me a favour and sms / email me (before 7pm) wrt how trivial you think it'd be to get to your hotel around/after midnight [12:27]
pharveyI thought %WIKIPREFSTOPIC% = SitePreferences [12:27]
packoman_huh? where would you put this? [12:27]
SvenDowideitas i'm likely to arrive midnight, and then would have to find the hotel - after a day of looking after the girls, so a 6am start, and likey lots of running around :) [12:28]
pharveyah, it's %LOCALSITEPREFS
packoman_: it's nothing important, I'm trying to update the doc, and be a good Foswiki-citizen by not hard-coding the web names ;-)
[12:28]
packoman_ah. argh. I am breaking my whole, fragile cite-thingies
it still needs a lot work, it seems :(
[12:29]
SvenDowideitSvenDowideit goes back to trying to get awake enough to deal with more store2 conversions [12:30]
pharveySvenDowideit, w00t :) [12:32]
SvenDowideitmmm, pharvey ever seen/used smalltalk?
i was reading up a little for a dev UI for federated wiki, and started to decide that a smalltalk-ish dev Ui should really be pretty interesting for wiki app dev
[12:33]
pharveyseen, yes, used, no [12:34]
SvenDowideitgiven that it'd encourage smaller components
:)
[12:34]
pharveyI briefly used an HDL for Virtex-II FPGAs that was supposedly inspired by SmallTalk [12:34]
SvenDowideiti also thought about prolog and forth, but had to decide my ostory might be a bit odd [12:35]
pharveyaren't the cool kids using OCaml, Haskell, Erlang, things like that? [12:35]
SvenDowideitcertainly the fed.wiki UI encourages writing, and i think refactoring of prose too [12:35]
pharveypackoman_: Foswikirev:13021 [12:36]
FoswikiBothttp://trac.foswiki.org/changeset/13021 [ Changeset 13021 – Foswiki ] [12:36]
SvenDowideity, but this is less a language choice, more a dev UI think
as it will start as pluggable js components
[12:36]
packoman_pharvey: I'll check it out! [12:36]
SvenDowideitSvenDowideit pops that thread off the stack and tabs to foswiki-test-unit [12:36]
packoman_(been meaning to bring that one :P) [12:36]
pharveypackoman_: so you will need to replace [[SemanticLinksPlugin_DEFAULTWEB::foo]] with * Set SEMANTICLINKSPLUGIN_DEFAULT_VALUEWEB = foo [12:37]
packoman_ok.
SEMANTICLINKSPLUGIN_DEFAULT_PROPERTYWEB already in there too?
[12:38]
pharveyyep [12:40]
packoman_cool [12:40]
pharveyyou should see mention of it in the doc :P [12:40]
packoman_hehe. I'll read ;) [12:40]
pharveytho I admit... that doc is getting mighty tall
need to split it up one day
[12:40]
packoman_So I would put those in Main.SitePreferences? [12:44]
***GithubBot has joined #foswiki [12:46]
GithubBot[foswiki] foswiki pushed 1 new commit to master: http://git.io/wPpHgg
[foswiki/master] Item11243: Add DEFAULT_VALUEWEB & DEFAULT_PROPERTYWEB - PaulHarvey
[12:46]
***GithubBot has left [12:46]
FoswikiBothttp://foswiki.org/Tasks/Item11243 [ Item11243: Use System.PreferenceSettings instead of [[...DEFAULTWEB::... notation ] [12:46]
SvenDowideitoh yay - 'Apple Faces Temporary iPhone, iPad Ban In Germany'
SvenDowideit contemplated buying a samsung tablet because of apple
but still don't feel like spending the money
[12:47]
pharveypackoman_: You could do that, but I'd avoid relying on all your property topics wanting to assume the same default value web (i.e. I'd expect normally you'd set the DEFAULT_VALUEWEB on the property topic)
i.e. set VALUEWEB in Cite topic, and PROPERTYWEB in Main.SitePreferences
SvenDowideit: there are other droid manufacturers other than Samsung, too
:)
[12:48]
packoman_ok. makes sense.
I'd wait for the next asus transformer
actually I am waiting for it :)
[12:49]
pharveypharvey reset his Galaxy SII yesterday to factory dfeaults. Stupid bloody thing still can't decide if it wants to use Telstra NextG internet in HSDPA, HSDPA+ or UMTS [12:50]
packoman_how do you do these cool "off" comments
I want that too
[12:50]
***dj_segfault has quit IRC (Quit: Leaving.) [12:51]
SvenDowideitmmm
in au i use tpg - optus
but thats enought to make me think waiting is always good :)
my ipony 3gs does still work - just the off button gets ignored alot
[12:52]
pharveysadly, most of my family live in obscure places that barely have electricity, let alone optus coverage ;-) [12:53]
SvenDowideiti want the lenovo - cos it has a trackpoint
the thinkpad tablet plus kb seems more to my liking - though it does thus end up weight something similar to the x61
and i presume that unlike the asus xformer, needs to be used on a desk
[12:53]
pharveySvenDowideit btw - did you ever see leik's bug regarding a missing metatext property? as in... 'web.topic'/metatext [12:59]
SvenDowideityes, and no - in that i noticed it, but didn't see it [12:59]
pharveypharvey didn't even know we had such thing as a metatext, and didn't seem to find any code for it [13:00]
leikding :) [13:00]
SvenDowideiti suspect i'll wish we didn't
and i hope its just an old legacy of nastyness we cna ignore
[13:00]
pharveythere's doc, and a test, but grepping for metatext yields little else
pharvey never got to checking why the test passes
pharvey pokes that right now
[13:00]
leikWe used 'all' with dbcache..it's useful [13:01]
SvenDowideitwossitdo? [13:01]
leiksearches formfields (meta) and the text of the topic [13:01]
pharveyOH [13:01]
SvenDowideiter, so 'raw' version of the topic? [13:02]
pharveythere's actually no test at all
it's just a variable named $metatext in Fn_SEARCH tests
[13:02]
SvenDowideitand no code either [13:02]
pharveyso, we only have doc [13:02]
SvenDowideitso it something that was broken many years before in tmwikiland?
thatd be nice :)
[13:02]
pharveyleik: can you use... text =~ 'foo' OR fields.value =~ foo ? [13:03]
SvenDowideitOR fields.name =~ foo OR fields.attr =~ .... etc [13:03]
leikI haven't had time to dig into it, just realized it doesn't work like it did with DBCache [13:03]
pharveyleik: and update Foswiki::Support.Faq40 [13:04]
FoswikiBothttp://trunk.foswiki.org/System/PerlDoc?module=Foswiki::Support [ (Foswiki login) PerlDoc ] [13:04]
pharveyFoswiki:Support.Faq40 [13:04]
FoswikiBothttp://foswiki.org/Support.Faq40 [ Faq40 ] [13:04]
pharveypharvey understands :)
pharvey still trying painfully to get trunk into a deployable state for trin.org.au
[13:04]
SvenDowideit:(
my fault i presume?
