#foswiki 2012-03-19,Mon

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

WhoWhatWhen
GithubBot[foswiki] foswiki pushed 3 new commits to Release01x01: http://git.io/pfsWeg
[foswiki/Release01x01] Item000: BUILD Foswiki-1.1.5-Beta2 at Sun Mar 18 23:47:28 2012 GMT - GeorgeClark
[foswiki/Release01x01] Item000: BUILD Foswiki-1.1.5-Beta2 at Sun Mar 18 23:48:02 2012 GMT - GeorgeClark
[foswiki/Release01x01] Item000: BUILD Foswiki-1.1.5-Beta2 at Sun Mar 18 23:49:51 2012 GMT - GeorgeClark
[00:00]
***GithubBot has left [00:00]
FoswikiBothttp://foswiki.org/Tasks/Item000 [ Item000 ] [00:00]
AlexisHazellIs there any reason i shouldn't create a plugin which essentially allows one to browse a file hierarchy within the FW file hierarchy? [00:07]
pharveyboo! [00:13]
gac410howdy pharvey [00:13]
pharveyhi [00:13]
gac410AlexisHazell: Usually admins want to hide the actual file hierarchy to prevent exposing things to the evil crowd. [00:14]
SvenDowideitAlexisHazell, nope - so long as you're ok with accepting the hacking risks :)
everyone run away, its pharvey ;)
[00:14]
gac410pharvey ... 55 files are removed in the TMCE upgrade. Any issue for upgraders leaving them around? [00:14]
pharveymuhahah
hrm
pharvey looks at which were removed
[00:14]
gac410pub/System/TinyMCEPlugin/tinymce/examples/translate.html
pub/System/TinyMCEPlugin/tinymce/jscripts/tiny_mce/plugins/foswiki/editor_plugin_src.js
pub/System/TinyMCEPlugin/tinymce/jscripts/tiny_mce/plugins/foswikibuttons/attach.htm
and 52 others.
[00:15]
AlexisHazellgac410, SvenDowideit: Okay, thanks. :-) The UI would only be exposed to logged-in users. [00:16]
gac410depends on how well you trust the logged in community. ie on f.o anyone can register & log in [00:16]
pharveygac410: editor_plugin_src is 'removed' but not really, it was simply relocated. As for the examples/* these are developer files, really. /me checking the rest [00:17]
gac410grep REMOVE ~gac/foswiki/115b2check (on foswiki.org) [00:17]
pharveyI'm reviewing the diff on Release01x01 to MANIFEST [00:17]
AlexisHazellgac410: Hm, good point. i should really disable user registration. [00:18]
gac410AlexisHazell: or modify it to submit registration verification email to an admin.
SvenDowideit: I've built Beta2 ... about to install on f.o
[00:18]
AlexisHazellgac410: *nod* Well, overwhelmingly anyone who needs a login will need me to add them to one or more groups first anyway. [00:19]
SvenDowideitgac410, brilliant!
i'll make a usbstick 1.1.5 beta2 today
maybe :)
[00:19]
pharveygac410: I've reviewed the MANIFEST, I guess my only concern is to whether or not stale files in pub/System/TinyMCEPlugin/tinymce/jscripts/tiny_mce/plugins/foswiki(image|buttons|)/* will confuse the plugin loader, but I'm 99% sure it won't [00:21]
AlexisHazellSo, further to the above, i gather it's probably a bad idea to directly copy a batch of attachments to the relevant topic page folder in pub? (We don't want to use the BatchUploadPlugin method of creating a single zip file.) [00:21]
SvenDowideitAlexisHazell, no, its fine
just use the UpdateAttachmentsPlugin to then register them into the topic's meta
[00:22]
pharveygac410: after some more review, I am now 99.9% sure there is no need to worry about stale files [00:22]
SvenDowideitshould contemplate making that core-ish too [00:22]
pharveygac410: if you want an extra 0.1%, poke CDot [00:23]
AlexisHazellSvenDowideit: Ah okay, excellent, thanks. :-)
(These questions follow on from a four-hour discussion i had with my, er, 'boss', as it were, last night.)
[00:23]
SvenDowideitSvenDowideit burns the latest andriod for x86 onto an sd card anf then tries to find out if he can boot from it
AlexisHazell, ha! send 'im here and we'll set 'im straight :)
SvenDowideit goes for coffee bbiab
[00:23]
AlexisHazell*chuckle* [00:25]
gac410AlexisHazell: if you are building a plugin, you could easily either require a group to allow view, or add a custom approval ALLOWTOPICBROWSE for example. [00:25]
AlexisHazellHe's incredibly snowed under with work .... our discussion finished at around 11.30pm, and he still had a couple more hours work to do.
gac410: Oh yeah, good point. *slaps forehead* i am obviously not yet awake enough.
[00:26]
gac4101.1.5 Beta 2 is now running on foswiki.org. It has the new TMCE, and I removed any files not in the new manifest. [00:26]
SvenDowideitSvenDowideit asks gac410 for the url :)
oh, and i hope you didn't notice, but cdot broke f.o last night
he deleted all the important topics, (ok, so he deleted one, SitePrefs >:}
[00:33]
gac410He was hacking in more antiwikispam stuff I think oooh... that's nasty [00:34]
SvenDowideitsily duffer did a grep fpr all topics containing MARKEDFORDELETION and did.
course, that was defined in... :)
[00:35]
AlexisHazellOopsies. :-) [00:35]
SvenDowideitmucho panic :) [00:35]
gac410Probably one of those "sinking feeling ... knot in the stomach moments" [00:38]
SvenDowideityup [00:38]
gac410That was like the time I tried to "sed" my apache Includes directory and rewrote all my vhost definitions to be the same line .. over and over...
And then discovered that my backup script had a typo and was not copying etc.
[00:39]
SvenDowideitmmm, interesting
looks like files attached to Main.SitePreferences also go through viewfile
[00:40]
gac410gac410 doesn't use sed much any more. [00:40]
SvenDowideitneed to change that, and add an option to not use viewfile for anything attached to WebPrefs i think [00:41]
gac410yeah. I thought I had exceptions like that . but trying to decide what to bypass auth checking rather treacherous [00:42]
SvenDowideitpharvey, can we watch?
where's the video feed?!
gac410, yeah, its fun watching a new foswiki come live
tail -f events.log :)
[00:43]
gac410more interesting tail -f error.log [00:44]
pharveySvenDowideit: nah, but I'll be making a youtube video demo'ing Temi's thing, which will be 90% of the presentation [00:44]
SvenDowideitmmm, we log edit, but not save? [00:44]
pharveywe do log save [00:44]
gac410our logging is *very* inconsistent. Different formats, etc. [00:44]
SvenDowideitoh shit.
mod_fcgid: HTTP request length 139264 (so far) exceeds MaxRequestLen (131072)
[00:45]
gac410IMO the *engine* (or dispatcher) should log that a script has been called, and not depend upon each script to log. [00:45]
SvenDowideitok, so we really do need to log save
gac410, agreed
i want log on success tho, not on the begining of a script call
or better, success, or failure either way
[00:45]
gac410And really, our log messages should be very consistent for content. (As an old mainframe IMS guy, I'd log txn start, each I/O event, render start, render end ... )
every search ...
(start and end)
:P
Hm... so do I flip the new switch that blocks duplicate registrations for the same email address ... on f.o (yeah I know it's handy for devs. )
[00:46]
SvenDowideitstuff devs :)
we can work around it
i have an infinite number of email addresses, and if someone really needs some...
[00:49]
gac410edit config, turn it off, reload apache, register, ... Yeah me to. It's the benefit of running my own email server. I usually use a unique email for every registration so I can see who if anyone is selling email addresses.
(and really surprisingly I thing across 5 years or so, I've only had one email get abused.)
[00:50]
SvenDowideitbut then, you don't go wild :) [00:53]
gac410sigh... looked at inbox. we *really* ought to update the LASTBUILD at the end after verifying the build worked. [00:54]
SvenDowideit| 2012-03-19T00:57:40Z info | sven | viewfile | Main.SitePreferences | XXX_style.css Safari | 2001:dc0:2001:11::246 |
impressive
ipv6?
[00:54]
gac410sure looks that way [00:55]
SvenDowideitso we actually log ipv6 addresses - well done! [00:55]
gac410You are runing your server on v6? [00:55]
SvenDowideitnot i :)
this is the tmwiki->foswiki conversion
they're just testing the conversion i did last night (in a haze, hate working when sick)
[00:56]
gac410SvenDowideit: Do you think you'll fix that attempt I reverted yesterday? Plan is to build RC1 and go public with it 48 hours from now. [01:02]
SvenDowideitgac410, i'd like to, but i'll be surprised
the pager one is more important
ask me again in er, 46 hours :)
danm.
x61's don't boot from sdcard
now to go find a micro usb key
[01:02]
gac410pharvey - most important to me now is are there any TMCE tasks tagged for Trunk / 1.2 / 2.0 that should be blockers for 1.1.5 now that TMCE has been merged.
SvenDowideit: Can you boot grub and have it load the sdcard for you?
[01:03]
SvenDowideitdon't want to mess with the HD :)
its got win8 on it
[01:04]
gac410boot grub from cd or floppy [01:04]
SvenDowideitew!
what i'm wanting, is somethign i can boot to android on this tablet computer
without having a big usb key hanging out
http://www.ebay.com.au/itm/Smallest-T-Flash-Micro-SD-SDHC-Card-USB-Reader-Writer-/190449971386?pt=UK_Photography_MemoryCardReaders_RL&hash=item2c57b3f0ba#ht_2289wt_804
there's a little 8mm overhand over 2 of the usb sockets
so, i may well be in a luck there (though sdcard woulda been nicer
[01:05]
gac410ah. Speaking of "tablets" was playing with the new Samsung Note today .. pretty cool phone. About the same size as my dell streak, but bigger display, and lighter. [01:06]
SvenDowideitwow, they're really hitting this wiki hard, very nice
yeah, my iphone3g is starting to die, so i'm looking forward to thinking about a samsung note :)
cept, i'm a little person, so it'll look like i'm holding a tombstone next to my ear :)
[01:06]
gac410From my streak experience, bluetooth headset is a must, along with a belt pouch of some sort. [01:07]
SvenDowideitRCS: /usr/bin/rlog -h %FILENAME|F% of .../Netops/SoulNet2.txt,v failed: at /va
blah!
[01:07]
gac410corrupted file? [01:08]
SvenDowideitmmm, that might be the thing that turns me off - mobile gets stuck into pokets :)
checking :)
[01:08]
AlexisHazellAlexisHazell makes a mental note to check if a bug report about SPACEOUT's behaviour already exists. [01:08]
SvenDowideithehe - ,v file that is zero size
but same on tmwiki, so legacy :)
[01:09]
gac410The thing that **really** gets me angry with phones is the delay with the carriers "blessing" new software updates. (or refusing to update altogether). Spend hundreds of dollars for an orphan. using an open source OS. and blocking features (tethering, wifi) [01:11]
AlexisHazell*sad nod* [01:12]
SvenDowideitSvenDowideit doesn't expect any after purchase care from companies
so basically i refuse to buy until it does what i want, and then i hope i can stop it from auto updating
but then i now buy notebooks that are ~5 years old from ebay,
when up til 2005 i used to buy next day onsite warrant from ibm
[01:12]
gac410Okay - f.o appears stable. I've pushed the mod_expires back out to 17 hours. [01:15]
SvenDowideitoh hit.
editing is toast
why am i editing webhome
[01:18]
gac410on f.o?
working for me.
