#foswiki 2012-03-28,Wed

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

WhoWhatWhen
harlanIs there a way to "force" people to change their password after a password reset? I'm OK with this action being a choice, but I've got a couple of sites where the users never change their new passwords and then complain when the passwords expire. [00:26]
.......... (idle for 46mn)
schoosharlan: password expiration is a bad idea IMO. it leads to insecure passwords [01:12]
foswiki_irc7User admin question - I want to re-order all the Webs - not using Alphabetical order as current. How do I go about this? [01:22]
gac410Set up a static web list. That's faster anyway than using the automatic lists. [01:23]
foswiki_irc7OK, cheers. So change WebLeftBarWebsList to a bulletted list of the Webs I want users to see? This also hides System and Main webs from users too - I like it. [01:27]
gac410foswiki_irc7: yes that 's the solution. (Sorry for disconnecting, my connection is pretty bad) [01:40]
............... (idle for 1h13mn)
foswiki_irc7Thanks gac410 - looks great. My departments foswiki setup is so impressive it has been adopted across the division now, each with their own Webs. Fantastic stuff. [02:53]
gac410Great! [02:53]
pharveyfoswiki_irc7: I also wrote up http://foswiki.org/Support/BestPracticeTip20 which might be educational (although you probably have everything you need already) [02:55]
foswiki_irc7Thanks pharvey - that's excellent - I'll have a read through.
Hadn't seen that before.
[03:00]
pharveyI try to write things up as http://foswiki.org/Support/FAQ or http://foswiki.org/Support/BestPracticeTips (I encourage everybody else to do the same :) [03:03]
FoswikiBot[ BestPracticeTips ] [03:03]
pharveynifty: http://foswiki.org/Support/BestPracticeTip27 is #5 and #1 in WebStatistics for Feb/March :) [03:05]
SvenDowideit: people really do like CSV :P [03:10]
I really need to update our prod wiki. These spammers are creating bogus accounts every hour. Git revert isn't hard, but it's annoying. [03:16]
....... (idle for 31mn)
harlanschoos: I think I heard that when folks reset their passwords the password/link they are sent expires after 30 days' time.
I could be mistaken...
[03:47]
............................. (idle for 2h22mn)
***ChanServ sets mode: +o MichaelDaum [06:09]
.......................... (idle for 2h9mn)
GithubBot[foswiki] foswiki pushed 1 new commit to Release01x01: http://git.io/HccMuQ
[foswiki/Release01x01] Item9526: Uncommitted translations in the last 4 hours - WikiTranslationGnome
[08:18]
***GithubBot has left [08:18]
FoswikiBothttp://foswiki.org/Tasks/Item9526 [ Item9526: Translation work (needed for Pootle auto-checkin - please donĀ“t close) ] [08:19]
.... (idle for 18mn)
***ChanServ sets mode: +o OliverKrueger [08:37]
GithubBot[foswiki] foswiki pushed 1 new commit to master: http://git.io/-u2pjg
[foswiki/master] Item10186: redesigned login screen - MichaelDaum
[08:48]
***GithubBot has left [08:48]
FoswikiBothttp://foswiki.org/Tasks/Item10186 [ Item10186: new NatSkin ] [08:48]
......... (idle for 40mn)
***OliverKrueger sets mode: +oooo ArthurClemens Colas Lavr JulianLevens [09:28]
OliverKruegerOliverKrueger pings MichaelDaum [09:34]
MichaelDaumpong [09:34]
OliverKruegerI think MetaCommentPlugin is broken again. [09:34]
MichaelDaumis it
why
[09:34]
OliverKruegercannot save comments; leaading to a raw jsonrpc call. [09:35]
MichaelDaumargh [09:35]
OliverKruegerlooks like the css (and/or js) is missing in normal view. [09:35]
MichaelDaumgot some work for a 1.11 release ... checkingin [09:35]
OliverKruegerI tested the svn version from yesterday. [09:36]
MichaelDaummea culpa [09:36]
OliverKruegerthanks. :)
another thing... at CeBIT you fixed a bug concerning the content of an action-var while saving.
I updated NatSkin et al. yesterday an experiencing this bug now.
[09:37]
MichaelDaumrelated to foswiki trunk changes in Foswiki::UI::Manage [09:38]
FoswikiBothttp://trunk.foswiki.org/System/PerlDoc?module=Foswiki::UI::Manage [09:38]
OliverKruegerim running 1.1.5rc1
yes, George (afaik) changed something in UI::Manage.
I encounter the problem when I try to edit/save topic settings.