[13:04]
pharveynah, lots of boring things
mostly jquery 1.6
[13:05]
SvenDowideitew [13:05]
CDotSvenDowideit: sure; remind me closer to the time. I'll be biking it. I already google-streetviewed it, and it looks easy enough (bike lanes most of the way)
CDot has gone for a shower, due to having had a very sweaty taiji class this a.m.
[13:05]
SvenDowideit:)
ta
[13:06]
leikusing jeditable and saving a formfield, it works but returns the whole page so it looks very strange, anyone know hos to make it behave? [13:06]
SvenDowideiti'm avoiding booking a hotel til the last minute
and am even tossing up just booking one for thru nite, and then booking the next 2 nites form cern
[13:06]
pharveyleik: any js errors? [13:07]
***Lynnwood has joined #foswiki
Lynnwood has quit IRC (Changing host)
Lynnwood has joined #foswiki
[13:09]
pharveySvenDowideit: Item11215 - did you get a change to look at Foswikirev:12938 [13:10]
FoswikiBothttp://foswiki.org/Tasks/Item11215 [ Item11215: Search pager broken when filtering on date ] http://trac.foswiki.org/changeset/12938 [ Changeset 12938 – Foswiki ] [13:10]
pharveychance* [13:10]
***dnavarro has quit IRC (Ping timeout: 260 seconds) [13:10]
pharveyMongoDB is now missing a filterByDate method [13:11]
***Colas has quit IRC (Quit: Ex-Chat)
AnneGilles has joined #foswiki
[13:11]
pharveypharvey created Item11244 :P [13:14]
FoswikiBothttp://foswiki.org/Tasks/Item11244 [ Item11244: [[Item11215]] introduced a filterByDate method, which MongoDBPlugin lacks ] [13:14]
***AnneGilles has left [13:14]
pharveyleik: FWIW, Release01x00 has no code for metatext either :/ [13:17]
CDotCDot smells slightly less bad now [13:19]
pharveyleik: would you be terribly offended if I simply removed metatext from the doc? Or to put another way: can you explain a use-case where an fields.value =~ 'foo' wouldn't be sufficient? Are you also searching in other types of meta, e.g. META:PREFS? [13:20]
CDotwhen a URL comes into Foswiki::Request, and contains %15 characters in the path, should those be expanded to their unicode equivalents for use internally? [13:20]
FoswikiBothttp://trunk.foswiki.org/System/PerlDoc?module=Foswiki::Request [ (Foswiki login) PerlDoc ] [13:20]
pharveyCDot: FWIW, URI::Query holds its URI in memory in its unescaped form [13:23]
***Colas has joined #foswiki [13:30]
CDotaye, which forces anyone manipulating that URI to know that, hence the reason for my q
CDot has gone to get his eyes tested
[13:31]
leiksorry was afk. I'll test the search you suggested and won't be offended in any way if we dont need 'metatext' [13:35]
pharveyCDot, ... I think requiring users of a $string to care about its encoding is asking for more punishment. We need to stop using byte strings... although I take the point that really, the escaped entities in an URI string aren't necessarily unicode entities, they really really really should be (unless somebody hand-built an escaped URI using something other than unicode)
but in that case you're screwed anyway; HTTP doesn't offer a way to indicate charset/encoding of a URI
[13:39]
CDotin the case of a Foswiki URI that includes a foswiki path, the % chars *are* unicode, by definition, since that's what FW speaks (now)
a similar constraint can be applied to any FW param
of course it's not true of the non-path and non-param sections of the URI; those should be left sacrosanct
CDot really has gone now
[13:40]
pharveyI seem to recall that that the protocol://user:pass part may contain %escaped bits, but I can't remember if that's supposed to be unicode entities, too. [13:42]
leikjsut a little stupid quiestion: in a SEARCH must I use URLPARAM to check if there is a param? I mean not like IF{"defined param" [13:43]
pharveyleik: defined is one of those few operations that are understood by IF and not by QuerySearch. Basically, I think in a SEARCH you can only use ops listed on the System.QuerySearch page
i.e. yes, you must use URLPARAM in a SEARCH
AFAIK
[13:44]
leikI think the search for ..OR fields.value =~ 'foo' works. You can remove it :)
it just gets a bit long..
remove 'metatext' I meant, but I guess you understood
[13:50]
pharveyleik: actually, CDot added this doc in 2009, according to git blame. CDot, if you say it's okay to delete mention of metatext, leave some feedback on Item11235 [13:53]
FoswikiBothttp://foswiki.org/Tasks/Item11235 [ Item11235: trunk SEARCH for metatext not working ] [13:53]
***pharvey has quit IRC (Quit: ChatZilla 0.9.87 [Iceweasel 7.0.1/20110930065748]) [13:55]
AndrewJones has joined #foswiki
gac410 has joined #foswiki
uebera|| has quit IRC (Ping timeout: 240 seconds)
[14:05]
.... (idle for 18mn)
MartinCleaver has joined #foswiki
MartinCleaver has quit IRC (Changing host)
MartinCleaver has joined #foswiki
MartinCleaver has quit IRC (Client Quit)
[14:28]
..... (idle for 24mn)
dnavarro has joined #foswiki
foswiki_irc1 has joined #foswiki
[14:54]
foswiki_irc1Wonderful, great, extraordinary, amazing, superb, fabulous.. ;)
Just checked http://foswiki.org/Community/FoswikiCamp2011 and most of participants prefer the unicode-int-the-core sprint topic… GREAT! :) Hopefully foswiki will understand unicode soon.. ;) ;) :)
[15:01]
SvenDowideitargh :)
it'll need people that test stuff
cos none of the people signed up use non euro text afaik
[15:07]
gac410Howdy SvenDowideit [15:08]
SvenDowideitheya gac410
i wish i'd never started this store re-arch
its depressing how much concentration it takes
and now, all the hidden groups tests fail, and so i have to work out wtf
[15:08]
gac410Hm. At least we have unit tests for this stuff. ;-) [15:09]
SvenDowideityes, and for that i thankyou :) [15:11]
foswiki_irc1Sven: non-euro text? I can't use forwiki because of bug filed as Item10635... ;( Unfortunately I haven't any idea how to write unit-tests for the utf8 stuff... ;( [15:13]
FoswikiBothttp://foswiki.org/Tasks/Item10635 [ Item10635: Wrong international character display ] [15:13]
***MartinCleaver has joined #foswiki
MartinCleaver has quit IRC (Changing host)
MartinCleaver has joined #foswiki
[15:14]
SvenDowideitfoswiki_irc1: excellent - i don't either :)
but i'm sure we'll be learning alot next w/e
[15:15]
foswiki_irc1:) [15:16]
SvenDowideitMartinCleaver: did you ever raise a feat req for logging in using email address? [15:19]
packoman_is there a way to rename WikiUsers?
ie. the wikiname
[15:20]
SvenDowideitthere might be - but it'll be untested [15:21]
gac410packoman_: yes but ... it leaves the history out-of-sync. also the entry in htpasswd
and a new user registring with the old wikiname will "acquire" the history (blame and credit) ;-)
[15:21]
packoman_ok. so I would have to get the name "right" a priory
hehe
ok. good to know
[15:22]
SvenDowideitmmm
gac410: huh?