[01:18]
pharveygac410: Item11312 is the one I'm desperately trying to find time for [01:20]
FoswikiBothttp://foswiki.org/Tasks/Item11312 [ Item11312: TinyMCE Corrupting HTML Tables ] [01:20]
SvenDowideitna, me
i'm busy watching and fixing things onsite-ish
[01:22]
gac410okay great. And it's tagged for 1.1.5. If we don't fix, do I have to revert TMCE? Or consider a setting to block TMCE editor when table html is found? [01:22]
SvenDowideitgac410, don't forget to tell me where the zip file is
so i can make a 1.1.5b2usb
[01:22]
gac410SvenDowideit: okay great. :P on my laptop.
I'll start an upload My uplink bandwidth is nearly dial-up
[01:22]
SvenDowideitoh, ok, dont' then :)
i can build my own, i presumed you'd use that zip to install on f.o
[01:23]
gac410Started the upload. No I used the tgz. zip upload running now. It will be in ~gac/foswiki [01:24]
SvenDowideitgads, i can use the tgz too :)
don't waste your BW when you can use it for looking at .... comics!
https://www.ohloh.net/p/Foswiki/analyses/latest
[01:24]
pharveygac410: I'm torn, because actually there are a lot of nice bugfixes in the new TMCE & WYSIWYG. I don't think the actual error-case in Item11312 will have a huge amount of impact on most end-users, but I am concerned it might be indicative of a more serious bug that has a broader scope than the case that SteveLin highlights. In any case, we've been using trunk WYSIWYG for over a year, and the...
...users prefer it ...
pharvey wonders where his day went
[01:26]
SvenDowideitto the dogs!
didja win? :}
[01:27]
pharveywin? [01:27]
SvenDowideitif you goto the dogs, you should bet, and if you bet, you should win!
why else would anyone goto the dog races!
at the bottom of every topic in the Main web (of this upgrade site) I get a groups add/remove UI
as though that topic was a group
and using it, it correctly fails to add you to the group of that topic's name
[01:28]
gac410view template setting issue? (or autoview?) [01:29]
SvenDowideitthats probably it, but stuffed if i can find it [01:30]
pharveySvenDowideit: or Web/Site/DefaultPrefs leakage? [01:31]
SvenDowideitpharvey, can you esplain? [01:31]
pharveywe had a problem in Support web where VIEW_TEMPLATE was set in WebPreferences [01:31]
SvenDowideitoh wow. [01:31]
pharveynever mind, just brainstorming [01:31]
SvenDowideitthese guys even just tested auth failure
pharvey, ah, nice :/
[01:31]
pharveytested or encountered?
pharvey decides to eat food
[01:32]
SvenDowideitsameo - don't care :) they're testing quite alot of things, and i like that
oh, so the 'rest' call when they edit means they're doing a wysiwyg edit
why doesnt' the log just say so ffs :)
[01:32]
pharveyyou wanted the handler to be logged? Hehe. I wish it would do that too. Because it's hard to see jqgrid vs wysiwyg hits. [01:33]
SvenDowideiti would feel happier if the SystemSiteChanges we ship didn't die so much
pharvey, corect
really trivial to do it the way gac410 suggested
engine logs, preffereably with details
ooo, sitechanges is working on here atm
[01:33]
pharveySpeaking of SiteChanges, Item11244 breaks MongoDBPlugin, I need to fix that one day [01:34]
FoswikiBothttp://foswiki.org/Tasks/Item11244 [ Item11244: [[Item11215]] introduced a filterByDate method, which MongoDBPlugin lacks ] [01:34]
SvenDowideityeah, or when 1.1.5 happens, and this contract's done, i'll be looking at 1.2.0 and cleaning lots [01:35]
pharveySiteChanges uses a date="foo" thing, which for some reason post-sorts the results, rather than chucking the sort into the query
or at least, that's my hazy memory of triaging it 6 months ago
[01:36]
SvenDowideity, but even so, on lots of sites, it 500's [01:36]
pharveyit 500s with mongo too
even before arthur's change which just broke it
if you do a date="foo", it seems to post-sort the entire resultset
[01:37]
gac410neat ohloh stats ... but it is a bit deceptive. A one-line comment change commit counts as equally to a 20 file store rewrite. [01:38]
pharveyor am I lying
pharvey forgets
food time
[01:38]
SvenDowideitgrin
gac410, yeah, gotta start somewhere
oh crud.
so we deprecate rdiff
[01:41]
gac410yeah. Though it has some tricks that compare doesn't offer. [01:41]
SvenDowideitbut we leave it in the typical webchanges diff
and others have copied them
it does?
crud
[01:41]
gac410yeah. it also is in the mailnotify changes. [01:41]
SvenDowideit:(
so i can't rewrite rdiff to compare?
[01:41]
gac410well ... i think it's the type history ... where it shows every change that's missing. [01:42]
SvenDowideitok, do we have a task?
mmm, ok, i take it back
[01:42]
gac410I don't thnk so [01:43]
SvenDowideiti wrote rdiff in 2003
so i should probly wait til 2013 before i re-do it all
mmm, should probably not use upload with fcgi either
[01:43]
gac410Just tuning - need to match the max attach size to the fcgi max request size. [01:45]
AlexisHazellSvenDowideit: Are there any reasons why users shouldn't be allowed to select a template for a Talk page upon that page's creation? [01:45]
SvenDowideitgac410, booo :(
imo there is no reason to limit user's from doing what they want
template - as in topic template, form template, view template or edit template?
SvenDowideit regularly is in template hell
[01:45]
gac410FcgidMaxRequestLen 20000000 You need to increase the default. Which is ... hmm [01:47]
AlexisHazellMm, edit template? The idea is that different types of documents would have Talk pages with certain sets of headings etc. [01:47]
gac410SvenDowideit: Default is 131072 (and in FCGI docs: Before 2.3.6, this defaulted to 1GB. Most users of earlier versions should use this directive to set a more reasonable limit. )
Or as you say - bypass fcgi for the upload script.
[01:48]
SvenDowideitmmm, how do we hide content?
is there verbatim hide?
mmm, i guess i have to style=display:none
as the html comment is.... getting killed by ---++ or something
[01:58]
gac410We need a STARTSECTION for hidden content so it actually doesn't send it to the client.
And pharvey FTW - Foswiki:Development/AddHideOptionToSTARTSECTION
[02:00]
FoswikiBothttp://foswiki.org/Development/AddHideOptionToSTARTSECTION [ AddHideOptionToSTARTSECTION ] [02:03]
kip3fyeah, I'm a fan of that proposal. Looks like there really isn't consensus about naming though. [02:07]
SvenDowideitnot that important :)
i also want to be able to hide verbatim
but thats all 1.2
[02:11]
gac410class=foswikiHidden ? [02:11]
SvenDowideitmmm, if that works
dammit
why can't i type verbatim into wysiwyg and have it dwim!
tada
thankyou george
[02:12]
gac410(it's listed as a solution in the addHideOption... proposal [02:13]
kip3fI think that naming is important [02:15]
gac410I didn't like many of the suggestions. Some too wordy, some confusing. [02:16]
SvenDowideitimo, as verbatim class=foswikiHidden exists
we could claim that STARTSECTION{"" class=hidden} woudl not suck totally :)
STARTSECTION{"" class="tml"}
STARTSECTION{"" class="verbatim,tml"}
who knows :)
[02:16]
gac410Ah... what about startSection expand="include|inline" default is inline. that the section expands inline. expand="include" says only expand when included. [02:17]
SvenDowideitreplace expand with show/display [02:18]
gac410The prroblem with any css solution is that it still gets sent to browser. And actually %MACROS% in the section also get expanded. [02:18]
SvenDowideitSTARTSECTION{"" show="tml"} - show in a verbatim class=tml with chilli [02:18]
gac410So even if you hide it, if it has a SEARCH you will still do the possibly costly search. [02:18]
kip3fI think that I like show="off" [02:18]
SvenDowideitSTARTSECTION{"" show="hide"} - don't show inline [02:19]
gac410:p [02:19]
SvenDowideithide being not a css hide, correct [02:19]
gac410But does show=hide still expand the macros? [02:19]
SvenDowideitwhy? [02:19]
kip3fperformance [02:19]
SvenDowideitactually, yes
in that show=hide might use css
show=off would elide it from render
[02:19]
gac410I'm asking. How about render="inline|include" NO BAD BAD> Need to not render [02:20]
kip3fno teh point of the proposal is NOT to just use CSS [02:20]
SvenDowideitso that its possible to do either [02:20]
gac410right it is explicitly rejected. [02:20]
SvenDowideitthe point of the proposal for me, is to add functionality - so i can elide, hide, verbatim, chilli or superslice
all just with that param
elide being the most performant
[02:21]
kip3felide would be perfect, but it's really not a common word [02:21]
gac410I think pharvey's purpose was to prevent macros etc. from expanding multiple times if a section is transcluded in a topic. [02:21]
SvenDowideitshow="verbatim" being next
show="verbatim,tml" being next
[02:21]
kip3fhide="on" [02:21]
SvenDowideitoh, elide~~off :)
show="popup" could use jqDialog
show="twisty" could use TWISTIE
etc
mmm
[02:21]
kip3fI don't htink that STARTSECTION is the right macro for that type of stuff [02:23]
SvenDowideitdo we have a standard advice for hiding the ChangePassword topic?
or reset?
[02:23]
gac410Delete it? [02:24]
kip3fI'm going to suggest hide="on" [02:24]
SvenDowideittaht won't survive an upgrade
kip3f, boooo
that misses the useful extras
we also deperatly need a show=verbatim/tml
as quite a few of us define sections to be used elsewhere, and surround them in verbatim so they are self doccoing
[02:24]
gac410I don't understand why. We have <verbatim class="tml"> why do we need another syntax. [02:25]
SvenDowideitbecause
<verbatim class="tml">%STARTSECTION{}%
sux
and i write that several times a day
in function topics
[02:25]
gac410But STARTSECTION show=verbatim ... is that saying show when INCLUDED as verbatim, or show when inline as verbatim. I think it's horribly confusing. [02:26]
SvenDowideitok, good point
doesn't change the need, just that we need to make the name better
[02:26]
gac410yeah. it needs specific behavior when inline, and a separate defined behavior when included. [02:27]
SvenDowideitpersonally i thing show= is about how to display the section there were it is [02:27]
kip3fSven: how about a wrap="verbatim,tml" [02:27]
SvenDowideitand says nothing about where it is transcluded
but ... its true
kip3f, same issue, but also good :)
[02:27]
kip3fsorry, what is the issue with wrap? [02:28]
gac410to me the first thing that comes to mind is wordwrap / line wrap [02:29]
SvenDowideitand it says nothing about where its applied - inline / included
but as i say
i dont' thing the SECTION should declare how it is rendered when it is INCLUDED
[02:29]
kip3fI agree with that [02:29]
gac410(having just spent forever trying to figure out how convince vim not to auto-wrap pasted lines. :-( Finally renamed the file to .asdf [02:30]
SvenDowideitso i don't entirely agree with gac410's concern [02:30]
gac410As long as it's very clearly defined fine with me.