I thought, you already fixed that, too. ;)
[09:38]
...... (idle for 27mn)
MichaelDaumq: Foswiki::Func::searchInWebContent is deprecated ... but what's its replacement? [10:07]
FoswikiBothttp://trunk.foswiki.org/System/PerlDoc?module=Foswiki::Func [10:07]
........ (idle for 36mn)
pharveyMichaelDaum: Foswiki::Func::query IIRC [10:43]
FoswikiBothttp://trunk.foswiki.org/System/PerlDoc?module=Foswiki::Func [10:43]
MichaelDaumtoo late found it out by myself [10:43]
pharveyI'll fix the doc
at least I will have a single checkin to show for what has otherwise been an unproductive day :/
[10:43]
SvenDowideitok, that was fascinating.
we do have a problem :)
at least half the company likes make their topics look fun&pretty
and don't like the idea of being forced to separate content from look&feel
[10:50]
pharveySvenDowideit, is that a problem? [10:59]
SvenDowideity, we don't support that very well
wysiwyg breaks it
and they want to use wysiwyg _and_ add pretty
[10:59]
MichaelDaumOliverKrueger, made a new release of MetaCommentPlugin, have a try. [11:00]
SvenDowideitplus that skin templates are a woeful omg more to learn thing
mmm, as i think further about the reactions, INCLUDE based skins make alot more sense for users
as its really hard to find the name of a skin definition that you want to augment/replace when you don't have file access to the templates dir
[11:00]
GithubBot[foswiki] foswiki pushed 3 new commits to master: http://git.io/yEZO8A
[foswiki/master] Item11700: - MichaelDaum
[foswiki/master] Item11700: preparing release - MichaelDaum
[foswiki/master] Item11700: why was that disabled? - MichaelDaum
[11:02]
***GithubBot has left [11:02]
SvenDowideitlet totally alone the confusing new syntax [11:03]
FoswikiBothttp://foswiki.org/Tasks/Item11700 [ Item11700: add rss feed for all comments in a web ] [11:03]
MichaelDaumINCLUDE -> runtime ... TMPL:P -> compile time [11:04]
SvenDowideitMichaelDaum, do you have an opinion about the idea of replacing the tmpl stuff with INCLUDE (assuming that skin path and that kind of thing is be aded to INCLUDE, and that its performs at least as well?)
why not totally remove TMPL:P?
[11:04]
eugenmayerHello guys [11:04]
MichaelDaumbackwards comp'ty [11:04]
SvenDowideitbasically - do we have a real business case for 1. an extra DSL?
MichaelDaum, ok, agreed - we can remove that by making a conversion script
[11:04]
MichaelDaumTMPLs are cool actually [11:05]
SvenDowideitmore?
only for the ~40 people in the world that grok it
[11:05]
MichaelDaumINCLUDE already does too much for my taste [11:05]
SvenDowideitmmm, taht would be good to explain
TMPL:P ~~ STARTSECTION ENDSECTION, TMPL:DEF ~~ INCLUDE
[11:05]
MichaelDaumthe cool thing about TMPLs is its magic searching for the right thing [11:06]
SvenDowideity, the skin path
and that we have a feature for INCLUDE that will do the same
(plus or minus little details)
[11:06]
eugenmayerWe are still "owning" / rather maintaining the foswiki.eu / foswiki.de Domains. We would like to hand over those to the association [11:06]
MichaelDaumwikiapp devs normally only overlay a hand full of tmpldefs: beforetext, content, aftertext, sidebar ... stuff like that. [11:06]
SvenDowideiteugenmayer, ooo, nice
MichaelDaum, do we have a doc for that?
[11:06]
eugenmayerWe registered them back then, to protect them of others [11:07]
MichaelDaumthe other thing is that section defs can't be redefined [11:07]
SvenDowideitMichaelDaum, they can :) [11:07]
eugenmayerSvenDowideit: they of course belong to Foswiki. How to arrange the handover quickly? [11:07]
SvenDowideitok, imo you're focussing on impl details there
eugenmayer, i don't know
as we didn't hand over the .com .org etc either
the board decided it wasn't desperalty needed
so we've not done it
[11:07]
MichaelDaumredefining things in tmpland is _the_ number one way to interact with the page anatomy [11:08]
SvenDowideiti think OliverKrueger and MichaelDaum will have to do something tho
MichaelDaum, y, at the moment
but i can't think of one reason why its superior to using in wiki familiar syntax
and one thats _more_ accessible to wiki learning people
[11:08]
pharveyhi eugenmayer. I wonder who has the keys to foswiki.org [11:09]
SvenDowideiti 'own' .org.com.net
and i have an agreement that i will hand them to the association if they ask
[11:10]
eugenmayerso i hand over the others to you also sven? [11:10]
MichaelDaumSvenDowideit, if it is missing docu only, then I don't see how this motivates a paradigm shift. [11:10]
pharveypharvey would've thought the association should own it/them :P [11:10]
SvenDowideiti'm pretty sure my registra does not do .de
pharvey, yes, but there's paperwork and effort required