[15:22]
gac410SvenDowideit can correct me if I'm wrong I'm sure. [15:22]
SvenDowideitif you rename from the cmdline?
or what
SvenDowideit wonders why renaming a topic would lose history
[15:22]
gac410Oh - rename the topic, and it will fix all WikiName references including that in WikiUsers topic ... but none in old revisions, right? [15:23]
SvenDowideitrename is rename? [15:23]
gac410It updates old revisions in the txt,v file too? [15:24]
SvenDowideityou will have to fix the entry in the passwd file if and only if you use WikiNames as login [15:24]
gac410right. [15:24]
SvenDowideitwhy would old revisions be changed?
never were before?
[15:24]
gac410They wouldn't be. So the back revisions would reflect the old WikiName - that no longer exists. [15:24]
SvenDowideitbut that doesn't mean '< gac410> and a new user registring with the old wikiname will "acquire" the history (blame and credit) ;-)'
unless :/ login == wikiname
gawd this stuff is fraught
[15:25]
gac410Oh - right. I'm thinking of a default topic-user-mapper install, where login == wikiname [15:26]
SvenDowideitwhen i worte it so many years ago, i assumed we'd ditch slow topic based usermappers, cos they have so many bugs
the db backed ones have cuid == an actually unique number
rather than deriving it from the login
[15:26]
gac410The history uses cUID which is encoded from the loginName, correct? So as long as the login name doesn't change, history remains intact / consistent. [15:27]
SvenDowideitzactly [15:27]
gac410But with default mapper. loginName and WikiName are same, and cUID changes if Wikiname is renamed [15:27]
packoman_ok. good to know. I am using default mapper [15:28]
SvenDowideitno
default mapper does not work that way
[15:29]
gac410no? [15:29]
SvenDowideitthe default mapper can be set to have wikiname == login _or_ wikiname != login
the default setting for the default mapping on the other hand....
[15:29]
gac410Okay - so default mapper with default settings. yeesh.
:-)
[15:30]
SvenDowideiti use the default mapper most of the time
but not set the way you suggested
so ...
and MartinCleaver suggested we even allow / set log == email
er login == email
which is imo even more fun (i have that on one of my old setups, but..
[15:30]
gac410This probably needs a FAQ, or supplemental doc on impact of using / not using loginname [15:31]
SvenDowideitit probly even exists
as its ancient
[15:32]
gac410Maybe a RenamingUsers document that discusses the side effects based upon mapper choices. [15:32]
SvenDowideitbut there are so many friggen docs, its impossible to find anyting
or stop doccoing, and start writing code that dwim :(
[15:32]
gac410Yeah - some of our Support docs really need scrubbing. Half-started efforts, stuff that is not maintained as foswiki changes. [15:33]
SvenDowideitfrom my pov, support doc == code broken
not that I / we have time to resolve the code to not need it
[15:33]
***MartinCleaver has quit IRC (Ping timeout: 245 seconds)
MartinCleaver has joined #foswiki
MartinCleaver has quit IRC (Changing host)
MartinCleaver has joined #foswiki
[15:34]
gac410Well we have indeed done lots of doc improvements. InstallationGuide has evolved with foswiki. But the supplemental install docs, like for windows, were very badly out of date in places [15:35]
SvenDowideitreally?
i noticed you did something to the activestate perl docs - but didn't read - i hope you write at top - don't use activestate perl :)
SvenDowideit wonders if mod_perl on win32 has been built with strawberry perl yet :(
[15:35]
gac410Yeah - Isaac Linn's rewritten install guide ... no didn't do that. But did fix up some of the instructions on running the rewriteShebang script for example. Things had totally changed. [15:36]
SvenDowideitaha - that kind of stuff should use INCLUDE :)
and the refactoring tool should point that out >:}
argh!
oh fffffffff.
SvenDowideit is going to hate this hidden gorup code
[15:37]
***MartinCleaver has quit IRC (Ping timeout: 240 seconds) [15:38]
SvenDowideitas it was written with the (imo) broken 1.1 store [15:39]
***MartinCleaver has joined #foswiki
MartinCleaver has quit IRC (Changing host)
MartinCleaver has joined #foswiki
[15:39]
SvenDowideitand which S**T decided to use BAaseUserMapper_{numbers} to make it hard to work out what user we're talking of
(me :( )
[15:40]
packoman_Is it possible to switch the mapper afterwards (i.e. now) to username != loginname? [15:44]
gac410Anything is possible. It's only a question of what topics, etc look like afterwards. [15:46]
SvenDowideityes, so long as every user is listed on WikiUsers, it should work ok [15:46]
packoman_hehe. well is it feasible... [15:46]
SvenDowideitand then when you rename the user topic [15:47]
packoman_so. could you point me to some information on how to do that? [15:47]
SvenDowideityou would need (i think) to edit the Main.WikiUser topic to map the new WikiName to the new WikiName
nope
[15:47]
gac410er.. If you switch to a mapper using loginname, don't you have to write the login name into the mapper topic? [15:47]
SvenDowideityou're well into territory where we throw up our hands and say either you are an expert, or you're willing to become one if things break
iirc, it will default to assume login==wikiname
as the passwd file contains the WikiName
[15:48]
packoman_ok. anyway. I think I am happy with how it is [15:48]
SvenDowideitit'll confuse the user nicely to log in as OldName, and then be displayed as NewName [15:48]
packoman_it's not worth the trouble [15:48]
***wdenk1 has joined #foswiki [15:49]
gac410Mapper has for example: JoeUser - JoeUser - 16 Oct 2011 First being wikiname, 2nd being the "login" name. [15:49]
packoman_just thought it might perhaps not be so complicated [15:49]
SvenDowideitwhich is another reason i played with login ==email
its complicated because topic based mapping uses... topics
[15:49]
gac410So if you switch using loginname - based upon unix accounts, or some enterprise login, the login name is probably not the same as WikiName [15:49]
SvenDowideitwich is nice, right up to the point where it really sux [15:49]
***wdenk has quit IRC (Ping timeout: 258 seconds) [15:50]
packoman_I do understand correctly that user mapping has only partially to do with how they are authenticated?
I am using apache as opposed to the templatlogin
[15:52]
gac410The mapper still is responsible for translating whatever string is used to authenticate into the Wikiname. [15:53]
packoman_ok. that's what I though
t
[15:54]
SvenDowideitermm
to be painful
[15:54]
gac410It's not the type of authentication - it's the mapping between whatever you login with to cUID to WikiName [15:54]
SvenDowideitthe (template|apahce)login etc is resposible for authentication (alone) [15:55]
***ThomasPHaeck has quit IRC (Quit: Leaving.) [15:55]
SvenDowideitthe mapper is responsible for authorisatoin & idenitfication
and worse
[15:55]
packoman_ok.
thanks for the info
[15:55]
SvenDowideitin apachelogin, that module is not responsible for authentication (as in checking that the login has the right pwd)
apache is
the mapper mapes the login (with the basic presumption that the login&pwd was tested (isn't that way authentication is)
there is an annoying exception to that simplistic seperation of course
in that the mapper API does deal with actually accessing the pwd :(
but you're better off not noticing that (unless you're like gac410 and esploding)
argh! hidden gorups will take me way too long to work out the code for
[15:55]
CDotSvenDowideit: what's you github address, so I can have a look at your new store api plz? [16:05]
SvenDowideitSvenDowideit !