So if we say it's inline behavior only, then the options are ... is it rendered but styled (hidden, verbatim, tml ...) or is it not rendered (no macro expansion)
[02:30]
SvenDowideitsweet :)
those are 2 options
imo there could be more in future
for eg, not rendered and made ready for lazy loading as a defered ajax section
or, addtoheaded as json content (ala a new json data exchange standard the mozilla people are doing stuff with
[02:32]
gac410Need to be careful to separate things. to me, addtohead, or ajaxLoad would be totally separate options. [02:34]
SvenDowideitmaybe - if you can figure out your reason to fear :)
for eg
[02:35]
AlexisHazellHm. Why is there a viewprint.tmpl which is simply included from view.print.tmpl? [02:35]
SvenDowideitwhen i do the jqDialog stuff that the create task thing is done via [02:36]
gac410ie style is verbatm, tml, addtohead=true So it displays inline as an example, but is also added to the head as active code. [02:36]
SvenDowideitits pretty pointless to have the extra stuff
when i could just do
STARTSECTION{wrap=jqDialogLink}
and have it work
but that kind of heavy magic needs more careful thought that i have atm :)
[02:36]
AlexisHazell(i ask because i want to customise the output of pages that are to be printed - add an org logo at the top, remove footers.) [02:37]
SvenDowideitAlexisHazell, i think its legacy
the printing used to be done to viewprint, but now we just cover=print
(i think)
[02:38]
AlexisHazellHm. [02:39]
gac410best bet is to not change any existing templates btw. override them with your own skin templates. [02:41]
AlexisHazell*nod*
i was just trying to work out which template i needed to override.
[02:42]
gac410ah.
yes view.print.tmpl is the cover=print used for the printable views.
[02:42]
AlexisHazellSo if i created say [02:44]
pharveykip3f, SvenDowideit, gac410 - STARTSECTTION{display="hidden"} came after a lot of back-and-forth.. if you propose something else, please be mindful of the existing discussion. hide="on" is bad because it smells like negative-positive madness like nofinalnewlines="on" [02:44]
AlexisHazellview.myorgprint.tmp
tmpl
[02:44]
SvenDowideitpharvey, yup, i think we can go with that [02:44]
AlexisHazellNo, that's not right. [02:44]
SvenDowideitas gac410 points out, he just needs it to be clear that it only applies to the inline place the text is
not where its included
[02:44]
AlexisHazellOkay, so i already have a skin i'm using, call it 'myorg'. [02:45]
SvenDowideitand hidden/verbatim/ [02:45]
kip3fpharvey - I'm fine with display="hidden". I think that hide="on" is better, but I certainly don't want to stir up trouble
I don't see a problem with double-negatives
[02:45]
SvenDowideitAlexisHazell, great, you worked out why you have to make viewprint.myorg [02:45]
kip3fSTARTSECTION is shown by default [02:45]
SvenDowideitbecause its impossible for the skin to work with cover=print [02:45]
kip3ftake another look at the name of the proposal, hide is what you want to do! [02:46]
pharveydouble-negatives is a common criticism of Foswiki's macros; but I take your point. [02:46]
SvenDowideitwe are broken
kip3f, hide=on doesn't help the other use cases
[02:46]
gac410gac410 goes running screaming off into the sunset. gahh.... not no double negatives . never never. [02:46]
AlexisHazellSvenDowideit: Yes, but how do i get the system to use viewprint.myorg.tmpl? [02:46]
SvenDowideitwhereas display=hidden, display=verbatim, display=magic does [02:46]
pharveykip3f: I actually have no strong opinion, I just want anybody contributing to the discussion to realise that it took a lot of wrangling to get CDot & Micha to agree on something [02:46]
SvenDowideitAlexisHazell, by fixing our core :( i fear [02:46]
gac410AlexisHazell: no... core changes not needed [02:47]
kip3fSvenDowideit: perhaps you could propose adding a wrap attribute? [02:47]
SvenDowideiti think some 'clever' skin-ter was too clever for us [02:47]
gac410You created view.myorg.tmpl. So you can override the print buttons that say cover=print to cover=myorgprint [02:47]
SvenDowideitkip3f, i actually like display=
gac410, nice workaround
[02:47]
AlexisHazellgac410: Aaah, okay, that makes sense, thank you. :-) [02:48]
gac410so you will have view.myorg.tmpl and view.myorgprint.tmpl [02:48]
kip3fit's pretty good - the only problem is the possibility of confusion with CSS [02:48]
SvenDowideitkhnz_, yup, but if, as pharvey says, it was discussed to death already [02:48]
gac410For css I'd prefer that we use an existing keyword and behavior ... like style [02:49]
SvenDowideitthen i presume they covered that :) [02:49]
pharveythe CSS display conflict was discussed, yes [02:49]
SvenDowideit:) course, i bet i had my fiddycents too :) [02:50]
kip3fpharvey: yes I re-read the discussion. I originally read it a while ago, when I was lurking [02:50]
pharveySvenDowideit, kip3f - oops, it is definition="hidden" - sorry [02:51]
gac410bleh. [02:51]
pharvey%STARTSECTION{"blo" definition="hidden"}%
I actually started implementing this, but found myself refactoring the include code
[02:51]
SvenDowideitpharvey, thats woeful
SvenDowideit reconds kip3f's wrap= is infinitly better than 'definition'
[02:52]
pharveyCDot vehemently rejected 'display'
"display is not fine. Anyone would immediately assume that would suppress the display of the section anywhere it is used - exactly like it would mean in CSS. "
[02:53]
gac410definition= tells me nothing. I really disliked all of the discussed options. tbh wrap sounds so simple now.
geeze... except we don't want to wrap for pharveys use case. We want to not render it at all.
[02:54]
kip3fSven: let me clarify. I want two attributes. hide="on" for sections defined in the same page, to be hidden. wrap="magic" for cool stuff like wrapping with <verbatim class='tml'> [02:54]
pharveykip3f: please update the proposal to use wrap... this will be the third proposal which will be accepted for the same feature :) [02:54]
SvenDowideitkip3f, yes, and i violently object to using 2 params
as wrap=hidden makes sense for eliding too
hide=on is not intention revealing enough for me
[02:54]
gac410But its two totally different functions. wrapping it vs. n ot rendering it. [02:55]
SvenDowideitbarely
both are rendering instructions
[02:55]
pharveykip3f: if you feel like taking ownership of it, feel free to set yourself as committed developer. It is regrettable that SvenDowideit & yourself didn't chime in at the time it was re-proposed :P
we will need to reset the date of commitment
[02:56]
SvenDowideitits a shame, but its unreasonable to expect that everyone has time at just the right point [02:56]
pharveyI know, I know. [02:56]
SvenDowideitif were implemented, then we probly wouldn't care that much :) [02:56]
pharveyI just don't look forward to re-convincing everybody, again. [02:57]
SvenDowideit:) [02:57]
pharveyBut we wouldn't have that problem, if I just coded features instead of talking about them.
snap.
[02:57]
SvenDowideitgrin [02:57]
gac410(tbh ... it's an accepted proposal. ... those who are objecting should have objected way back when.
(and I'm counting myself as one of them)
[02:57]
SvenDowideiter gac410 thats insane
we don't insist that a passed propsal can't be improved
[02:58]
pharveyit hasn't been implemented yet, so I guess it's fair game. [02:58]
SvenDowideitwe allow the developer leeway to improve stuff, even after its merge to core [02:58]
kip3fSven - you have mentioned a few things that you would want to do with wrap. 1. verbatim class=tml. 2. jqDialogLink. 3. twisty. [02:58]
SvenDowideitotherwise we have a process that is shit.
yup
and i'll bet others will think of more in 6-10 months time
[02:58]
gac410yes. true. But at some point someone has to stop the analysis and actually implement. Or each new person that comes along can reset the timer again. [02:59]
kip3fso wrapping is somewhat open ended [02:59]
pharveyI'm a bit nervous about covering all those use-cases in this instance. Can we separate those concerns out somehow? [02:59]
SvenDowideity, pluggable
pharvey, why?
[02:59]
kip3fI agree with pharvey [02:59]
SvenDowideitseparating means we limit syntax ideas [02:59]
kip3fwrap is just different [02:59]
SvenDowideitbetter to talk about possibliltes
and then reduce scope of what the initial task is
gac410, no
[02:59]
pharveyYes, I like possibilities, but in the past when we overload a proposal with too many things it's more likely to be discussed to death, and this one has already had that twice over now [03:00]
SvenDowideitwhen a commited developer actually sits down to write a feature
they are not required to stick to the proposal
especially when they find a better way
[03:00]
kip3fwhy do you object to having two different attributes, wrap and hide? [03:00]
SvenDowideitwe are contributor based, not proacess based [03:00]
pharveyso if you think wrap="hide" is good enough to cover the requirement to 'hide', and isn't incompatible with the other use-cases you'd like, I'd prefer to keep it focused [03:00]
SvenDowideitkip3f, because they are btoh attr's for a rendering instruction
they are not 2 separate things
and we already have way too many macros with way too many attr's
[03:01]
pharveyI actually think definition="hidden" isn't that bad [03:01]
gac410SvenDowideit: I agree. I'm just saying that once the committed developer has been given the green light, the rest of the world has said their due . [03:01]
pharveyand this conversion we are having now, we already had exactly this fleshed out with CDot & Micha. [03:02]
SvenDowideitgac410, that basically says that once there's a greenlight, you stop accepting improvements
which is really naf
the final rule is pretty much entirely with the __implementor__
and atm, there is no-one
[03:02]
gac410No. You stop resetting the timer and re-hashing and re-analyzing. It's up to the developer to make that decision. Otherwise why have a timer at all. [03:02]
SvenDowideitprocess is there to help the implementor
no, we don't stop thinking
[03:02]
gac410gac410 thinks we are saying the same thing. [03:03]
SvenDowideitwe continue
we _want_ to encourge people to re-examine things
but that does not stop the implementor of an accepted proposal from ignoring / using the ideas
it really bothers me when people say that 'its an accepted proposal, you don't have the right to talk about changing it'
[03:03]
gac410totally agree. Just that once he/she has the AcceptedProposal, they are not mandated to keep re-analyzing re-hashing, [03:04]
SvenDowideitwe are all mandated to constantly re-examine everything
otherwise you may as well encrypt your source code and hide it
[03:05]
pharveyso, in summary. I went to implement Arthur's original proposal, and people chimed in that it could be improved. So I reached consensus with those participating at the time. If kip3f wants to do the implementation now, he is free to revise the spec yet again, as long as it goes through the feature proposal process. [03:06]
SvenDowideitwe want to figure out the best way to enable everyone's things, and at the same time allow people that want to implement something to haev a way to reduce their scope to reality
yup
well, ish
[03:06]
pharveyeven if kip3f doesn't want to do the implementation, I'd encourage him to please add a comment so we don't have this conversation all over again again again [03:06]
SvenDowideitif kip3f used the existing syntax, but did it differently/better [03:07]
pharveySvenDowideit's use-cases also. [03:07]
SvenDowideitthen he could also sidestep re-proposaling
ie, if definition="" were magically pluggable, there'd be no need - though thats up to the implementor's confidence in blagging it :)
[03:07]
gac410speaking of blagging. Need to blog about 1.1.5 Beta 2 [03:08]
SvenDowideitworse
i need to do 3 weekly newsletters and a monthly summary
guess i'll just have to admit i can't do weekly
[03:09]
kip3fhi I don't have a problem with definition="hidden". I think that hide="on" is better, but It's not a big deal. Furthermore, I think that the wrap ideas mentioned by Sven have merit but they should be in a separate attribute (and a separate proposal) [03:09]
SvenDowideitcan you elaborate why you want to have 2 separate rendering instructions?
in that both are really just the same as display=elide and dislpay=verbatim
[03:10]
gac410kip3f: the issue with "hide=on" is that it's single purpose. definition=" " is "open" to further uses. [03:11]
pharveyplease add that comment to the proposal. BTW my recollection is that hide="on" confirm-the-negative and its friends are to be avoided. [03:11]
SvenDowideitbut separate because display= is a little hairy?
i'd like to know what makes you _like_ hide= and wrap=
[03:11]
pharveySvenDowideit: IIRC there was concern that a 'display' attribute might imply to the user that it would affect how it looks when INCLUDEd [03:11]
SvenDowideitpharvey, yes, and imo that is crazy ;) [03:12]
pharveywhereas definition="foo" is more explicitly indicating that it will only affect the section where it's defined. [03:12]
SvenDowideitnot to me it doesn't :)
it makes me wonder - is that a label for the definition
or what?