so we decided there was no need to
[11:11]
MichaelDaumbesides there are some fundamental differences between INCLUDE and templates: the first is a function call, the second is a view on data ... not the same [11:11]
pharveyah, yeah. Actually, I'm having a bigger problem getting SSL certs out of a domain we're adopting at $work :( [11:11]
SvenDowideitMichaelDaum, i will try to figure out what that means [11:12]
OliverKruegerback [11:12]
SvenDowideitas it might be the important thing, though i can't see it atm
MichaelDaum, OliverKrueger
can you please work out how to receive eugenmayer's donation of domains
[11:12]
OliverKruegereugenmayer: are these domains prolonged, soon? [11:12]
SvenDowideiti only hold the ones i do because i'm happy to - though i too would rather xfer them to the association _before_ someone gets annoyed at my power
see, my awesome power :)
[11:13]
OliverKrueger;) [11:13]
***OliverKrueger sets mode: +o pharvey
kip3f has left
[11:13]
SvenDowideitMichaelDaum, the core reasons imo to simplify to one language, is that we drastically reduce the learning curve, especially for newer view_template app devs
i'm not totally sure that INCLUDE is only a function - as i've seen enough uses of INCLUDE as a view on data - basically where the parameters are used to make the section be a 'format'
[11:17]
MichaelDaumI see the motivation, though I don't necessarily agree to a templating language based on INCLUDE to be any easier at the end of the day [11:18]
SvenDowideitin the _large_ it probly won't be [11:19]
MichaelDaumau contrair: we are missing essential features in INCLUDE part of TMPL:P... [11:19]
SvenDowideiter, please explain - parse error [11:19]
MichaelDaum*if* we were kicking out our old template language, then we'd aim for something commonly known on cpan. [11:20]
SvenDowideitas i say tho, this is predicated on INCLUDE getting enough extra juice to do what TMPL does
if we want to support _users_
then we would choose somethig we make them learn
and INCLUDE is (from what i've seen) the first and promary tool that users want and come to grips with
[11:20]
pharveyI always thought of adding 'path' support to INCLUDE as being an iterative step towards doing *some* of what TMPL:P is currently used for [11:21]
MichaelDaumsee, there is always a "learn" item when composing data views [11:21]
SvenDowideitpharvey, it is
that was the motivation
[11:21]
pharveyright [11:21]
SvenDowideity, and i want to reduce the number of things a non-programmer has to learn
so same syntax as they've been using for years is a plus
[11:21]
pharveyI see Michael seems to be concerned that we shouldn't overload INCLUDE too much and make it all SEARCH-tastic just for the sake of ditching TMPLs [11:22]
MichaelDaumfor now, it seems you are only replacing one item with another people must grok [11:22]
SvenDowideitits all learning, but instead of 3 things
they only learn 2
[11:22]
MichaelDaumthats cheating ;) [11:22]
SvenDowideitand INCLUDE (seems) to be something they learn much earlier
no, thats what its about
reducing the amount of things that are needed so its easier
[11:22]
pharveyI think there's no harm in making INCLUDE support fall-back/'path' type behaviour. But not as a mechanism to replace the whole skinning infrastructure. Just to aid wiki-apping. [11:24]
SvenDowideitpharvey, its to aid people taking the next step [11:24]
pharvey... unless, of course, we do this and then somebody makes a TMPL-less skin... [11:24]
SvenDowideitimo, once INCLUDE has the fallback feature [11:24]
MichaelDaumalso remember: INCLUDE will always be executed on runtime, that is on every click. TMPL:Ps arent: they are done on compile time and cached from there on. [11:25]
SvenDowideiti think there are only corner cases
and I can compile TMPL's into topics of INCLUDEs
MichaelDaum, na, thats only in our current crap impl
TMPL's aren't actually cached well, and INCLUDEs aren't either
[11:25]
pharveyINCLUDEs are horribly in dire need of some TLC for perf [11:26]
SvenDowideitso are TMPL's actually [11:26]
pharveyI added a unit test to show some timing tests at how slow they can be :P [11:26]
SvenDowideitand if they are synonyms
then one fixes the other
[11:26]
pharveyI'm a bit terrified of making them synonyms. [11:26]
SvenDowideitok, so, we have no known strong reasons that it can't be done
pharvey, bah - you'll get over your small minded fears when you see it working :p
[11:26]
pharveyI like that INCLUDEs do exactly as they say. You don't need any out-of-band knowledge to determine what it's doing. [11:27]
SvenDowideity, and that will remain [11:27]
pharveyTMPL:P OTOH, requires knowledge of what the path is, and where things are [11:27]
SvenDowideitand?