CDot: as in https://github.com/SvenDowideit
i'm getting close to giving up having it done in time for next w/e
which will make alot more work for me merging whatever we do at the camp :(
i hope to work out the groups acl stuff tonite/tomorrow, but argh
[16:05]
***denisr has quit IRC (Quit: Parti)
uebera|| has joined #foswiki
[16:08]
CDotwtf happened to =begin TML? Or do you particularly hate PerlDoc?
'changeDefaultUser' sounds jolly nasty
[16:09]
SvenDowideitit is, but we've had that forever
i'm just making it explicit
forever ~~ since before my user mapper code
[16:10]
CDotdoc for sub save == sub __internal_save. So what should I read, code or doc? [16:11]
SvenDowideitas loading groups used to rejig the logged in user to == admin, to bypass auth
but it did it by directly accessing the session
[16:11]
CDotay, ay [16:11]
SvenDowideitthen in 1.0, we used a local
and then in 1.1, that code stopped being necessary :(
which means.... that all the code written in 1.1 to deal with hidden gorups and metacache has to be re-visited - /me has a small heart failure
[16:11]
CDotlooks sensible so far, much as I'd envisaged. [16:13]
SvenDowideiti hope so :)
shame making it works sux as much as we'd envisaged too
[16:13]
CDotCDot would quite like to get rid of this distinction between store object types - as in, a store object is a web, a topic or an attachment, and each should have essentially the same methods over it [16:14]
SvenDowideitat which point?
cos i'm trying to do the opposite
[16:14]
CDottho of course a web is only a collection, a topic is a collection + some other shit, an attachment is only other shit. [16:14]
SvenDowideitin that imo F::Meta needs to become several seperate classes [16:14]
CDotmind you, a web has always needed/wanted other shit too (metadata) [16:15]
SvenDowideitone for each store type [16:15]
CDotCDot goes to read ::Meta before commenting further [16:15]
SvenDowideitnothing much has been done to seperate them for now
as just getting the new API wired up and functional has been more work than i seem to have energy for
[16:16]
CDothmmmm, not finding a subdir "Meta" [16:16]
SvenDowideitsee above [16:16]
CDotaha, ok
you are in the same place I ended up in at least 3 times
[16:17]
SvenDowideitthe bug you made in 1.1 store takes up alot of time
nothing else has caused me much trouble so far
[16:17]
CDotI'm glad my work has helped you focus [16:17]
SvenDowideiti'm not glad atm :)
i really really wish i had a good debugger for perl :(
[16:17]
CDoty [16:18]
SvenDowideitkomodo is actually crap [16:18]
CDotwish I had some way to track down and kill wide characters [16:18]
SvenDowideitmmm, maybe i should see is the new one is better
i have komodo5 - it does not do watches or local vars other than the top of stack
so to ask the depressing q
if you were at this point 3 times before
what stymied you?
[16:18]
CDot(16:16:30) SvenDowideit: as just getting the new API wired up and functional has been more work than i seem to have energy for [16:19]
SvenDowideitmmm, the thing you did in 1.1 was the same amount of work as i'm doing [16:20]
CDotat the time, i was fighting with many tunnelers
most of which have been killed off now
[16:20]
SvenDowideitand the 1.0 api was actually more compatible with what i'm doing
ok, thats 80% a lie, but this store returns topics the current user does not have perm for really makes for alot of work
and clearly, along the way to 1.1, you imrpoved the non-api bit
but the api is argh
[16:20]
CDoty, I know you & others don't like it; but it worked for me at the time, and allowed me to move stuff to what I felt were better places [16:23]
SvenDowideitf. i wonder if i can get a trail version of komodo6 :( [16:23]
CDotso, I'm happy to see it change, if it helps advance the cause of science
CDot is *much* happier with the store since fixing the "floating change" problem
[16:23]
SvenDowideitSvenDowideit always ignored it as a corner case [16:24]
CDotI feel much more confident that the store code isn't going to bite my ankles every time i change something [16:24]
SvenDowideitand tbh, i avoid changing topics without using the API like the plague [16:24]
CDotand now the same needs to be made true of Meta
cos I acknowledge that Meta has ankle-biting capability
[16:24]
SvenDowideitoh, i think i
'm stuffed
shit.
[16:25]
CDottrussed, spiced, herbed and stuffed? or just stuffed? [16:26]
SvenDowideitSvenDowideit just pondered - how does getPrefs work when you want to read a preference as a different user
someone simplified that code to hell too
[16:26]
CDotahhhhhh..... that old chestnut [16:26]
SvenDowideitit used to work [16:27]
CDotCDot has been afraid of prefs since the bitmask stuff cam along [16:27]
SvenDowideitbut gilmar told me someone else removed the code that dealt with different user stuff
and threw up his hands and gave up for the same reasons i wimpered
[16:27]
CDot"someone else" usually means me [16:27]
SvenDowideiti can't recal [16:28]
CDotI don;t recall it, TBH
but storing a prefs tree for a number of different users => a big nest of squirmy things
as prefs reading is about the least efficient thing we do
[16:28]
SvenDowideitcept its necessary
SvenDowideit waves a profiler at it
and totally dissagrees
[16:29]
CDoty, sure. IIRC I tried to define a "per user" prefs structure that allowed for different prefs to be loaded on a per-user basis
CDot wonders what happened to that
[16:29]
SvenDowideitmmm, ok, good to hear
that might be there
and might be more stuff gilmar said 'not my code, didn't make sense to me, so i stopped looking'
SvenDowideit will feel much happier if he could afford to sit down and work on this for a few months without interruption
but my life doesn't allow that without someone paying for pam to be on holidays :(
[16:30]
CDotthe biter is when preferences in the tree are *protected* from read by certain users [16:31]
SvenDowideiti'm sure i'm not alone in that :(
zactly - thats the issue with the hidden groups thing
or better said part of the issue
the other part is that it was coded in 1.1 to take advantage of the looser store
[16:31]
CDoty. That's why I had to init a prefs tree each time from a full session obj [16:32]
SvenDowideit(from what i can grok so far) [16:32]
CDotcos nothing else had the right info to init it
code deep in the prefs handling gropes right into the session obj for user info
[16:32]
SvenDowideitSvenDowideit sees if thats still the case [16:33]
CDoty, if you look at the way the prefs "singleton" is inited, it does it subject to the session being sufficiently configured to support user access control checks [16:34]
SvenDowideiti can't see it [16:34]
CDotFoswiki.pm1726 and down [16:35]
SvenDowideitits more that in my store2 branch [16:35]
CDotsetUserPreferences [16:35]
SvenDowideiti can't see the Prefs code actually using the session much [16:36]
CDotok, that's good, I think
IIRC the only requirements for it was as described; access controls
[16:36]
SvenDowideitdunno - cos itmight just have more todo with the now store singleton and the hidden uid there
and the access obj ..