[03:12]
pharveywell, to CSS users it would [03:12]
SvenDowideityes
as i said, display= has a reason to be out
[03:12]
kip3fYes. hide="on" is 'closed functionality' - it deliberately does only one thing. whereas wrap implies that the content is *still there*, but something special is wrapped around it. a verbaitm block, twisty, etc. [03:13]
SvenDowideitbut that imo is no reason to make the render instruction 2 attr's [03:13]
pharveyanyway, at the end of the day I really don't care, I just wanted consensus and that's what we got.. [03:13]
kip3fThere should be a way to add additional "wrappers" [03:13]
SvenDowideitkip3f, ok, so its due to english, and due to not using a better word that allows both
so if instead we had an acceptable synonym for 'display' and 'elide'
then we could use one attr to tell the SECTION how to render itself inline?
[03:13]
pharveywrap="delete"... ? Hehe :) [03:14]
SvenDowideitpharvey, y, though maybe less un-elucidating :)
ls
mmm, i think my drugs are wearing off
[03:15]
kip3fI'm not sure how "wrap wrappers" should be specified [03:16]
SvenDowideittbh, i try not to enter into some feature proposal discussions in the hope they get implemented fast
its not necessarily a wrapper
its just a render instruction
ie it could be just as plausible to say display=adjustheader
[03:16]
kip3fwhat does that mean? [03:18]
SvenDowideittake the content inside, and adjust the ---+++'s in some way [03:18]
pharveyI think there's a point that it's a worthy goal to try to pick something that highlights the point that it only applies to the definition, and not any usage of the section [03:18]
SvenDowideitup to the implementor of that rendering thing
imo there is no place for this attr to apply to the trancsclusion usage of the section
and so (for me) that is just a distraction from its purpose
i'm quite happy to say display= is out because somepeople are unsure what it applies to (for the same reason)
i don't grok this: i can buy a wifi router that does '300mb wifi' but it only has 100mb ethernet ports
[03:18]
kip3fI don't like the idea of including anything that can transform the contents of the section - that seems a bit much [03:22]
SvenDowideitkip3f, because...?
SvenDowideit doesn't agree, because the flexibility is useful to some things i actually do
it'd rock to be able to deplay=function and have it highlight the %INCLUDEVARS% and other things used -
more self doccoing INCLUDE function stuff
[03:22]
kip3fhmm yes, starting to grow on me [03:24]
pharveySvenDowideit: because who uses Ethernet? Isn't wireless like, totally faster than cables? CSMA/CD !!!! [03:24]
SvenDowideitwebscale!
SvenDowideit higs his gigabit network and looks forward to upgrading to 10gig
and yet i have >6 wifi devices here too
[03:25]
gac410:P CSMA/CA and rate scaling can become much slower. [03:26]
pharveyI had to explain to a CCNA (in my previous job) why, in fact, we should prefer cables [03:26]
SvenDowideitSvenDowideit knows nothing :)
mmm, let me see, er, point to point duplex, or shared 'BW', let me guess, um, um
[03:26]
gac410Everyone demanded switching, and cut-through switches, and zero latency packet forwarding. .... and then fell in love with WiFi. which is purely shared media. [03:27]
kip3fSven: I think that you are talking about an attribute called 'transform'. so we could have transform='hide' or transform='function' [03:27]
SvenDowideiti still remember using lots of coax
kip3f, y
just too many chars and too geeky :)
[03:27]
gac410Do you go back to thick-net ... and drilling the holes in the coax for the vampires [03:28]
SvenDowideitthankgod no [03:28]
kip3fok, now we're getting somewhere ... [03:28]
SvenDowideiti started networking in 1990
before that it was nullmodem all the way
[03:28]
pharveygac410: exactly [03:29]
kip3ftransform=
"hide" is fewer characters then definition="hidden"
[03:29]
SvenDowideitkip3f, ya - any 4 or 5 letter synonyms for transform?
>:}
[03:29]
kip3ffunc
ugh
[03:29]
SvenDowideitmmm, func=hide :/
show=tuna
(where tuna just replaces all uses of '%' with an image of a fish?
[03:30]
kip3fwell I checked an online thesaurus [03:31]
gac410And everyone wants their 11g or 11n, and then some clown walks down the hall, rate scales to 1 Mbps while streaming a video. and everyone waits for the slow packets. [03:31]
kip3fthey had 'cook' [03:31]
SvenDowideitlol
gac410, :p
[03:31]
kip3fmutate
alter
adjust
[03:31]
SvenDowideiti recon start with transform, though i hates it, and see what the peanuts think [03:34]
gac410I still can't figure out how definintion="hidden" ( send to client but hide ie display:none ) vs. definition="hidden" (hide it from the rendering pipeline) [03:34]
SvenDowideitgac410, yeah
imo hidden is to mean elide
as most times, when someone wants to hide something
they just don't want to see it
they don't actually care if thats using css or server side
[03:35]
gac410okay i looked up elide which was synonym for omit. ... I disagree [03:36]
kip3fso how would one add a new section transformer? A mechanism similar to include handlers? [03:36]
SvenDowideitare you sure that someone that wants to hide content actually cares how its hidden? [03:36]
gac410If it's a %SEARCH ... it may be *very* important to hide on the server. so that the search is actually not executed. [03:36]
AlexisHazellThis is probably a stupid question, but doesn't overriding a skintemplate def that's been set up in foswiki.tmpl simply involve redefining that def in one's own template? [03:36]
SvenDowideitgac410, thats what elide is
remove from render cue at begining
[03:37]
kip3fAlexisHazell: yes [03:37]
SvenDowideitAlexisHazell, y, in your own foswiki.myskin.tmpl
gac410, sorry - to me elide means remove from rendering before we do anything
[03:37]
gac410It's a new word to me. Online dictionary said it was a speech - omitting vowels [03:38]
SvenDowideitnot 'hide with css after we've done all the pointless hard work [03:38]
AlexisHazellSvenDowideit: Ah, that's what i'm doing wrong then .... i thought i could just include foswiki.tmpl, and then redefine certain defs that got included. [03:38]
SvenDowideitAlexisHazell, ah :/ more docco needed
gac410, cdot uses it when talking about replacing reprev
in that we save all rev's, but elide some
[03:38]
gac410AlexisHazell: Actually that is what you do. include a template, and then override it. [03:39]
AlexisHazellgac410: Well that's what i was doing. [03:39]
SvenDowideitand we'd do the eliding in the store, so the rest of the core thinks reprev is still as it is now [03:39]
AlexisHazellgac410: So i've included foswiki.tmpl
gac410: And then defined 'standardfooter' to be empty.
gac410: But the contents of 'standardfooter' as defined in foswiki.tmpl are getting displayed anyway.
Here's exactly what i've got:
http://pastebin.com/jdXwyxp8
[03:39]
kip3fand this is in foswiki.myskin.tmpl? [03:42]
AlexisHazellNo, that's view.myorgprint.tmpl [03:42]
kip3f1. create foswiki.myorgprint.tmpl 2. %TMPL:INCLUDE{"foswiki"}%%TMPL:DEF{"standardfooter"}%%TMPL:END%
3. profit!
[03:44]
AlexisHazellHm. [03:45]
gac410so if you view your topic with cover=myorgprint you still get the standardfooter? [03:45]
AlexisHazellgac410: Correct.
kip3f: That doesn't work.
From what SvenDowideit was saying, i'd need to create a copy of foswiki.tmpl - foswiki.myorg.tmpl - which i edit accordingly, then include that in the view.myorgprint.tmpl file.
[03:45]
SvenDowideity - so what kip3f says should work
mmm, or not
[03:48]
kip3fI've never seen a 'view' template include a 'foswiki' template [03:48]
AlexisHazellPresumably because foswiki is not a script? [03:48]
SvenDowideitok, too many things going on
foswiki.*.tmpl
is the base of everything
[03:48]
AlexisHazell*nod* [03:49]
SvenDowideitso view.*.tmpl includes foswiki.*.tmpl at the top
so, if you haev a skin my org
er
so, if you have a skin myorg
then foswiki.myorg.tmpl can contain anything that you want to change in foswiki.pattern.tmpl and foswiki.tmpl
_if_
you set Skin=myorg,pattern
and that way, cover=print would result in a temporary skin path of SKIN=print,myorg,pattern
where, as its calling view script
it would read view.print.tmpl, which should have a TMPLINCLUDE{view} at the top
so that include then would read view.myorg.tmpl
[03:49]
AlexisHazell*nod* So, in template A, i can't override definition b included from a template C by including a new definition for definition c in template A /after/ the include? [03:51]
SvenDowideitwhich should also have a TMPLINCLUDE{view} at the top
so that it reads view.pattern.tmpl
y
its basically macro/functional programming, and so just as confusing
[03:51]
AlexisHazellBut, if i remove the footer from foswiki.myorg.conf, doesn't that /also/ then remove it from /all/ views, not just the print cover?
i only want the footer removed in the print cover.
[03:53]
SvenDowideitdammit, good question - tis true
perhaps print is a context you can test for
its clear that cover=print is a bit of a problem
[03:54]
AlexisHazellWell, let me ask this: [03:55]
kip3fwhy not just add it to foswiki.print.tmpl? [03:55]
AlexisHazellWhy doesn't a definition after an include statement redefine any definitions resulting from that include statement? [03:55]
SvenDowideitAlexisHazell, it does [03:56]
AlexisHazellSvenDowideit: Well, that's not happening, is what i'm saying. [03:56]
SvenDowideitthat means one of us is misunderstanding :)
and i'm sick :(
[03:56]
kip3fAlexis: If you just want it to be removed from the print skin, then you can accomplish that by creating foswiki.print.tmpl [03:57]
AlexisHazellWell, here again is what i have in view.myorgprint.tmpl:
http://pastebin.com/e80VtdwM
[03:57]
SvenDowideitbut how is that tmpl supposed to be activated?
are you sure its being read and used at all?
[03:58]
AlexisHazellYes, i am.