SvenDowideit wonders what odd implementation you're dreaming
[11:27]
pharveythat's the exact magic which people hate about TMPL [11:28]
MichaelDaumhate-love [11:28]
SvenDowideitok, pharvey i think you've gotten lost in a bad branch [11:28]
pharveyan INCLUDE{"This,That,TheOther" is far, far less magical than TMPL:P{"form:footer" [11:28]
SvenDowideitTMPL:P{banana} would be synonymous with INCLUDE{path="%SKIN%" section="banana"}
or something similarly explicit
[11:29]
JulianLevensFWIW, I agree with the principle of moving TMPL -> INCLUDE. There has always been an action at a distance anti-pattern feel about it all [11:29]
pharveyI am wary of 'moving'... as long as we keep the 'action-at-a-distance' aspect, out of INCLUDE, then SvenDowideit I think I'll be mostly happy [11:30]
SvenDowideitbasically, i was hit today with a room full of non-technical users that were perfectly happy to use INCLUDE and STARSECTION
but once i mentioned the idea of VIEW_TEMPLATE, glased over and ran away
[11:31]
pharveyyeah, teaching TMPL:P is a nightmare. [11:31]
SvenDowideitthey would have been fine with the syntax being familir [11:31]
pharveyI've never even mentioned VIEW_TEMPLATE. FooForm, FooViewTemplate, topics only :} [11:32]
SvenDowideitbut given that i was asking them to 1. consider dataforms, 2. consider SEARCH 3. consider ... and then added a new DSL
y, FooViewTemplate, == VIEW_TEMPLATE
[11:32]
pharveyit's not just the syntax though. It's the concept of the n-dimensional, non-deterministic tmpl-search-path [11:33]
SvenDowideitthe problem was that rather than introducing one new thing, i had to add 4
na,
that can be washed over
as its a COVER
[11:33]
pharveyyou gotta be kidding [11:33]
SvenDowideitno
for an indrodcution to it
you can wash over details like that
but you can't wash over new language at all
[11:33]
foswiki_irc2hi i was wondering how to get to links to lotus notes working in foswiki at the moment the links just come up as normal text [11:34]
pharveyPeople spend hours and give up poring over tmpl files trying to add "HERE I AM" debug text just so they can figure out which DEF something comes from [11:34]
SvenDowideitwhat do the links look like? [11:34]
foswiki_irc2let me get an example [11:34]
SvenDowideity, and that is the wrong thing for view_template [11:34]
foswiki_irc2[notes:///8025730000314FE5/A9307C2B01E89DCB802571EE0052C167/BF80203FA4D3810480257323005276FF] [11:36]
SvenDowideitfoswiki_irc2, in configure (the admin UI) there is a setting telling foswiki what is a link
you need your admin to add notes:// to it
[11:36]
pharveyas long as we keep INCLUDE, as MichaelDaum says, function-call-tastic (explicit, self-contained, deterministic...) [11:36]
SvenDowideiti think its a misc expert option
i don't propose to change the existing goodness
personally, i think templates are highly deterministic
[11:37]
MichaelDaumoh I don't want to hard-code the search path into every function call! [11:37]
SvenDowideitMichaelDaum, no, you would not
that would be crap
[11:37]
MichaelDaumy [11:37]
SvenDowideitthere would have to be some way to tag the INCLUDE as using the SKIN path [11:38]
JulianLevensWhile I am sure templates are highly deterministic, it's often a lot of effort to determine what will happen [11:39]
SvenDowideitthe impl detail is that it has to all work at least as well as what we have, which includes that TMPLs have to be automatably converatble
JulianLevens, agreed :)
[11:39]
pharveymassively agreed [11:39]
SvenDowideitunless you're, er.... me or MichaelDaum
i just breath in the air from your server, and i know what it is
ok, one more person :)
ArthurClemens, can you think of a show stopper reason that we could not improve INCLUDE and then replace the TMPL language with it
[11:39]
pharveythat's the biggest problem I've faced; I've trained two graphic designers and two software developers and a couple of student software engineers Foswiki templates, and I can't say it was fun in the least. [11:41]
SvenDowideiti like the feel of MichaelDaum's INCLUDE~function TMPL~view, but sadly, i don't agree that either DSL enforce that, so its just a style choice
pharvey, thats why i use no-skin for them
[11:41]
pharveyexcept they're maintaining our skin! [11:42]
SvenDowideitwhere the skin == foswiki.them.tmpl, and contains 2 tmpl:def's
one for 'before topic TEXT' and one for 'after TEXT'
and they go do what they do :)
makes twiki/foswiki people shudder, but it works well for those types of places
cmcrossroads.com 's wiki is an eg
[11:42]
pharveyI've thought about it, but I just haven't seen any skin structured that way before. And they still must fight templates to do things with NatEdit customization, SolrPlugin defs, FooForm/ViewTemplate... [11:43]
MichaelDaumin confluence every little fart is a template. they massively embraced the view concept. [11:43]
SvenDowideity, it means you are using default skin
we can't quite do that, cos some ASDSD coders like to use CGI to create html constructs
[11:44]