[16:36]
CDotposs [16:37]
SvenDowideitSvenDowideit got rid of the meta->{session} in about 10 mins work [16:37]
CDotgood man! [16:37]
SvenDowideitand that made debugging so much easier
SvenDowideit gets pissy about Data::Dumper being useless due to local caches of objects that are trivially accessible via globals
especially as seeing the global used can then raise a flag to others to remove them
[16:37]
CDoty. I originally did that local ref as part of the process of eliminating the hundreds of globals that were floating around, unchecked
the only good reason for *not* killing that is multiple co-existing peer sessions
[16:39]
SvenDowideiter no
thats the main reason for killing it
you can't cache and reuse an object across session if it contains a session
same as ou can't store a prefs tree in meta
havn't killed that yet
[16:40]
CDotwell, not entirely true, but I'm not going to argue. So long as you can maintain multiple concurrent sessions - partition the data space on a per-session basis e.g. with locals - then i see no problem [16:42]
SvenDowideita meta obj is a representation of the data on disk
adding session related cruft directly to it, mixes concerns and leads to woe
i want to move away from 12 years of woe
[16:42]
CDotagreed [16:43]
SvenDowideitone sub goal is to be able to use store&meta without session, but all i can do atm, is to try to make the API capable of it [16:43]
***Colas has quit IRC (Ping timeout: 258 seconds) [16:44]
SvenDowideiti keep forgetting to hassle ColasHome
he should be coming next w/e to see particles dissappear
or be created - not sure which
[16:45]
CDoty, th api is a great starting point
I have a horrible feeling that the unicode mission is doomed, so store2 is looking good for the sprint
[16:48]
SvenDowideit:(
i'd rather that store2 mostly works before the camp
and that unicode is it - at least lotsa tests
but i am prolly dreaming too
[16:51]
CDotlotsa tests = ideally, 5X the number we have now by my pessimistic estimate [16:53]
SvenDowideitnope. seems komodo7 beta1 is just as pointless as 5.
so to debug this, i can either learn some painful gdb like perl -d, spend weeks adding debugging to padre, or scream and give up for the day
[16:54]
CDotwhat are you trying to debug? prefs? [16:56]
SvenDowideitna, i'm not even that far
all i want is my debugger to give me the correct values for local vars
at every point in the stack...
so i can get back to properly knowing what the code does
wow
[16:56]
CDotin perl? pfffft [16:57]
SvenDowideitand between komodo5 and 7 they decided to change the function keys for debugging
pelr can doit
[16:58]
***foswiki_irc6 has joined #foswiki
foswiki_irc6 is now known as webfork2
[16:58]
webfork2Babar: dunno if you're online still, but hopefully you read the chat logs...
sorry i dropped out yesterday, I have a spotty connection
[16:59]
SvenDowideitthat might be enough reason to change - 7's are the same as my now finger memory - msdev :(
geee, thanks - it tells me '%args' isa HASH
like, duh - i want to know whats in the hash you pos
[16:59]
***AnneGilles has joined #foswiki [17:05]
SvenDowideitah, yay, a solution present itself - more work on the pile, but progress [17:18]
***webfork2 has quit IRC () [17:24]
leikwho knows jeditable? It's cool to make a "field" an input and submit but it returns the whole page so it looks very wrong until you reload the page
of course I'm doin' it wrong
[17:27]
CDotleik: I made extensive use of it in EditRowPlugin. See EditRowPlugin/pub/System/EditRowPlugin/erp_src.js
make if behave well is not easy
[17:28]
***MartinRowe has quit IRC (Quit: Leaving.) [17:29]
.... (idle for 15mn)
leikok, I give up :) [17:44]
***MartinCleaver has quit IRC (Ping timeout: 245 seconds) [17:49]
CDotleik: quitter!
anyway, it sounds like you are trying to do a whole-topic save. What you probably want is to use it with a server-side REST service; not sure what state the RestPlugin is in, but that should provide the service you need.
[17:53]
leikfor sure! glazing at the appelsiini doc it's not quite clear [17:55]
***dnavarro has quit IRC (Quit: Leaving) [17:55]
CDot(the EditRowPlugin defines an API for REST calls to manipulate tables in topics, so it's a more specific problem) [17:55]
leikbut the editable POST's and sends the data for a formfield, it gets saved, its only that the return data is the whole topic. make it go > /dev/null [17:56]
CDotif you want to stay in JS land, then just ignore the response from the server (unless it's an error) and manipulate the DOM field from JS (see erp_src.js for examples of doing that) [17:56]
leikyes. how?
it think it defaults to this behaviour and I need to override....guessing
[17:57]
CDotsee var editCallback = function(value, settings, val2data) {
in erp_src.js
that callback is looking at a specially compiled server response, but the same principle applies for what you are doing
[17:58]
leikthanks, must try with an editor instead of 'less' :) didn't I see Komodo 7 mentioned :)
next up is JsTreecontrib :) examples of using icons from System/DocumentGraphics
[18:05]
CDotyou're on your own with that one. But one last point on jeditable - I added a JEditableContrib you can use.
doesn;t do much, just adds it to the JQueryPlugin model
[18:12]
leikthanks, am stutying, sometimes take a while u know, and maybe more than one try :)
studying
[18:13]
....... (idle for 32mn)
***Lavr has quit IRC (*.net *.split)
Lavr has joined #foswiki
card.freenode.net sets mode: +o Lavr
[18:45]
.... (idle for 16mn)
leik has quit IRC (Quit: Leaving) [19:02]
CDot has quit IRC (Quit: Leaving.) [19:12]
....... (idle for 33mn)
ArthurClemens has joined #foswiki [19:45]
foswiki_irc1_ has joined #foswiki [19:55]
foswiki_irc1_hey! I'm really new to wiki so this might be a stupid question...- anyone in here who knows how to set up one table and link it to several pages? meaning you only have to update the table in one place and it gets updated in all pages [19:59]
gac410You can put the table in it's own topic, and the %INCLUDE it on the other pages. [20:00]
foswiki_irc1_ah! thnx - will try it :) [20:00]
gac410Look at Foswiki:System/VarINCLUDE for details [20:01]
FoswikiBothttp://foswiki.org/System/VarINCLUDE [ VarINCLUDE ] [20:01]
foswiki_irc1_TY - didn't find that one [20:02]
foswiki_irc1or you can wrap the table in the CurrentTopic with STARTINCLUDE/STOPINCLUDE and in the another topic use INCLUDE("CurrectTopic")...) check: VarSTARTINCLUDE [20:07]
foswiki_irc1_thanks a lot - works great! [20:18]
***Colas has joined #foswiki [20:20]
ArthurClemensstrange, pasteword is default enabled by TinyMCEPlugin, but not the button to use it [20:34]
BabarArthurClemens: I think it's disabled per browser
pharvey would know better
[20:35]
***leik has joined #foswiki
packoman_ has left
[20:37]
ArthurClemensI need to write * Set TINYMCEPLUGIN_ADDITIONAL_BUTTONS3 = pasteword to see the button
that does not look like disabled
[20:39]
BabarBabar seems to remember either pharvey or MTempest making tests and saying it was broken on some browsers. But... maybe he/they decided not to make it visible by default... dunno, sorry
crap... pootle...