When i access:
http://mydomain.com/foswiki/bin/view/Main/TopicPage?cover=myorgprint;
[03:58]
gac410:P this one is ugly
the standardfooter is both defined AND expanded directly in foswiki.tmpl
[03:59]
SvenDowideity, so needs to be over-ridden in foswiki.myorgprint
sorry AlexisHazell brain fart
[03:59]
gac410or the template that in turn expands standardfooter needs to be overridden. [04:00]
SvenDowideityou do mostly have to over-ride things in the same script.yours.tmpl tmpl file
otherwise you're toasted by other messeseses
[04:00]
AlexisHazellSo the file i currently have as view.myorgprint.tmpl should be renamed foswiki.myorgprint.tmpl? [04:00]
***EnriqueCadalso has left [04:01]
AlexisHazellOr do i leave that file as-is, and create foswiki.myorgprint.tmpl containing just the line "%TMPL:DEF{"standardfooter"}%%TMPL:END%"
?
[04:02]
gac410grepping for standardfooter, the slideshowplugin overrides it. But I don't see how It's expanded in TMPL "main" which is expanded in
foswiki.tmpl
gac410 gets totally lost in our template system.
[04:02]
SvenDowideitgac410, i think you've forgotten something that i think is true
we read the files
and then replace TMPL:INCLUDE's with txt
and only _then_ process the TMPL:DEF's
[04:04]
AlexisHazelli must say i can't understand why it's possible to override e.g. standardheader but not standardfooter.
Because the standardheader override i have in my view.myorgprint.tmpl seems to be working.
[04:04]
SvenDowideitAlexisHazell, me neither yet
its subtle i guess :/
[04:04]
AlexisHazellSvenDowideit: Ah, okay, it's not just me then. :-) [04:05]
SvenDowideitna, there are some frustratingly fascinating things [04:05]
AlexisHazellFirst , off the top of my head guess: [04:06]
SvenDowideitmm, ok
big thing:
you should avoid TMPL:INCLUDE{foswik}
as you break the tmpl INCLUDE SKIN path
[04:06]
AlexisHazellOh okay. [04:06]
SvenDowideityou _shoudl_ generally TMPL:INCLUDE{script} [04:06]
AlexisHazellSo i should be creating foswiki.myorg.tmpl in any case. [04:06]
SvenDowideitthat way your tmpl will include the next in the SKIN path all the way to the default view.tmpl
i don't think you need to
i don't think you /should/ need to
[04:07]
AlexisHazellWell, if i don't include it, things /really/ don't work. [04:07]
SvenDowideit'it' [04:07]
AlexisHazellNot just "Oh why is this footer still being included", it's "Why is nothing being displayed?" [04:07]
SvenDowideitwere you TMPL:INCLUDE{view} ing? [04:08]
kip3fOff to bed with me. But I started writing up some of the ideas in http://foswiki.org/Main/KipLublinerSectionalTransform [04:08]
SvenDowideitcos _that_ will, at the base if the nest of wurms TMPL:INCLUDE{foswiki} for you
kip3f, sleep well :)
[04:08]
kip3fbye [04:08]
AlexisHazellOh okay. [04:08]
gac410view.print.tmp does include of viewprint. viewprint.pattern.tmpl does include of viewprint.tmp. viewprint.tmpl includes foswiki.tmpl [04:09]
SvenDowideitSKIN path is implemented by the code 'magically' iterating INCLUDE's with Skin names [04:09]
AlexisHazellYes, i'd based my view.myorgprint.tmpl on viewprint.tmpl. [04:10]
gac410yeah. it is indeed magic. [04:10]
SvenDowideitit kinda has to be magical too [04:10]
AlexisHazellHowever, changing INCLUDE{'foswiki'} to INCLUDE{'view'} produces even less of what i want in terms of printable output.
Because doing so just recreates the standard view page.
[04:10]
gac410I modified viewprint, immediately after the tmpl include of foswiki.tmpl, I nullified out standardfooter And it still prints the standard footer. [04:11]
AlexisHazellgac410: Yes, that's what is happening to me. [04:11]
SvenDowideitit looks like i modified viewprint.pattern.tmpl to get a mod to the topicinfo
SvenDowideit needs to go rest and drug myself
getting hazyer :(
[04:11]
AlexisHazellFair enough!
Sorry to have brought all this mess up. :-(
[04:12]
SvenDowideitna, its important
i _think_ that its the cover=print and viewprint stuff that is the problem tho
[04:12]
gac410gac410 needs to sleep too. g'night all no worries AlexisHazell Templates are indeed a quagmire. and not well documented either. [04:12]
SvenDowideitsee if you can make it do your bidding in normal view mode
and then we'll see
sleep well gac410 :)
[04:13]
***gac410 has left [04:13]
AlexisHazellWell, i'll keep playing around - thanks all! [04:13]
..... (idle for 22mn)
Well here's some fun.
Deleting the standardfooter entry from foswiki.tmpl has no effect - the standard footer still gets displayed!
[04:35]
SvenDowideitAlexisHazell, that just means its defined elsewhere
and y, we have tons of redundancy in there, we clean it once in a while, but its argh, and full of misunderstandings
[04:46]
.... (idle for 16mn)
AlexisHazellSvenDowideit: Sorry, was on phone. - Elswhere in templates/, or elsewhere somewhere else entirely? :-_
:-)
[05:02]
SvenDowideityup
elsewhere in templates mostly
but you can define skins in topics too
but at the age pattern was done, we didn't do that as much
whereas arthur's latest work is in topics only
aka - lalalalalala, too much info :)
[05:02]
AlexisHazellWell, there's nowhere else in the templates folder that standardfooter gets defined as anything other than empty. [05:05]
SvenDowideitah, there's a neacky
i just remembered something
there's 2 footer's
standardfooter and >mumble<footer
simplefooter i think
and i have no idea atm, which is used when or why
thankfully, arthur has learnt a lot about coding since the heady days of pattern skin (like >7-8 years ago i think
[05:05]
AlexisHazellThere's no simplefooter defined anywhere in templates/ [05:07]
SvenDowideitsearch for footer - don't trust my memory :) [05:08]
AlexisHazell:-) Just did that,
no plain ol' "footer" - closest is NEWS:footer.
contentfooter is defined as empty in foswiki.tmpl
[05:08]
It wouldn't be anything to do with DBCachePlugin, would it? [05:14]
pharveyAlexisHazell: has anybody mentioned Foswiki::Template's TRACE mode to you [05:14]
FoswikiBothttp://trunk.foswiki.org/System/PerlDoc?module=Foswiki::Template [05:14]
AlexisHazellNo, but i found that myself - doesn't work.
At least not as advertised. :-)
[05:14]
pharveyhrm
well, it says what defs are being printed
[05:14]
AlexisHazelli enabled it, and yet only one def was mentioned.
In the output.
[05:15]
pharveyI agree it's a bit anemic, but has helped me guess in the right direction in the past
oh. Is NatSkinPlugin enabled by any chance?
[05:15]
AlexisHazellJust checked - no, not installed. [05:16]
pharveyit's just that it strips <!--comments-->
and you are checking view page source, rather than the DOM
[05:17]
AlexisHazellHm, okay. Just reenabled TRACE, let me look again.
Ah i see.
My mistake!
So, it seems bottombar is what i'm seeing.
And overriding that in my view.myorgprint.tmpl file worked!
[05:18]
SvenDowideitSvenDowideit calls pharvey a legend :p
i don't think i've used tempalte TRACE in years..
[05:22]
AlexisHazellThat was tremendously useful, now that i actually have a clue about how to use it. ;-) [05:23]
SvenDowideitSvenDowideit wonders if you can write a paragraph of docco on it
in er, whatever topic in System web where it should be mentioned
cos you only get to be a beginner once :(
[05:23]
AlexisHazelli definitely will!
Actually, will do so now.
Should i change it in my dev install, or on f.o?
[05:25]
SvenDowideitdev and commit :) [05:26]
AlexisHazellOkay. :-) [05:26]
SvenDowideiti'm pretty sure if you write it in Swahili, it'll be more useful than what isn't there now [05:26]
AlexisHazelllol [05:26]
SvenDowideitthis is my depressed face >:P
total waste of a day for me, too sick to concentrate :(
[05:27]
AlexisHazell*sad nod* i know that feeling all too well. :-( [05:29]
SvenDowideitcept in my case its an ephemeral thing that i know will pass
so i'm whineyer about it!
[05:30]
AlexisHazellHeh, fair enough. :-) [05:31]
SvenDowideitvery generous of you :P
i personally wouldn't put up with it! ban them i say!!
[05:32]
AlexisHazellWell, i'm feeling in a generous mood right now, what can i say. ;-) [05:33]
SvenDowideitgrin [05:33]
AlexisHazellDo i need to create a Task for this commit? [05:37]
SvenDowideitya, 'its the law' [05:38]
AlexisHazellHeh, okay. [05:38]
SvenDowideitaka, there's an svn pre-commit script that will reject it [05:38]
AlexisHazell*nod* [05:39]
pharveyfunny, I'm fixing a legend now
it says "stars" and "box" but they're all squares AFAICT
[05:42]
AlexisHazellSvenDowideit: Btw, i like the new Create Task dialog. :-) [05:43]
SvenDowideitpharvey, lol
AlexisHazell, :) do you get the whiney strikeone screen when you save the task from it?
[05:47]
AlexisHazellMm, not that i've noticed .... [05:48]
SvenDowideitjust to make sure - this is the one where you save directly from the popup
not the older one where the popup takes you to edit?
mmm, oh, it'd be the second time you use it
cos it looks like its cacheing the html including the strikeone numpty
[05:48]
AlexisHazellYes, saving directly from popup - but that is indeed the first time i've seen it. [05:49]
pharveyor, we abandon HTML and embrace the world of semantic-free JSON [05:51]
SvenDowideitpharvey, eh?
na, its actaully simple
fix jqDialog to obey {cache: off} like the docco says (and micha says works)
i havn't had time to work out why its caching when it shouldn't
so it could be my side (kinda unliekly, given that i'm getting it on ie8, ff11, chromium etc,but who knows)
[05:54]
pharveybrowsers seem to aggressively cache everything, even ajax requests
so I'm not sure what cache:off actually does
[05:57]
SvenDowideitah
in simplemodal
it actually refuses to cache
as both are implemented using jquery ajax stuff
and to do it, you change one small bit of code
cd bin
[05:57]
pharveyheh, according to the doc this is achieved by adding an "_=[TIMESTAMP]" query param
which is ?t=foo by another name, I guess.
[05:59]
SvenDowideityup
and that appears to fail too so far
which is _really_ surprising
[05:59]
pharveyindeed
so your firebug net panel confirms the same token keeps getting delivered?