JulianLevensAs you have ideas about how to compile/cache templates, presumably you have something similar in mind for INCLUDE handling? With my SQL head on I've been vaguely wondering how that might help either of these, without actually being the way it should be done of course [11:44]
SvenDowideiter, s/ASDSD/ACCCC/
tmpl compiling i've written before - there's a wikiring blog post about how much speedup it gives
INCLUDE - no preconceived ideas
just needed to test the idea that we could actually reduce our DSL count
[11:45]
JulianLevensDSL = ? [11:47]
SvenDowideitdomain specific language [11:47]
JulianLevensthanks [11:47]
SvenDowideitas we have TML, HTML, CSS, JS, TMPL, DataForm definition and more
and each reduces the cadre of users willing or able to take the next step
for 1.2, i intend/hope to reduce DataForm + simpler search + edit to a wizard style
but it hits a big wall when you want to make a view/edit template for that app
[11:47]
pharveyACCCC? Almost a couple of codons [11:50]
SvenDowideitas MichaelDaum points out, there are only very few TMPL:DEF's that will cover alot of uses
so i'll probly go with that
but longer term, if we can simplify
[11:50]
foswiki_irc2thanks sven that worked ! [11:51]
SvenDowideitfoswiki_irc2, sweeeet :)
btw pharvey that https/http cookie patch that i put into 1.1.5
hasn't helped these guys
[11:52]
pharveydamn [11:53]
SvenDowideiti'm 99% certain that its because they have about 2million dns names for that server
and hardcoded url's in their topics
so its exhibiting the same symptoms, for the same cookie reasons
but gads its painful to try to track
[11:54]
pharveyif they're different domains, they shouldn't clobber the other domains... or you mean they're all mixed http/s? And they're using all forms of both? [11:55]
SvenDowideitall forms of things [11:55]
foswiki_irc2ah just come across another issue, the jquery plugin has a known issue with it and its asking me to install the latest release from a certain address, how do i do that? [11:55]
SvenDowideithttp://one.server.domain, http://someother.server.domain, https://one.server.domain, https://someother.server.domain, times 3
so i imagine they log in to http://one.server.domain, which gets them some req's to the others
that toasts the session somewhere
so they can reasonably reliably get both the login screen in edit
and have edit be OK, hit save and it saved the tmpllogin html to their topic
somewhere in the apache conf, someone started adding mod_rewrite stuff that matched directly to the scripts - to avoid stacks of redirects - and that really is uhg
foswiki_irc2, configure has a 'install extensions button
unless you're using debian / yum repo
[11:56]
foswiki_irc2aye it does sven, but if you look at the description of the jquery plugin it says Note. The newer versions of this plugin expose an error in Configure which will prevent the configuration from being saved. [[Tasks.Item11136]]. Install a new [[http://trac.foswiki.org/export/12631/branches/Release01x01/core/lib/Foswiki/Configure/Valuer.pm][lib/Foswiki/Configure/Valuer.pm]] to resolve. [12:01]
SvenDowideitSvenDowideit hopes that the persons that write that can help :)
basically, grab the file linked to, and copy it over the file on your server
[12:02]
foswiki_irc2Thanks Sven i think that fixed the configure issue, youve been a star !
thanks for all the help
[12:08]
SvenDowideitno worries :)
we're here lots :)
[12:08]
flexibeastArgh. Why am i suddenly getting "Attachment '?' does not exist" errors? [12:22]
pharveyflexibeast: when you view attachments?
there was an old RestPlugin bug which got the UI switchboard mixed up, but I guess you're not running hairy unreleased stuff from svn? :)
[12:30]
flexibeastOr try to. - Think it's a problem with the RewriteRule in my .htaccess, i.e. that i'm not understanding how viewfile works properly. Trying out something now.
No, 1.1.4. :-)
Hm.
Better, but what if the filename i'm passing to viewfile contains spaces?
They seem to be getting stripped.
[12:31]
pharveyflexibeast: I think there's an issue there. I think Foswiki tries to sanitize the attachment name... I forget how/where (it depends on SiteLocale? I think. Or UseInternationalizatoin... MichaelDaum might know) [12:35]
flexibeastpharvey: Ah, okay. [12:35]
pharveyyou will notice that Foswiki tries to replace spaces in attachment names with underscores
if you upload with the UI
[12:35]
flexibeast*nod* Very sensible indeed. :-) [12:35]
pharveynah, it's lame
CDot removed most of this topic/attachment-name mangling stuff in the unicode branch
but that's not going to be ready for Foswiki 1.2, I think
[12:35]
flexibeastOh? Well, given that i had to spend some time earlier putting underscores in certain folder names so that they passed taint checks, auto-underscores seems good to me .... [12:36]
........ (idle for 37mn)
fsfsI'm testing redirecting with POST passthrough, with TRACE_CACHE set to 1 in lib/Foswiki/Request/Cache.pm - and it seems that when running as cgi-script, the cache is not read in again?