[20:41]
***foswiki_irc1_ has quit IRC () [20:49]
Babarhum... weird... pootle seems to imply nothing changed
which seems illy
*silly
great... I can't find back the backup I made on my disk...
[20:56]
.... (idle for 18mn)
***TJCRI has joined #foswiki [21:15]
Babargac410: I finally managed to update pootle to see the new strings.
now the translators can work for 1.1.4
sending an email...
[21:25]
...... (idle for 26mn)
***pharvey has joined #foswiki
pharvey has quit IRC (Changing host)
pharvey has joined #foswiki
AnneGilles_ has joined #foswiki
AnneGilles has quit IRC (Read error: Operation timed out)
AnneGilles_ is now known as AnneGilles
[21:52]
SvenDowideitgmc: boooooooooooo! :(
that is all
[22:00]
gac410Babar, excellent! Thanks. I guess that really makes us into string freeze. [22:01]
pharveySvenDowideit: I've been in those crazy hidden group tests before, if you want a second pair of eyes [22:01]
SvenDowideitpharvey: i fear they're just no enough
i just pushed some changes, and now they pass :(
and i really didn't expect them to
cos if i read the code i commited right, as soon as groups are expanded, the store thinks the session isadmin
[22:01]
pharveyArthurClemens: TMCE's paste plugin tries to do the right thing automagically (i.e. detect word content) [22:03]
SvenDowideiti'm moving on to the next set of tests that crash first tho
which ironically is the problem with the AdminOnlyAccessTests - we're testing isAdmin before we search for groups, so it calles exists recursivly
[22:03]
gac410yeah - isAdmin for the super admin is an override for everything - nothing else is checked if that is true. [22:07]
SvenDowideitthis is lower down in the new code
when the hidden groups code was written
it was written with the 1.1 assumption that it could acces topics that the session user did not in fact have access to in 1.0 (or 2.0)
[22:08]
gac410We have to be careful though that the isAdmin for the super admin needs to work even if mappers and all are defunct. [22:09]
SvenDowideitand so is excessivly simplistic [22:09]
jayen@pharvey i want, when users visit the top page, users to be presented with a login page that looks like it's part of the wiki. the same thing template login gives is preferred. [22:10]
SvenDowideitfrom what i have seen isAdmin has never worked when mappers etc are defunct [22:10]
ArthurClemenspharvey: you are right [22:10]
SvenDowideitbut yes, one goal i have is for the store to consider cuid==undef == admin
but right now, that broke - once i get closer to having the existing unit tests running (as in, not crashing, and some portion actually passing too)
[22:10]
gac410Thats a huge issue which you and I disagreed with at one point. the access checks need to be fully functional even if the current user has no access to the mappings. Any "utility" Func tools like group membership, email addresses, etc. needs to cut through. [22:11]
SvenDowideiti'll start adding more medd
the access checks need to be fully functional even if the current user has no access to the mappings
i agree
[22:11]
gac410Otherwise notifications, etc. won't work [22:11]
SvenDowideitbut its not as simple as all that [22:11]
gac410If things like notification plugins and contribs can't read the email addresses due to authentication issues it all falls apart. [22:12]
SvenDowideitits getting easier in trunk, because the acl code is not in meta [22:12]
ArthurClemensArthurClemens starts to give attention to 150 words [22:12]
SvenDowideitthe issue is that <2.0 caches data that the current user should not get, but isn't good at keeping track that they should not get it
and that <1.0 tried not to pass data across the store boundry
[22:13]
gac410If the "intent" of the access is "operational" (email notify, etc) then access checks are unnecessary. If "intent" is "informing" - display group members, etc. then access checks are critical. [22:13]
SvenDowideitFunc however isn't allowed to cut through [22:14]
gac410It has to be or you wipe out email contrib [22:14]
SvenDowideitas the 10 year old 'theory'
was that users of Func should only get what they have permission to get
email contrib used to run as a privelidged user
but that too is shite.
oi, wtf?
' (email notify, etc) then access checks are unnecessary.'
[22:14]
ArthurClemenspootle is showing an overkill of icons [22:15]
gac410Not always - And ImmediateNotify, the notification runs as a afterSave [22:15]
SvenDowideitare you trying to suggest that the fact that atm, paul's secret topic info is notified to people that should not see it is good? [22:15]
gac410NO.... not that. [22:16]
SvenDowideitIMO, access checks are always necessary [22:16]
gac410I'm saying that if paul's existence is secret. But he is subscribed to get notifications, when I save, he should still get notified. [22:16]
SvenDowideitwhy would that not be the case? [22:17]
gac410I can't see him, but the code running under my control still must notify hem. [22:17]
SvenDowideitno, i don't agree [22:17]
BabarArthurClemens: overkil of icons? Ah, google translate and the like? [22:17]
SvenDowideitits convenient to do that in aftersave, but a really bad idea [22:17]
gac410If you make Func so it can only access based on my permissions, how can the core still use Func to get his email address to notify him. [22:17]
SvenDowideit'expedient, given our shit arch'
no
[22:17]
gac410Not all sites have ability to run cron. [22:17]
SvenDowideitsecurity says, don't write code that is inheriently risky [22:18]
gac410for better or worse, in some cases, the only available option is an exit. [22:18]
SvenDowideitthere are generally better options
for your strawman, i would shell out to a new req
[22:18]
gac410So you wout rather tell sites without cron access thaty they should just suck it up and find another hosting. [22:19]
SvenDowideitno, i would rather find a safe way to secure the code
rather than making a mess
(i am conveniently ignoring some pre-existing mess)
[22:19]
gac410*Some* API needs access to protected information. You can't stop it 100% -- otherwise some things can't work. [22:19]
SvenDowideityes, some, but not the API to extensions [22:20]
gac410Not a chance. Then extensions will just bypass the api. [22:20]
SvenDowideityes, thats a positive [22:20]
gac410not a chance. [22:20]
SvenDowideitthat way we continue to say - thats not a supported case [22:20]
gac410that's utterly totally bogus [22:21]
SvenDowideitno difference to howit is in any program
i give you a c API that is safe, if you want to go unsafe, you can poke my memory
its perfectly normal to give a safe API, but accept that real hackers know what they are doing
i've done it myself
[22:21]
gac410thats crap. There needs to be an API that has access to privileged access. [22:21]
pharveyArthurClemens: If you have time can you review http://www.tinymce.com/wiki.php/Plugin:paste - maybe I've misunderstood the description (I read it as: paste-from-word button being unnecessary, but then the example they give has an pasteword button) [22:21]
SvenDowideiti have not heard any one eg that can't be solved in a simpler fashion without polluting the poorly coded internals
my thoughts are coloured by the reality that we have lots of shittly little caches in this code that assume that it can leack anything it has to the current session
metacache is one that probably almost works ok now
[22:22]
pharveyArthurClemens: btw any feedback (even if boring, like "it didn't crash on Safari") of trunk TMCE & WysiwygPlugin would be cool :-) (uploaded to Extensions/Testing) [22:23]
SvenDowideitbut its an obvious one that we made more obvious
unless you _know_ that the core is clean of those leaks, a convenient ACL busting API is just asking for your persistent perl session to leak to guest and give them your passwords
(it scares me how many people have host, login, pwd lists in their wiki's)
[22:23]
gac410I guess I'll wait to see the results because IMO breaking the func Group membership and Email APIs will totally break EmailContrib and ImmedateNotification type plugins. [22:25]
pharveyseems a few other projects came to the same conclusion as me, http://dev.contao.org/issues/1199 [22:26]
SvenDowideitfunc Group membership and Email APIs didn't used to break out of the ACLs
but then, that was before hidden gorups were evaluated too :)
[22:26]
gac410I was going to say - they didn't break them - because there was no checking in the first place. [22:27]
SvenDowideitit was checking ACLs [22:27]
BabarArthurClemens: so I was almost right. Not that it was buggy, but useless. I rememberred both, but... hey, chose the wrong one :( [22:27]
SvenDowideitbut what a group was changed when that was implemented - what is a pain, is that that work was built on the ACL-slack 1.1 [22:27]
gac410Email contrib depends on Func returning ALL email addresses that need notification. Topic access for each user is checked. But the user membership of a group is not. [22:27]
SvenDowideitmaking it work in the 1.0 / 2.0 ACL tighter version, requires the same as the groups SEARCH thing [22:28]
ArthurClemensdone translating. next time we need to review the english first [22:28]
SvenDowideitEmail Coontrib - you mean MailerContrib? or the other?