[05:59]
SvenDowideiti'm resisting the urge to make this a release blocker
yes, it confirms that its served from the cache!
er, chromeium dev panel, but, like , whatever :)
[05:59]
pharveyI'm not sure that this is a Foswiki problem... do you see the _ param in the URL? [06:00]
SvenDowideitSvenDowideit isn't near the issue atm
and opening the dev panel causes browser to crash out of memory :)
[06:00]
pharveyoh, SvenDowideit
Default: true, false for dataType 'script' and 'jsonp'
on second thoughts, that probably means nothing
[06:01]
SvenDowideity [06:01]
pharveyI'll go back to fixing my own red herrings [06:01]
SvenDowideithttp://foswiki.org/System/JQueryUIDialog
<a href="http://...." class="jqDialogLink {cache:false}">Open Dialog</a>
but its not urgent today, it might be urgent before we release 1.1.5 to make sure its jsut my f**kup
[06:02]
GithubBot[foswiki] foswiki pushed 1 new commit to master: http://git.io/yedk1A
[foswiki/master] Item11668: Modified System/SkinTemplates 'Debugging' section. - AlexisHazell
[06:03]
***GithubBot has left [06:03]
FoswikiBothttp://foswiki.org/Tasks/Item11668 [ Item11668: Clarify how TRACE works when debugging SkinTemplates ] [06:03]
pharveySvenDowideit, one last thing... do you ahve the metadata plugin loaded ? [06:03]
SvenDowideitgod no [06:03]
pharveythen that's why [06:03]
SvenDowideiter, huh
this is foswiki core functionality fer gods sake
http://foswiki.org/System/JQueryUIDialog does not mention any non-core deps
argh! whitespace changes in all files in 1.1.5?
probly 'just' unix vs dos line feeds, but oh dear.
[06:03]
pharveySvenDowideit: it doesn't _depend_ on metadata.js, but your usage of it does [06:21]
...... (idle for 26mn)
SvenDowideit: quick, teach me CSS! [06:47]
SvenDowideitpharvey, eeeek
mmm
pharvey, should the merge cdot did delete all the foswikiimage and foswikibuttons in tinymce?
and foswiki/editor_plugin.js
[06:51]
pharveySvenDowideit: yeah, he moved them. There's a new foswiki_plugins_src JSON attribute in the default TINYMCEPLUGIN_INIT now, IIRC
hopefully the plugins still work in their new location (they did for me)
it basically allows us to finally add MCE plugins without having to plonk them into the pub/System/TinyMCEPlugin directory
[07:02]
........ (idle for 39mn)
Babarso, do you guys still need me for anything? [07:42]
pharveyBabar: are you taking wishlists? :D [07:47]
Babarnope
only production issues
[07:47]
pharveyI think things are fine [07:47]
Babarwe're way too far from Xmas, and the bunny costume doesn't fit me anymore [07:48]
pharveyhehe
firebug is so damn buggy
[07:48]
.... (idle for 15mn)
SvenDowideitBabar, we do
but its not nice :)
we need to know what to do about the error number that f.o's sendmail gives foswiki
but argh ):
and for your good looks - we always ..... yeah, ok :)
JQueryPlugin/jquery.noconflict.js was also deleted in 1.1.5?
and lib/Foswiki/Contrib/MailerContrib/Constants.pm ?
[08:04]
........ (idle for 37mn)
GithubBot[foswiki] foswiki pushed 3 new commits to master: http://git.io/t3gjSA
[foswiki/master] Item11670: initial checkin - MichaelDaum
[foswiki/master] Item11670: preparing release - MichaelDaum
[foswiki/master] Item11670: preparing release - MichaelDaum
[08:47]
***GithubBot has left [08:47]
FoswikiBothttp://foswiki.org/Tasks/Item11670 [ Item11670: externalize page optimizer from NatSkinPlugin ] [08:47]
pharveyThis is no good, this is not right. My CSS sticks out of its container all night. [08:58]
.................. (idle for 1h25mn)
***ChanServ sets mode: +o OliverKrueger [10:23]
............. (idle for 1h0mn)
kip3f has left [11:23]
............ (idle for 58mn)
padraig_lennonHey all.. Anyone here making use of textboxlist form types? [12:21]
OliverKruegerpadraig_lennon: occasionally. [12:22]
***OliverKrueger sets mode: +oooo Colas gac410 LarsEik MartinCleaver
OliverKrueger sets mode: +o padraig_lennon
[12:23]
padraig_lennonHi Oliver.. I'm having an issue with defining my own json lists.. I am able to use the lists fine in forms but I run into issues when the formfield has the H,M flags set in the DataForm definition. When I try to save the topic it complains that the field is empty where it is actually defined.. I must be doing something wrong.. If I post a list definition could someone have a quick look at it? [12:25]
Babarpadraig_lennon: you dare coming it after what you guys did on St Patty's day?! :)
(haven't seen the game yet btw)
[12:26]
padraig_lennonDreadful performance (by both France and Ireland!!)
Here is a typical list definition.. http://pastebin.com/dwXp9Yxw
I must be missing something..
Works fine when the field is not mandatory though.. Weird..
[12:26]
Babarvery [12:30]
padraig_lennonI use NatEditPlugin but I just tested and its the same using pattern
Anyone seen MichaelDaum?
[12:31]
BabarWikiRingBot: seen MichaelDaum [12:33]
WikiRingBotWikiRingBot has last seen MichaelDaum 17 Mar 2012 - 12:00 GMT on #foswiki
WikiRingBot has last seen MichaelDaum_ 15 Mar 2012 - 07:53 GMT on #foswiki
[12:33]
BabarFoswikiBot: seen MichaelDaum [12:33]
FoswikiBotBabar: MichaelDaum was last seen in #foswiki 2 days ago joining the channel. [12:33]
padraig_lennonThanks WikiRingBot!! :0
;)
I have a feeling its a bug with JQueryTextboxList - I will raise a task there - Unless anyone has it working correctly with the M flag set?
[12:34]
..... (idle for 20mn)
fsfsis it on purpose that CGIs push the paths in bin/LocalLib.cfg into @INC (via bin/setlib.cfg), but tools/mod_perl_startup.pl does not (ignores bin/LocalLib.cfg)? [12:55]
Babarhum... it shouldn't ignore it
because mod_perl will run the code from bin/* anyway
it will just be delayed until an action is performed
[12:59]
***ChanServ sets mode: +o MichaelDaum [13:00]
fsfshmm, that's true... but some things don't work reliably here, and I suspect they would if @INC was already modified in PerlPostConfigRequire... [13:02]
Babaryou can try. If you find something that seems to be more reliable, feel free to submit a patch :) [13:03]
fsfs(this is about a plugin requiring Perl modules in non-standard locations; doing 'use lib "..."' in the plugin doesn't work reliabley under mod_perl) [13:03]
Babaruse lib with a full path should work reliably, even under mod_perl [13:04]
OliverKruegerpadraig_lennon: sorry, was afk, reading irclog now. [13:05]
fsfsBabar: I was reading http://www.perlmonks.org/?node_id=615123, second last entry:
"That's because @INC is reset to the contents valid on Apache startup on each run. The use..bla statement is only executed the first time the script is run. On subsequent runs of the same script in the same Apache process (or it's children) the use statements are not executed again.
At least this is my understanding from what I read."
and that seems similar from what I'm seeing: 'use lib' sometimes works, more often doesn't (modules cannot be found)
[13:06]
gac410fsfs I wonder if you could also add the scripts to Foswiki config SafeEnvPath under securith
er.. security environment tab.
sorry add the paths, not the scripts
gac410 needs more coffee
never mind.
[13:07]
Babaroh, interesting [13:08]
gac410that's PATH not INC.
gac410 goes back to hibernate. clearly not enough coffee.
[13:08]
Babarfsfs: still, perlmonk says I'm right
the second-to-last comment is misleading and wrong, sorry :)
the third comment on that page has all your answers
the second to last comment is kind of useless. If he can harcode it in the startup script, he would hardcode it in the script, and it would just work
anywa
*y
[13:09]
fsfsBabar: but I could hard-code it, I use a plain string in the 'use lib' statement. It works as CGI, but it _mostly_ doesn't work under mod_perl [13:14]
Babarhum... I'd need to look at the code, but I think it's just the order of things which is wrong [13:15]
fsfsI'll do a bit more experimenting to see if pushing those paths to @INC during mod_perl_startup.pl would make a difference... [13:16]
Babarmeaning, if you: use lib "/somepath"; use MyModule; should work, but if MyModule also needs the path, it could die...
programming under mod_perl is tricky :(
[13:16]
............. (idle for 1h1mn)
gac410Howdy CDot [14:18]
CDotGeorge! Good to see you [14:18]
gac410I think we really need someone BSD knowledgeable to fix that email issue. It is really deceptive to report a mail failure during registration when the real issue is sendmail has queued the message. [14:19]
CDotthe obvious thing is to detect error code 17152 and ignore it!
CDot googled and binged but found nothing about it
CDot just found this: "It looks like the local mailer has the m mailer flag specified, which indicates that the local mailer can deliver to multiple users in one transaction. It probably can if there are no errors. If there are errors, there is no way to indicate that just one user had an error or which user that was. The solution is to remove the m mailer flag. You will have to consult the documentation to see how to do that. "
Another reason may be: http://www.linuxquestions.org/questions/linux-software-2/sendmail-procmail-ex_tempfail-506821/
[14:26]
gac410if os =~ /freebsd/i && error 17152, ignore ? [14:33]
CDotre: the first answer, more: "Another solution to this problem, perhaps slightly more efficient, is to
have sendmail talk LMTP to mail.local (available since sendmail 8.9) -
in .mc, before the MAILER() lines :
FEATURE(local_lmtp)"
CDot has no idea what or where .mc is
[14:34]
Babarit's in /etc/mail usually :)
# grep mtp /etc/mail/freebsd.mc
FEATURE(local_lmtp)
so it's already in :)
[14:35]
CDotok, so clearly that doesn't help :-(
thanks Babar
[14:36]
gac410and it's not just local destinations. gmail messages "fail" (but work) [14:37]
CDothttp://objectmix.com/sendmail/207350-mail-local-exited-ex_tempfail.html is what I was reading
and I have not seen the 75 error code myself; just going on what Sven reported in the code
(at least I think it was Sven)
where is the sendmail log?
[14:37]
Babarusually /var/log/maillog
on all SysV :)
any example / timestamp I could look at?
[14:38]
CDotBabar: check the foswiki.org apache error log
there are various FAILURE messages there, should have timestamps
[14:39]
Babaryes, I can see them. Digging
. ( $? << 8 )
what the heck is that?!
that's supposed to be a >>, right?
ERROR: Exit code 4390912 (17152) ... right. Could someone please patch that to be a >>? :)
I know, I could, but I'm digging the code now to find the bug
[14:41]
CDotok, it's a >>8 now [14:45]
***ChanServ sets mode: +o OliverKrueger [14:45]
Babarthx :) [14:46]
.......... (idle for 47mn)
GithubBot[foswiki] foswiki pushed 1 new commit to Release01x01: http://git.io/p8reHw
[foswiki/Release01x01] Item11675: correct boolean op and comment - CrawfordCurrie
[15:33]
***GithubBot has left [15:33]
FoswikiBothttp://foswiki.org/Tasks/Item11675 [ Item11675: Teach Sven how boolean ops work ] [15:33]
OliverKruegerOliverKrueger pings MichaelDaum [15:35]
MichaelDaumHi Oliver :) [15:35]
OliverKruegerHi Michael.
I saw your PageOptimizerPlugin.
[15:35]
MichaelDaumyep [15:36]
OliverKruegerI suggest that I deprecate my BrowserBoosterPlugin in favour of POP.
It supersedes the functionality.