(and when running under mod_perl, the cache is read in again, but somehow the name of the topic I want to create is lost, which is why I'm looking at it in conjunction with FormPlugin)
that is, foswiki tells me: The "WebHome" topic already exists
[13:13]
..................... (idle for 1h43mn)
***ChanServ sets mode: +o OliverKrueger [15:01]
.......... (idle for 45mn)
GithubBot[foswiki] foswiki pushed 1 new commit to master: http://git.io/PUW7hA
[foswiki/master] Item11250: fixing typo, JSON::XS 2.27 not 2.270 - OliverKrueger
[15:46]
***GithubBot has left [15:46]
FoswikiBothttp://foswiki.org/Tasks/Item11250 [ Item11250: JsonRpcContrib doesn't depend on JSON ] [15:46]
.............................................. (idle for 3h47mn)
Rich_MorinSvenDowideit: yt? [19:33]
........... (idle for 53mn)
***ChanServ sets mode: +o OliverKrueger [20:26]
gac410fsfs: We are in "release candidate" state. Please don't commit changes to the release branch unless it's an urgent core bug. We should not be committing anything into 1.1 branch unless it's a urgent core bug. [20:35]
GithubBot[foswiki] foswiki pushed 2 new commits to Release01x01: http://git.io/_XzMPA
[foswiki/Release01x01] Item11690: add $status to Foswiki::Func::redirectCgiQuery and use only that to issue a 307 redirect - FlorianSchlichting
[foswiki/Release01x01] Item11690: do not double urlEncode param name, and urlDecode trace output in Foswiki::Request::Cache to match appearance of input - FlorianSchlichting
[20:35]
***GithubBot has left [20:35]
FoswikiBothttp://foswiki.org/Tasks/Item11690 [ Item11690: FormPlugin: global variables, double Location HTTP headers a problem under mod_perl ] http://trunk.foswiki.org/System/PerlDoc?module=Foswiki::Func
http://trunk.foswiki.org/System/PerlDoc?module=Foswiki::Request::Cache
[20:36]
gac410Any changes now that we've built a release candidate ... should really go through another round of testing and a new release candidate. [20:36]
GithubBot[foswiki] foswiki pushed 3 new commits to master: http://git.io/3PfAgQ
[foswiki/master] Item11690: add $status to Foswiki::Func::redirectCgiQuery and use only that to issue a 307 redirect - FlorianSchlichting
[foswiki/master] Item11690: use amended Foswiki::Func::redirectCgiQuery() API to do a proper 307 redirect - FlorianSchlichting
[foswiki/master] Item11690: do not double urlEncode param name, and urlDecode trace output in Foswiki::Request::Cache to match appearance of input - FlorianSchlichting
[20:36]
***GithubBot has left
OliverKrueger sets mode: +ooo gac410 Colas ArthurClemens
[20:36]
gac410Thanks OliverKrueger [20:37]
OliverKrueger;) [20:37]
fsfssorry, I asked about that in Item11690 -- feel free to revert those commits rather than delaying the release.
but the changes are minimal, really, can't see how they could break anything...
[20:39]
OliverKruegerFamous last words... ;) [20:40]
fsfsyeah... ;-) [20:41]
gac410We've had other releases with issues after "minor" last minute changes. So unless it's really an urgent bug in core reported against the release candidate build, I'd rather defer.
Before you revert, let's see if Babar can review this in some detail. Have you written a unit test.
[20:42]
BabarBabar already commented on the bug report
and as Florian implemented it exactly the say I said he should, I have no objection
I said it should require a feature request, though :)
[20:43]
fsfshmm, did you? [20:44]
gac410That's what I was wondering ... It is an API change and those are subject to review. [20:44]
BabarI vote for the latter: fix redirectCgiQuery. I have to admit I don't really understand why the browser must wait for the user's confirmation before getting the page, but I trust you on this.
Please, if you code the fix, add a basic check that the status is a 3XX, like: $status = $status ? $status =~ /^3[0-9][0-9]$/ ? $status : 302 : 302;
-- OlivierRaginel - 23 Mar 2012
ok, I never said it required a feature request
must have only mentionned this here then, sorry
[20:45]
gac410If you are all really really sure of this one, I'm not going to play the "Draconian Release Manager" and force a revert. But I'd hate to have this one be the issue that forces a withdrawal of 1.1.5. We are around 5 days away from release. [20:46]
BabarBabar reviews the code once more [20:48]
fsfsthanks [20:49]
Babarhum... there might be one thing... lemme double-check [20:51]
gac410If there has been an API change, do we have to increment the API version ... (And how does that impact 1.2, which is already using a new API version. :( [20:58]
Babarno, ok, the code is fine. It was already in as you wrote, you just wrapped it up
as for your questions gac410, I have no idea. Lavr or CDot or Sven might...
[20:58]
gac410I think we really do though - otherwise how can an extension know if the new API option exists.
How bad is the issue if we don't make the API change in 1.1.5?