either way - if there are unit tests, it'll help me :)
[22:28]
Babarlet's see if I can sync that back to SVN :) [22:29]
gac410Thanks ArthurClemens [22:29]
SvenDowideitseriously, i was expecting the admin group membership thing to bite me much harder than it just did
but i've much more togo
SvenDowideit is still confused where email isa problem, those are not stored in topics
[22:29]
ArthurClemenspharvey: I am now editing Sandbox.TestTopic44740 on trunk. what should I look for? [22:30]
gac410I tell contrib notify "SecretGroup" Membership is secret. But I need the membership to get their email addresses to do the notify. [22:31]
pharveyArthurClemens: breakage :) I usually test image/attachment handling, and play with tables, headings, bullets [22:31]
gac410For each member, then determine if that member is permitted to view the topic and therefor be notifited. [22:31]
BabarBabar has lost the cronjob for pootle... looking...
found it
[22:31]
pharveyArthurClemens: I just thought, if you were already playing with trunk TMCE, you could let me know how it goes. You don't have to do my basic testing for me :) [22:32]
gac410So mailer needs to resolve SecretGroup (and all sub-groups and sub-group members) to a list of email addresses permitted to view TopicX [22:32]
SvenDowideitnow, consider, why exactly should _you_ be given the email addressresses of a secret group you are not even supposed to know exists?
i can fully justify a mailer req getting that info
[22:32]
gac410I DO NOT NEED the address.. The internal functtion needs it to do the notification. [22:33]
SvenDowideitbut thats a seperate and secure request [22:33]
gac410The user never sees it. [22:33]
SvenDowideitexactly - so there's no need for it to pass across the boundry to your session
and thus, there's no reason for it to be 'in process'
[22:33]
gac410Hello, wall. NO NO NO NO.... The ImmediateNotify runs under control of the current user . [22:33]
SvenDowideity, and that is NOT A GOOD THING
(tm)
[22:34]
gac410We have no other place to do it. [22:34]
SvenDowideitwe do, we have many
incuding calling system(./rest imediatenotify)
[22:34]
gac410None that are public API that get control during save that dont need a cron job. [22:34]
SvenDowideitfor eg
i just gave you one eg that doesn't need a cronjob
i'm sure that there are other imaginative ways to make it inheirently safer
[22:34]
gac410So user updates topic and then changes url to rest Immediate notify? huh [22:35]
SvenDowideitno
the _plugin_ shells out to its resthandler
thus avoiding polluting the current session with info it shoudl not have
its one simple (and bruteish) way to ensure that you're not writing code that requires mixing of session permissions
take off the expedient hat, put on the security guy hat
and yes, i expect there are better answers to the issue, but i'm going for as simple to code an answer as possible atm
[22:35]
gac410I don't get how that is more secure, If plugin can shell out, it should still pass through the current user auth otherwise I could shell out and do evil stuff. [22:37]
pharveyArthurClemens: I posted a question at http://www.tinymce.com/forum/viewtopic.php?pid=96186#p96186 [22:38]
Babarok, Arthur, your translations have been committed [22:38]
SvenDowideitits more secure formthe pov that the currently running exe does not have much more info than it should [22:38]
ArthurClemenspharvey: generally it works well. I found 2 bugs so farmultiline verbatim is not picked up [22:38]
SvenDowideitits never going to happen, but the point is there [22:38]
ArthurClemensmultiline verbatim is not picked up [22:38]
Babarwe'll see in 1.5 if my cronjob is working :) [22:38]
ArthurClemensand blockquote gets an empty p before the text [22:38]
SvenDowideiti hope that the Tasks API thing from timothe is clevererer than that to [22:39]
ArthurClemensalso, blockquote doesn't get parsed when re-entering edit mode [22:39]
gac410SvenDowideit: You make it secure, and the rest of us will figure out ways to actually make it work. [22:39]
SvenDowideitcos emailing from a cgi req is really naf
especially when the email program hangs, dispite it not supposed to be able to
[22:39]
gac410SvenDowideit: 100% in agreement. But if the hosting site doesn't permit background, then *today* we have no other option. [22:40]
SvenDowideitat this point, i'm working on making it work [22:40]
pharveyArthurClemens: bugger, you'd think I'd have some failing tests for that [22:40]
SvenDowideitbut if you can work on that instead, i cna go home :)
today, we also don't have a safe store
and we also don't have a safe meta
and hell, we don't have alot of things
[22:40]
gac410for now my focus is 1.1.4 - we need more releases to get our freshmeat aka freecode vitality up. [22:41]
SvenDowideityup, and i'm glad of that - extremely glad
cos 2.0 is depressing me enough, without being told what i'm trying to do won't work
cos if you're right, i've just wasted 4 months
[22:41]
gac410And I think I fixed your WikiName@blah.com issue with findElsewhere [22:42]
SvenDowideitoooooo, brilliant!!!