[15:36]
MichaelDaummaybe...hm [15:36]
OliverKruegerWhen do you plan to release it? [15:36]
MichaelDaumI first tried to merge it into BBP but it turned out I had to rewrite most of it.
it is already out
[15:36]
OliverKruegerAh, in the meantime... :) [15:37]
***OliverKrueger sets mode: +oo CDot MartinCleaver [15:38]
MichaelDaumthere are still some loose ends, e.g. log css/js files used per topic for statistics. [15:38]
OliverKruegerto optimize disk space usage?! [15:38]
padraig_lennonMichaelDaum - I installed POP as well but it did not pick up my custom css. [15:38]
MichaelDaumand then add some means to group or exclude some css files from being mangled together
padraig_lennon, how did you add your custom css?
and: is it media="all" ?
[15:38]
padraig_lennonVia a custom skin.. Well JQueryMenuSkin.. I just installed the plugin.. I did not see where to set that media="all".. [15:41]
MichaelDaumcheck the html sources then [15:42]
padraig_lennonYup.. <style media='all'>
Let me dig deeper..
[15:42]
MichaelDaumtry _removing_ media="all" [15:46]
GithubBot[foswiki] foswiki pushed 1 new commit to master: http://git.io/WTx0bQ
[foswiki/master] Item11675: correct boolean op and comment - CrawfordCurrie
[15:46]
***GithubBot has left [15:46]
MichaelDaumthings aren't that "fault" tollerant yet [15:47]
padraig_lennonActually my custom skin is being picked up fine.. in the custom skin i import the base patternskin colors.css @import "../../System/PatternSkinTheme/colors.css"; Things like foswikiSubmit etc.. look off.. [15:48]
MichaelDaumany 404 file not found? [15:49]
padraig_lennonNot that I see.. I'm looking at the new css file.. There does not seem to be any color attribute for foswikiSubmit/Button etc.. Shouldn't there be? [15:53]
MichaelDaumtry this: disable POP; check where the missing styles are defined; enable POP again, this time with debugging switched on; see what happens then [15:55]
padraig_lennonWill do.. [15:55]
MichaelDaumPOP parses all css and follows @import statements importing these css files as well. maybe there's an error guessing the file path of that secondary css. [15:56]
padraig_lennonfoswikiSubmit - When POP disabled defined in http://k-portal/foswiki/pub/System/PatternSkinTheme/colors.css ... When POP enabled they are missing.. [15:57]
MichaelDaumwhere are they @import-ed? [15:58]
padraig_lennonThey are imported in JQueryMenuSkinBar/colors.css - I have set the pref variable FOSWIKI_COLORS_URL to this file in SitePreferences.. [15:59]
MichaelDaumhow does the @import statement look like?
seems it hasn't been parsed correctly
[16:00]
padraig_lennon@import "../../System/PatternSkinTheme/colors.css" When I disable FOSWIKI_COLORS_URL it works but I lose my customized JQueryMenuSkin colors.. [16:01]
MichaelDaumtry an absolute url path
debug output should show how these urls are processed ...
[16:01]
padraig_lennonNope.. That makes a mess of the menu.. absolute url should be http://k-portal/foswiki/pub/System/PatternSkinTheme/colors.css right? [16:04]
MichaelDaumy [16:04]
padraig_lennonNo doesn't work.. [16:05]
MichaelDaumah
this @import statement is invalid
must be @import url(...
and?
[16:06]
padraig_lennonYup.. That fixed it.. Correct syntax is @import url('../../System/PatternSkinTheme/colors.css'); needs single quotes.. Thanks Michael..
MichaelDaum - Up for another challenge?
[16:11]
MichaelDaumnaw ;)
MichaelDaum up for nother mug of coffee
[16:13]
gac410Babar: CDot: Documented the Sendmail issue in http://foswiki.org/Support/Faq69#Common_61sendmail_61_issues [16:14]
padraig_lennon:( ... When you get a chance (Whenever..) I opened this this morning.. Good idea about the coffee..
http://foswiki.org/Tasks/Item11674
[16:14]
MichaelDaumdunno whats wrong with mandatory textboxlist formfields... [16:15]
padraig_lennonSo its known? [16:15]
MichaelDaummaybe there's some AI saying: "why should an item be part of more than one wiki app"
and "why should it be user-editable" ;)
[16:15]
CDotmaybe there's some AI whispering "MichaelDaum, why don't you fix me, instead of complaining" ;-) [16:16]
MichaelDaumMichaelDaum trying to avoid mandatory formfields whereever possible ... that's why this combi never showed up 4 me [16:16]
CDotgoddam, why is this BugsContrib wikiapp so damned inside-out? [16:17]
MichaelDaumCDot, oh I am happy. zero complains over here ... umpf ;) [16:17]
CDotshit. OK. So here's the problem. When you define a list-type field, such as a =select=, in a foswiki form, then it loads the topic correspoding to the field name to get the values for the field
it does this on view as well as edit
when it loads that topic, it does an expandMacros on it to try and perform any SEARCH to build the table
then it loads the first table it finds in the included topic.
[16:18]
gac410Why does it need to do all that on view. Isn't the list of valid values only used during edit? [16:20]
MichaelDaumeven for a view? wow. [16:20]
CDotOK so far? In the BugsContrib, the AppliesTo topic is a field name topic, but also has a number of other active sections
yes, I was surprised too, I didn't think it did that. But it does.
Anyway, the other parts of the topic get expanded *before* the table gets parsed out (makes sense, really)
[16:20]
MichaelDaumbtw guys. gr8 work on AntiWikiSpam. excellent to see things been picked up. [16:21]
gac410I suspect that could improve performance then. As would a STARTSECTION definition="hidden" so it doesn't get expanded. [16:21]
CDotunfortunately, another part of the AppliesTo topic has a %INCLUDE..... oooh, could you be more inefficient? I think not. [16:21]
gac410MichaelDaum: threshold of pain got too high I guess. [16:21]
CDotquestion is, why does AppliesTo have all these extra sections? Can they be moved? [16:21]
MichaelDaumit really frightened me seeing this many tweets matching "foswiki" ... lots of them pointing to spammed foswikies [16:22]
CDotwell, we can block new spammers, but can't cure already-spammed wikis [16:23]
MichaelDaumthat's why we should advertise the new 1.1.5 + AntiWikiSpamPlugin + SafeWikiPlugin as a major improvements for public foswikies [16:24]
gac410Well SafeWiki is still got issues. it breaks f.o esp. the Home/WebHome landing page. [16:24]
MichaelDaumBlackListPlugin isn't really an option for high traffic sites
it does not protect against DoS, even though it says so
[16:25]
gac410When I was using it, I would typically DOS myself [16:26]
MichaelDaumy [16:28]
gac410CDot: http://foswiki.org/Home/WebHome?debugenableplugins=AutoViewTemplatePlugin,CommentPlugin,CompareRevisionsAddOnPlugin,EditTablePlugin,HistoryPlugin,InterwikiPlugin,JQueryPlugin,PreferencesPlugin,RenderListPlugin,SlideShowPlugin,SmiliesPlugin,SpreadSheetPlugin,TablePlugin,TinyMCEPlugin,TwistyPlugin,WysiwygPlugin,SafeWikiPlugin [16:29]
MichaelDaumit puts a foswiki-fcgi to sleep discovering bad actions...so blocking all of the process serving good requests
we told pth a decade ago. he did not listen.
[16:29]
CDotgac410: what, you want me to click a really dodgy-looking link like that? You think I'm mad? ;-) [16:30]
MichaelDaumbest would be to maintain some .htaccess file's deny entry ... for apache setups [16:30]
gac410iirc it also would ban every buildcontrib upload for a while ... resulting in the excessive sleep time beetween successive uploads.
CDo - :P It's a way to view webhome will SWP enabled. But have to list all plugins.
we needs a "debugenableplugins=default,SafeWikiPlugin
[16:30]
CDotCDot considered using AWSPI to inspect the IP of the source of the registration request, and auto-add it to a "suspects" list
y, thanks, I realise what it is; I was just joking. Still trying to de-%CLASS% the BugsContrib
[16:31]
gac410I was thinking our rejected registrations should be logged to the events log, not errors. So it gets IP logged, and fail2ban could process and add IP to firewall. [16:32]
GithubBot[foswiki] foswiki pushed 1 new commit to master: http://git.io/hzi1cw
[foswiki/master] Item11676: Fix incorrect import statement in menu.css - PadraigLennon
[16:32]
***GithubBot has left [16:32]
FoswikiBothttp://foswiki.org/Tasks/Item11676 [ Item11676: Correct css import on JQueryMenuSkin ] [16:32]
gac410CDot: okay. I figured you would know. [16:32]
CDotI can't work out WTF all these sections of AppliesTo are meant to do :-( [16:36]
gac410I don't even find where the form_select section is referenced. [16:41]
CDotI don't think it is. I think it's purely internal to that topic
CDot has split off Tasks.AppliesToParts to try to simplify
[16:42]
gac410And it appears I cloned AppliesTo to AppliesToParts ... but I have no idea why or what for. [16:42]
CDotno, i cloned it
CDot is masquerading as gac410 >:-)
[16:42]
gac410Ah.
:p
[16:43]
GithubBot[foswiki] foswiki pushed 1 new commit to master: http://git.io/LnjItQ
[foswiki/master] Item11676: Bump release num for new package - PadraigLennon
[16:45]
***GithubBot has left [16:45]
foswiki_irc0we have a small pilot foswiki site with 200 or so topics, and are thinking of switching from CamelCase to pure forced links [16:45]
gac410AppliesTo and AppliesToParts are identical [16:45]
foswiki_irc0Are there any tools to automate the conversion of existing CamelCase links? [16:45]
CDotfoswiki_irc0: GlobalReplacePlugin?
sed? awk? perl?
[16:46]
foswiki_irc0CDot: thanks. GlobalReplacePlugin looks promissing.
Are there any issues with editing the txt files in data directly with respect to versioning?
is a save cycle in the wiki required?
[16:50]
CDotgac410: zowie; try viewing a topic in Tasks now.....
foswiki_irc0: at the very least you should ensure all .txt,v files are touched after any mass change of .txt files
unless, of course, you *want* your modification to appear as a "real" revision
if you do, then you would probably be best to use rcs to check in your changes
see lib/Foswiki.spec {RCS} options for a list of the commands Foswiki uses to do checkins
[16:51]
foswiki_irc0CDot: thanks again. If that's the only risk, then I'm not too worried. I don't know the internals of versioning at all, and didn't want to break it. [16:54]
gac410Wow CDot ... it does feel faster [16:55]
CDotit *is* faster!
visit AppliesToParts for a feel for what it was doing *on every page view in Tasks web*
[16:55]
gac410This is like the SubscribePlugin disable improvement [16:59]
Now I need to get my changes to SubscribePlugin released so I can install on f.o and exclude the Tasks web without all the "disabled" messages
So still does it make sense for the select values to be searched on every view, or is there more room for optimization there?
[17:05]
CDotit makes sense
the problem was the *extra* searches it was performing, that played no part in the view
[17:20]
gac410I Uploaded and installed new SubscribePlugin on f.o. What webs do we want subscribe links on. So far I've set Blog,Community,Development,Extensions,Main,Sandbox,Support
Bleh... So with SubscribePlugin enabled, but set to exclude the tasks web, it still renders a null button:  Attach |  | Print version 
But the message | 2012-03-19T17:25:00Z warning | SubscribePlugin has been disabled | is eliminated
[17:22]
***foswiki_irc0 has left [17:30]
gac410phew. the error log finally seems quiet
CDot ... maybe that side effect of form field definition topics loading during view should get captured in a Performance topic somewhere.