Ooh... maybe not a problem. Foswiki 1.1.4 is API 2.1, Foswiki 1.1.5 is API 2.2 (Someone already made a change). Trunk is API 2.3
[20:58]
fsfs...seems like I may end this day lucky... [21:03]
gac410Yeah. pharvey un-deprecated Foswiki::Func::getScriptUrlPath() [21:04]
fsfsgac410: I'm really sorry I'm causing you such an amount of trouble! [21:05]
gac410nah no trouble. Just making sure the i's are all dotted, and t's all crossed ;)
I've stretched the rules a bit here and there as well. well. occasionally slaughtered not just stretched.
fsfs, please make sure that the API change is thoroughly documented in Func.pm with dates of changes, etc. I'll check back later.
[21:05]
fsfsdone, hope that's ok. I'll be back tomorrow [21:21]
.... (idle for 16mn)
GithubBot[foswiki] foswiki pushed 1 new commit to Release01x01: http://git.io/swGEFg
[foswiki/Release01x01] Item11690: add a *Since* date to API extension - FlorianSchlichting
[21:37]
***GithubBot has left [21:37]
gac410thanks fsfs [21:38]
FoswikiBothttp://foswiki.org/Tasks/Item11690 [ Item11690: FormPlugin: global variables, double Location HTTP headers a problem under mod_perl ] [21:38]
GithubBot[foswiki] foswiki pushed 1 new commit to master: http://git.io/JQSIGg
[foswiki/master] Item11690: add a *Since* date to API extension - FlorianSchlichting
[21:38]
***GithubBot has left [21:38]
gac410Hm Foswiki::Func::redirectCgiQuery() has no unit test coverage :( [21:45]
FoswikiBothttp://trunk.foswiki.org/System/PerlDoc?module=Foswiki::Func [21:45]
..... (idle for 22mn)
ArthurClemen[13:40:56] <SvenDowideit> ArthurClemens, can you think of a show stopper reason that we could not improve INCLUDE and then replace the TMPL language with it
INCLUDE would need to read the skin path
but I guess you had that in mind already
is the idea to deprecate TMPL:DEF in favor of STARTSECTION?
there is also TMPL:INCLUDE
wich is actually a kind of import
so would a top INCLUDE make available the other "imported" STARTSECTIONS?
[22:07]
but this needs a brainstorming topic obviously [22:17]
......... (idle for 43mn)
harlanSo if we send registration emails to a moderator, I know it's easy to approve the reg by clicking on a link. I don't recall seeing a similar easy way for the "registrar" to click on a link saying "Delete this reg and the users' registration page" or perhaps a method saying "Delete all non-approved registrations that are more than 30 days' old"
I'm also thinking we might want to add a text block to the registration form that we'd use to hold the answer to "Please tell us why you want to use our site"
[23:00]
gac410AntiWikiSpamPlugin has a new remove user tool. However needs 1.1.5 to do a complete job,. [23:04]
pharveyharlan: I thought the user topic wasn't created until it had been approved [23:05]
gac410(the TopicUserMapping code before 1.1.5 doesn't have the routine to remove a user from the WikiUsers topic. ) [23:06]
pharveyor is that just the htpasswd entry
pharvey wonders why the peinding approvals file has all the submitted form data if that's the case :)
[23:06]
gac410No I think you are right pharvey. If the registration is pending approval no topics are created. [23:06]
pharveyjust wiped out 50 spam registrations overnight :/
can't upgrade prod until I do some natedit fixing
[23:07]
gac410foswiki is down to 6 or so a day now - all rejected by the new filters. [23:10]
harlancool - thanks guys! [23:11]
gac410AntiWikiSpamPlugin version in svn will also remove users from group definitions.
harlan ... I think there is a disabled Comment field that could be enabled in the registration. You could use that to hold the "Why you want to use ..." question. Fields like that are spam/abuse magnets.
[23:11]
harlanThat might be fine - I'm looking more for a note that would only go in the registration email so the moderator could see why the registrant says they want to join. [23:14]
gac410yeah I don't think there is anything like that [23:15]
harlanFor us, that would likely be "to edit/maintain my clock entry" or "I want to fix typos or add content to the Support pages" for normal folks, and who knows what lame-ass reasons we'd get from the spammers (if anything). [23:15]
pharveyon our wiki we have a 'Project' field, which isn't mandatory. A comment thing would be great for those users who aren't sure which project they want to join.
But I have seen the spammers fill the Project with words apparently scraped from the main webhome
[23:21]
harlanI bet. Right now we are seeing registrations from throwaway accounts, and we are recognizing those domains.
If we could get additional support for working on this problem (perhaps as a plugin that would extend some of the reg info?) we'd be in a better position to do something post-1.1.5
[23:22]
pharveyit sounds as if George's work should go a long way to improving the situation
nothing like a spam storm to highlight the inadequacies of our OOB user admin features :)
[23:27]
harlan:) [23:27]
gac410And CDot's work too.