SvenDowideit wishes he owned blah.com :)
[22:42]
pharveySvenDowideit: it's not a waste. I really really want store2 :P [22:42]
gac410No - It will work - but may need an "auth" API FuncAuth or something. [22:42]
pharveyand I'm willing to help get it working
actually I'll be forced to get it working, given that I'm perpetually using trunk as a production thing
[22:43]
ArthurClemensblah.com is registered but not in use
Creation Date: 20-mar-1995
[22:44]
SvenDowideitArthurClemens: depressing huh - what a waste [22:44]
gac410SvenDowideit: As long as the unit tests work, things will be good - unless you fix the tests. ;-) [22:44]
ArthurClemensit's not even offered for sale [22:44]
SvenDowideitworse, i was considering becoming a domain squatter in 93-94, but pam convinced me it would be evil
gac410: boooo - the unit tests are not enough :)
[22:44]
pharveyArthurClemens: if you have some time can you add bugs to Item10724 [22:45]
FoswikiBothttp://foswiki.org/Tasks/Item10724 [ Item10724: Update to TinyMCE 3.4.5 ] [22:45]
SvenDowideiti wanted home.com at the time, oh well - its still squatted [22:45]
gac410Enough - no - but better than nothing. MUCH [22:45]
***TJCRI has quit IRC (Remote host closed the connection) [22:45]
pharveyBabar: FoswikiBot caches page titles? Nice :) [22:45]
Babarpharvey: what do you mean? it remembers when it already said a link, and doesn't say it again within some time, 30 min iirc [22:46]
SvenDowideitmuch much, yup, just don't talk to kent beck about them - there are no unit tests in there :) [22:46]
BabarSven told me to implement such a feature... [22:46]
pharveyBabar: it's just that the current HTML <title> for Item10724 is out of date (I changed it ~10 hours ago) [22:47]
SvenDowideitBabar: ignore that bossy fellow, he's not nice [22:47]
***GithubBot has joined #foswiki [22:47]
GithubBot[foswiki] foswiki pushed 1 new commit to Release01x01: http://git.io/DxoZJw
[foswiki/Release01x01] Item9526: Uncommitted translations in the last 4 hours - WikiTranslationGnome
[22:47]
***GithubBot has left [22:47]
FoswikiBothttp://foswiki.org/Tasks/Item9526 [ Item9526: Translation work (needed for Pootle auto-checkin - please don´t close) ] [22:47]
Babarpharvey: no you did not
click the link, you'll see the bot is right
ah no
[22:47]
gac41010742 says Update to 3.4.7 bot says 3.4.5 [22:48]
Babarah no...
wait
[22:48]
pharvey:) [22:48]
Babartitle_expire => 86400
Babar thinks we can even ask the bot what value it has
but... don't remember how :)
so, yes, it caches titles for a day
[22:49]
SvenDowideitSvenDowideit decides he should be in bed [22:49]
pharveythat's nifty [22:49]
SvenDowideitBabar: what are we doing first thing next friday? [22:50]
Babarurl_delay => 1800, title_delay => 1800, [22:50]
SvenDowideitbreakfast @ cern? [22:50]
BabarSvenDowideit: can be. To be honest... I have no clue :)
Babar still needs to send the list of registered people to the relevant person
but, as I'm on holiday, and Pete Jones hasn't replied...
[22:50]
SvenDowideitdoh
so we might be having fw ga in the hotel where the ex-el-president resides
[22:51]
Babarnah... I can always make you guys come in
I just wanted it to be more friendly... especially as I'm on-call that week
[22:51]
SvenDowideiteeegads
so long as nothing goes wrong :)
[22:51]
jayenpharvey: i want, when users visit the top page, users to be presented with a login page that looks like it's part of the wiki. the same thing template login gives is preferred. [22:56]
***Demosthenes has quit IRC (Ping timeout: 258 seconds) [22:56]
pharveyjayen: hi, let me see if there's an easy-to-use template def for that [22:57]
jayeni tried to use %TMPL in the wikitext, but that didn't work [22:57]
SvenDowideitwow. thats sooooo ouchie
SvenDowideit goes to bed to sleep on the latest isAdmin, isn't admin issue
[22:57]
***Demosthenes has joined #foswiki [22:58]
jayengnite [22:58]
Babaroh, Pete registered on the FoswikiCamp page!
crap.. they're asking for nationalities of participants too...
[23:00]
***terceiro has quit IRC (Ping timeout: 240 seconds) [23:02]
Babarlet's hope the AUTHORS is up-to-date :)
SvenDowideit: you're Australian or German?
any idea what nationality Padraig Lennon is?
ok, sending emails...
[23:03]
ArthurClemenswhat would them make feel more secure? [23:08]
BabarBabar thinks he's Irish [23:11]
ok, mail sent. [23:16]
pharveyjayen: http://foswiki.org/Sandbox/TestTopic44752 [23:18]
jayenwow that looks so broken [23:18]
pharveyoh, it does
sorry, I didn't notice the junk at the bottom on my monitor
got to run to a meeting
[23:19]
jayenok
well i see %TMPL workes for you
i'll see if i can get it to work for me
thanks
[23:19]
***ArthurClemens has quit IRC (Quit: Leaving...) [23:23]
jayenhrm, %TMPL:P seems to work fine on f.o, but not on my installation [23:24]
Babarseems my guest room just got free :) [23:30]
pharveyjayen: TMPL:P can only print TMPL:DEFs that have been TMPL:INCLUDEd somehow
so it's likely you're trying to TMPL:P a def that isn't defined
[23:31]
jayenso how are they included on f.o?
if i remove your <!-- --> comment and hit preview, it looks ok
but only in the preview. hitting save and it disappears
[23:31]
pharveyoops, I trashed our topic :)
I know the problem.
jayen: the trick is the * Set VIEW_TEMPLATE = Sandbox.TestTopic44752
that causes Foswiki to act on the TMPL:INCLUDE directives
without it, those TMPL:INCLUDEs are ignored
because by default, the BASETOPIC is not part of the Foswiki:System.SkinTemplates search path
[23:33]
FoswikiBothttp://foswiki.org/System.SkinTemplates [ SkinTemplates ] [23:37]
jayeni copied your wikitext, r4, and set the view template to main.webhome and the whole page is very broken
no top bar, botton bar, edit button
just a series of login boxes
https://test.foswiki.science.unsw.edu.au/Main/WebHome
er, i restored it
and it keeps quiet saving, so r3 isn't what was broken, either. ergh
ack, your r4 keeps changing. sometimes i really hate quiet save
[23:39]
pharveyoops, sorry. still hacking
I think it works now
but it's quite convoluted
I think this approach is too fragile. Probably better to reproduce the HTML separately.
[23:45]
jayenhttps://test.foswiki.science.unsw.edu.au/Main/WebHome
i'll wait until you see it before i revert it
[23:48]
Babarnice :) [23:50]
pharveyjayen: yeah, this approach is flawed. The login templates are depending on %VARS that don't exist in a view.
jayen: let me try again.
[23:51]
jayeni'm happy to use an iframe, if that works [23:52]
pharveyI guess that's another option [23:54]
jayenyeah, when i tried that, it broke the page as well
well, in the preview. let me try saving this time
https://test.foswiki.science.unsw.edu.au/Main/WebHome - lol
https://test.foswiki.science.unsw.edu.au/Main/WebHome?rev=8
[23:54]
gac410Hm Seems a bit concerning if the login templates can be used anywhere but by bin/login - does it make it easier to collect passwords in a spoofed login? [23:56]
Babarfunny.. how come this uni is called NSW, and it includes Canberra (which is in ACT)? [23:57]
jayeni believe it does
we have a campus there
we had a campus in singapore at some point, too
[23:57]
Babaryeah, like Monash in Melbourne, which has a campus in south africa, and I think one in Asia too :) [23:58]
jayenyeah, unis rename when they merge, not expand [23:58]
Babartrue [23:58]

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