[17:33]
CDotI was wondering about that. I should really be part of the system documentation. [17:35]
gac410btw - did you understand where I was coming from on the AWSPI register patch. Just a timing issue. Want f.o to run as close to possible as what will be in the RC I build tomorrow.
(unless you were planning on checking in the Register patch for the RC ... I suppose )
[17:40]
CDotyeah, sure, no problemo; I just wanted to point out why I'd done the hack, is all. [17:44]
gac410right I did understand it. And the overhead to remove a registered user is pretty large. So it's a good hack.
The subject "AWSPI even *more* enabled" threw me - I though you had patched it again hours after the beta2 upgrade. I didn't recognize it as a continuation of the old thread.
The fix for the recursive error will probably have *big* benefits when the bots are indexing tasks ... we do really slow down at times.
gac410 wonders. Should Foswiki:Development/ImplementAddMeToAdminButton really be a generic "AddMe" button rendered on any group page where the current user has change authority.
[17:45]
FoswikiBothttp://foswiki.org/Development/ImplementAddMeToAdminButton [ ImplementAddMeToAdminButton ] [17:51]
gac410Or "Join this group" / "Leave this group" button. Easier than clicking edit groups, find your id, uncheck, and submit. [17:54]
....... (idle for 31mn)
CDotgac410: there are REST handlers for addtogroup/leave group, so it's a wikiapp question. [18:25]
gac410right. Just wondering if it makes sense to put them in the Group view template by default, rather than shipping a "MakeMeAdmin" helper topic.
(or in addition to)
[18:26]
CDotoic, Yeah, makes sense [18:28]
................................................. (idle for 4h3mn)
AlexisHazellIs anyone else getting an Oops exception on AccessControlTests::test_setInSetAndMETA? [22:31]
gac410trunk or release11 [22:31]
AlexisHazelltrunk [22:31]
gac410running them now.
works for me.
[22:32]
AlexisHazellAck.
i added one line to Foswiki.pm, in spaceOutWikiWord
tried running the tests and got that.
Removed the line.
Still got that. :-/
[22:34]
gac410git status -uno ??? [22:34]
AlexisHazelli've been getting this for last 24 hours, even after git svn rebases - did one just now, tried running tests, no luck. [22:35]
gac410and status -uno shows nothing modified? [22:36]
AlexisHazellgit status -uno reports:
"# On branch master
# Your branch is ahead of 'origin/master' by 13 commits.
#
nothing to commit (use -u to show untracked files)"
[22:36]
gac410Ah ... so you have a lot of other commits that you have not committed back with "git svn dcommit"
git bisect is your friend.
[22:36]
AlexisHazellHm, thought i'd commit'd everything that needed commiting.
*committing
[22:37]
gac410Commit saves to the git tree. svn dcommit sends them back to svn. [22:38]
AlexisHazellWell, yes, but i meant both.
There's nothing i'm aware of that needs to be dcommit'd.
Anyway, bisecting now.
[22:38]
gac410Otherwise I'm not sure wht "ahead of origin/master" [22:39]
AlexisHazellHm, what label do i use to pass to git bisect good? [22:41]
gac410do a checkout of a commit you know works. (git log and just go back a bunch - not important which. [22:42]
AlexisHazellBut, using the git commit ID? [22:42]
gac410yeah. the hash. (or abbreviated hash) [22:43]
AlexisHazell*nod* Thanks. [22:43]
gac410git log --oneline gets you a quick list.
Something is not quite rithe with your branch being "ahead" . I normally see # On branch master ... nothing to commit
[22:43]
AlexisHazellYes indeed .... i'm going quite a way back now, still getting the same failure.
(Last checkout: 63c9a99)
[22:46]
gac410hm. strange I don't see that hash in my log. [22:48]
AlexisHazell"Item11317: perltidy" [22:49]
FoswikiBothttp://foswiki.org/Tasks/Item11317 [ Item11317: Allow an Include to increase/decrease header levels ] [22:49]
gac410The other possibility is something in your LocalSite.cfg or a topic issue. For example some of the tests will fail if a topic has unexpected history files, The SemiAuto tests are notorious for that. [22:50]
AlexisHazellHm. [22:50]
gac410What line is failing in your tests? [22:50]
AlexisHazell"Foswiki::UI::Register::_validateRegistration('Foswiki=HASH(0xad110f0)',
'HASH(0xad7dfac)', 1) called at /var/www/foswiki/core/lib/Foswiki/UI/Register.pm
line 413"
[22:51]
FoswikiBothttp://trunk.foswiki.org/System/PerlDoc?module=Foswiki::UI::Register [22:51]
AlexisHazellresulting in:
"OopsException(attention/already_exists web=>TemporaryAccessControlUsersWeb topic
=>UserRegistration params=>[ScumBag])Error at /home/alexis/src/perl5/foswiki/cor
e/lib/CPAN/lib/Error.pm line 148"
[22:53]
gac410did you make changes to the validate topic code? [22:53]
AlexisHazellThe only code i touched was in Foswiki.pm, to spaceOutWikiWord .... one line, which i then removed. [22:54]
gac410Are yoiu running unit tests witht the "-clean" option/ [22:54]
AlexisHazellNo .... where in the line do i add that option?
between TestRunner.pl and FoswikiSuite.pm?
[22:54]
gac410../bin/TestRunner.pl -clean BlahTests [22:55]
AlexisHazellJust did that, same error. [22:55]
gac410check in your data directory. Are there any Test* webs that you don't expect / didn't create? [22:56]
AlexisHazellThere's TemporaryAccessControlTestWebAccessControl, TemporaryAccessControlUsersWeb, TestCases [22:56]
gac410yeah.. dump them and try it again.
Not test cases
It's the TemporaryAccessControlUsersWWeb that's the issue.
[22:57]
AlexisHazellOkay, did that, same failure.
Oh no, wait, different failure.
"OopsException(attention/problem_adding web=>TemporaryAccessControlUsersWeb topic
=>UserRegistration params=>[ScumBag,~/src/perl5/foswiki/core/working/
htpasswd.lock open or create password lock file failed -check access rights: Per
mission denied at /var/www/foswiki/core/lib/Foswiki/Users/HtPasswdUser.pm line 2
09."
Do i need to remove a stale htpasswd.lock file?
[22:57]
gac410Okay - looks like you have a permissions problem on you rpassword file. [22:58]
AlexisHazellHm, odd. [22:59]
gac410Are you running as the same user as your Apache cgI ? some sort of ownership issue I expect. [23:00]
AlexisHazellNo, running as myself. [23:01]
gac410and Apache hasn't owned some files or directories on you? [23:01]
AlexisHazellThought the docs implied i shouldn't be running as the web server, but maybe i misread ....
Still, i'm a member of the group of the data/.htpasswd file.
Which has perms 664.
[23:01]
gac410heya pharvey - AlexisHazell has an issue - her repo says # Your branch is ahead of 'origin/master' by 13 commits. I've never seen that.
AlexisHazell: I suspect everyone does that a bit differently. I set up ownership using my personal user/group. And use Apache SuExecUserGroup directoive to have apache use my ID for cgi ownership.
[23:03]
AlexisHazell*nod*
Currently reading through http://stackoverflow.com/questions/277077/why-is-git-telling-me-your-branch-is-ahead-of-origin-master-by-11-commits-a
[23:05]
gac410Ah. you are using github, right? Probably need to pull.
or whatever the command is.
fetch. something. (I'm not all that knowledgable)
[23:06]
AlexisHazelli know that git fetch doesn't automatically do a merge ....
Well, i'll pastebin git status as it stands right now.
http://pastebin.com/fYYS3mfc
[23:07]
gac410Ah... you are not on any branch right now. you need to end your bisect and check your branch back out again ... master or whateer you were using.
Add this to your .bashrc
# Put Git branch name it shell prompt
parse_git_branch() {
git branch 2> /dev/null | sed -e '/^[^*]/d' -e 's/* \(.*\)/(\1)/'
}
gitbranch="\$(parse_git_branch)"
PS1="\[\033[01;32m\]\u@\h\[\033[00m\]: \[\033[01;34m\]\w\[\033[00m\] $gitbranch\$ "
(I should have pastebin'd that .. sorry)
[23:10]
AlexisHazellHm, i use zsh - does that have the same prompt syntax? [23:11]
gac410no idea. [23:11]
AlexisHazellOkay, will come back to that shortly. :-)
Done a git bisect reset
git status reports the same thing i already pastebin'd, except i'm now "On branch master".
[23:11]
gac410It doesn't say you are 13 commts ahead any more? [23:12]
AlexisHazellIt does.
So, i should try a git pull?
(which does fetch-and-merge)
[23:13]
gac410TBH I have absolutely no idea [23:13]
AlexisHazellOkay, i'm going to try it, on a hiding to nothing here i reckon!
That reported:
"Already up-to-date".
i just tried su'ing to the www-data user and running the tests - same failure.
No wait, different failure again. :-P
Hm.
Is there any reason i can't just completely wipe the working/tmp folder?
[23:13]
gac410no - except you'll loose web sessions. [23:20]
AlexisHazell*nod* "Acceptable losses". ;-)
.... and now the tests are running, from www-user.
gac410: Thank you for your help, and your patience. :-)
[23:20]
gac410glad to help. The learning curve is a bit steep. [23:22]
AlexisHazellStill not sure about the "13 commits ahead" thing of course.
Well, i should be recording the 'fun' i've been having with permissions, adding notes to the Development web on f.o.
[23:22]
gac410That link you pointed out had some reasonable descriptions. [23:22]
AlexisHazellSorry, which one? [23:23]
gac410since with git svn I'm not using a remote git repo, I'd never see the issue. [23:23]
AlexisHazellAh okay. [23:23]
gac410the stackoverflow one. [23:23]
pharveyAlexisHazell, gac410 - you will get the "ahead of master" because git thinks your master branch is tracking github.com/foswiki master, but in reality you will normally use git svn rebase to keep in sync with foswiki [23:30]
AlexisHazell*nod* [23:30]
pharveyit's not an error, just informative. A git fetch should make that message go away [23:30]
AlexisHazellpharvey: Implying a git pull should as well? [23:31]
pharveydon't git pull on any branch you plant to git svn dcommit with
svn doesn't cope with git merges very well
[23:31]
AlexisHazellOh okay. [23:32]
pharveyplant? plan
that's annoying fact of life until we fully migrate to git
if you prefer to use github.com instead of svn rebase, sometimes I do:
git checkout <something else>
git branch -D master
git checkout master
and that automatically creates a local master branch tracking origin/master
without using any merge commits which confuse svn
[23:33]
AlexisHazellHm, okay. - Might leave doing that until i understand things a bit more (translation: until i've broken my dev setup a bit more :-) . [23:35]
pharveyhehe [23:35]
gac410BTW changing how SpaceOutWikiWord works ... will probably cause lots of test errors in places.
That one might rise to level of neading a feature proposal.
[23:36]
AlexisHazellgac410: i suspected it might, which is why i was determined to make sure i could get the unit tests working properly. :-) [23:37]

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