We've both done pieces of the AWSPI work.
[23:27]
pharveySvenDowideit: you demonstrated a way to easily render forms for edit without FlexFormPlugin's RENDERFORDISPLAY, something we should do is update the System.FrequentlyAskedQuestion topic to use that pattern instead of hand-crafted HTML [23:28]
SvenDowideitgac410, are you ontrack for a release this w/e? [23:29]
pharveyI was just reading VickiBrown's collection of help docs http://wiki.cfcl.com/bin/view/Help [23:29]
SvenDowideitmy main worry right now is that strikeone issue when using the 'New TASK' popup
but i don't think i can do anything about it in the next 7 days
[23:29]
pharveyoh, sure [23:30]
gac410SvenDowideit: Early next week. [23:30]
SvenDowideitand as its somewhat new - in that no-one else is doing that
its probly ok to skip it
gac410, sweet :)
pharvey, not until it works :)
i get strikeone's every time (except the first) i use the dialog
[23:30]
gac410SvenDowideit: unless of course another blocker shows up. [23:31]
pharveyI wasn't really thinking of the ajax-tasticness, but rather just a simpler way to embed a dataform's edit fields into a 1999 style topic page [23:31]
SvenDowideitsomewhere something is using an old/misatched strikeone value
i'm told it works perfectly in chrome
[23:31]
pharveyargh
SvenDowideit, do they have a reliable way of reproducing?
[23:32]
SvenDowideityes
use firefox
and goto foswiki.org/Tasks
log in
and you should see 'create task' and NEW task'
[23:32]
pharveyI'm not referring to the strikeone issue for your tasks thing
or are you saying it's the same problem
[23:32]
SvenDowideityes [23:32]
pharveyI thought you were saying the session cookie was getting stomped
but in that case it's re-using the strikeone value more than once, innit?
[23:33]
SvenDowideitoh, y, that i can only repro on site
losing session i'm reasonably convinced is due to there being lost of hacked in rewrite's and too many hostnames
but i can't test until i get a client ssl cert
so that too might be lae ext week :/
[23:33]
pharvey:( [23:34]
SvenDowideitlate next
i'm broken
[23:34]
pharveyb0rken [23:34]
SvenDowideitsick for the 3rd week running, all day training course yesterday and now looking after sick kids
insert sleep :
ok, bbl
[23:35]
gac410How did the course go? [23:35]
harlancan you use a cert from cacert?
feel better SvenDowideit ...
[23:35]
pharveyI assume sven means an X.509 client cert
sven briefly mentioned WYSIWYG hassles http://irclogs.foswiki.org/bin/irclogger_log/foswiki?date=2012-03-28,Wed&sel=57#l53
[23:36]
harlanI think cacert stuff can be used to generate an x.509 certificate.
yes, I found several links by searching for "cacert x.509"
[23:38]
pharveywouldn't the customer site admins have to make one for him? [23:38]
harlanah, I wasn't sure who the "players" were. [23:39]
pharveyneither do I :P [23:39]
gac410WikiRingBot: seen lavr [23:39]
WikiRingBotWikiRingBot has last seen Lavr 27 Mar 2012 - 20:37 GMT on #foswiki [23:39]
pharveyFoswikiBot: seen pharvey [23:40]
FoswikiBotpharvey: pharvey was last seen in #foswiki 0 seconds ago saying "seen pharvey". [23:40]
pharveyhaha
FoswikiBot: FoswikiBot: seen pharvey
d'oh
pharvey makes coffee
[23:40]
gac410FoswikiBot: seen Lavr [23:40]
FoswikiBotgac410: Lavr was last seen in #foswiki 1 days 3 hours ago joining the channel. [23:40]
harlanWordpress has some interesting upgrade options - I can take a site that is "live" and reinstall the current version - it downloads http://wordpress.org/wordpress-3.3.1-no-content.zip and reinstalls that, which leaves existing content alone. Good for repair jobs. [23:51]
pharveywe have kicked around the idea of releasing core packages, I'm not aware of any technical reason why it can't be done, except of course trashing any shipped Topic.txt - which is why I guess the manual upgrade process is recommended [23:55]
harlanI still believe that if we think about it enough we can find a way, perhaps involving the VHC stuff.
The interesting thing is that if a topic under VHC is based on something in the 'main' instance, one could see that the main instance is "newer" and generate a "diff" which could be presented to the maintainer to see about merging in with local content.
[23:56]
SvenDowideitthe only reason upgrade is still manual [23:58]
harlanI'm also wondering if VHC sites might make it easier to have the USERSWEB be different from Main. [23:58]
SvenDowideitis that we need to do more :)
imo yes
I'm using a core pkg for the rpm's
and will move to that for 1.1.5 / 1.2.0 debs too
and will need to separate things in a similar way to make the usbstick upgrades
[23:58]

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