#foswiki 2012-02-29,Wed

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

WhoWhatWhen
SvenDowideitpharvey, i really don't care about perfect, if we have a plugin, we can try it until we have a better one [00:00]
gac410SvenDowideit: well if things go okay tonight/tomorrow we can blog about 1.1.5beta being on f.o hopefully. Drats I forgot to reduce the mod_expires time. But not many js/css changes I don't think. [00:00]
SvenDowideitargh!
root@quiet:/home/sven/src/foswiki/core/bin# apt-get install liblocale-maketext-lexicon-perl
Reading package lists... Done
Building dependency tree
Reading state information... Done
liblocale-maketext-lexicon-perl is already the newest version.
gac410, most excellent :)
[00:00]
pharveyand by "scratchings" I mean, I spent a few hours reading doc and modifying a hello-world syntax highlighter def... so there's almost zero value to my checkout right now [00:01]
SvenDowideitpharvey, the value is in the 'might prompt someone else to scratch for an hour more' [00:01]
gac410Ooh... firing squad for gac410. Never increased the mod_expires after 1.1.4 [00:02]
SvenDowideithehe
ok, so i don't get it at all
I have 2 servers building trunk, and one rel1.1
but i get one email for trunk and 2 for rel1.1
SvenDowideit thinks his world needs new servers :(
Babar, do you have a quick answer (and over simplified) for what i should do about getting a NAS for my important data and vm's?
[00:03]
Babardepends of budget
drobo or sonologic, I'd say
[00:05]
SvenDowideiti'm wavering between an enterpricey-server+with raid from ebay (5 years old), build my own pathetic box, or sonologic
when i looked at drobo it sounded like people got less throughput
and thats one of my problems
though the biggest problem would be that i have 15+ years of FS mess
[00:05]
Babarproblem with enterprise servers is that it needs pricey old disks [00:06]
SvenDowideitmy main server still has stuff from my first server - an IPX from 1996
ish - i was going to go with one that has a SAS/SATA backplane
like the dell 2950's with 3.5 inch disks
but the noise at home worries me
as does my lack of time and experience with that kind of HW
[00:06]
gac410gac410 wonders why Lavr's TML Release Notes query pulls all fixes that are not closed. Shouldn't it be waiting for release? [00:08]
SvenDowideitpharvey, you're always doing not-easy stuff, which is why i'm hopeing to get you to write a chapter on non-easy stuff :p
probly to prompt him to check them too
[00:08]
BabarBabar has... hum... lemme check... I'd say around 1000 2950's at work [00:09]
SvenDowideityup :) [00:09]
Babarand had to exchanged all of their perc5 controller, because the condos on them were crap [00:09]
SvenDowideiterkle! [00:10]
gac410yeah. though I now have TasksByRelease. [00:10]
Babarand yes, it's pretty noisy [00:10]
SvenDowideithttp://elvery.net/demo/responsive-tables/
i like the no-more tables option :)
[00:10]
Babarsome co-workers had drobo/sonologic. But I'm afraid they're on holiday now, so I can't tell you right away [00:11]
SvenDowideithehe - many thanks tho, all input is useful
ok
so i run the trunk unit tests
and there are no missing DEPs listd in the DEPS file
just optionals
and the optinoals are SASL, Encode::compat, Lingua, Win32 and Sleenium
SvenDowideit runs for the hills
[00:11]
...... (idle for 27mn)
Babarthey're not required
but.... they're required for some features, so for some unit tests
yes, in theory, it would be better to make the unit test skip
[00:40]
SvenDowideitno, i'm not curious why these are not non-optional deps for UnitTestContrib
and thus would be a more sane error
let alone - i need one of these for Fn_MAKETEXT::test_underscore ?
[00:43]
Babaryes, the Lexicon one
but in fact, you need it for everything
it's just that instead of doing anything, MAKETEXT will just return the string
and test_underscore tests that it throws an error
[00:45]
SvenDowideitok, so if i 'use Lexicon' at the top of that test [00:46]
Babarwhich it's not throwing as it's not enabled [00:46]
SvenDowideitit'll fail more obviously?
er - which Lexicon one?
there's no Lexicon in the list i gave you
[00:46]
BabarI guess we should eval "use Maketext::Lexicon::..." or $this->skip, or whatever [00:47]
SvenDowideitSvenDowideit mumbles about being ignorant [00:47]
Babarcheck my mail [00:47]
SvenDowideitno, i'd like BOOM!!
that one _is_ installed
[00:47]
Babarwhich module was that? [00:47]
SvenDowideitLocale::Maketext::Lexicon [00:47]
Babaryes, that one
it's installed on the build server?!
in the right perl?
[00:47]
SvenDowideiti use only one perl
foswiki@hp2:~/trunk/trunk/core/test/unit$ apt-cache search Locale::Maketext::Lexicon
liblocale-maketext-lexicon-perl - lexicon-handling backends for Locale::Maketext
foswiki@hp2:~/trunk/trunk/core/test/unit$ su
Password:
root@hp2:/home/foswiki/trunk/trunk/core/test/unit# apt-get install liblocale-maketext-lexicon-perl
Reading package lists... Done
Building dependency tree
Reading state information... Done
liblocale-maketext-lexicon-perl is already the newest version.
just to make sure that i'm not miss-seeing
so either its something else, or its something else, and either way I don't have enough info from the test to tell me
[00:48]
***SvenDowideit sets mode: +oooo AndreU gac410 pharvey Rich_Morin
SvenDowideit sets mode: +o terceiro
[00:52]
pharveySvenDowideit, Babar: I made most tests which depend on optional things, skip already - if I missed something let me know - there's a skip summary at the end with a reason for why test was skipped [00:56]
Babarpharvey: I just said what needs skipping and why
but ok, so it's something else
crap
Babar had the same test error on OS-X, and the reason was that I didn't have this module
[00:57]
SvenDowideiti'm not sure it should be skipped!
else devs will break stuff, and see zero failures
when all thats happened is they don't have an important module
[00:57]
Babarok, then is should be made more obvious [00:58]
SvenDowideitwould adding a use Locale::Maketext::Lexicon(); make it in your face? [00:58]
BabarBabar goes back to being sherlock holmes [00:58]
pharveySvenDowideit: if you look at the tests skipped summary, it says why... so it's not silent failure [00:59]
SvenDowideitpharvey, it is siletn [00:59]
pharveybut I guess we don't have an intelligent skip for MAKETEXT [00:59]
gac410hm. any last minute commits before I build 1.1.5 Beta 1? [00:59]
SvenDowideitif it says zero fails then most people will go 'ok, it works'
gac410, not from me
[00:59]
Babarnot from me either
or...
[01:00]
pharveyso, are we saying that Lexicon is a hard dependency? or that skip summary isn't loud enough? [01:00]
BabarItem11224
Item11506
[01:00]
FoswikiBothttp://foswiki.org/Tasks/Item11224 [ Item11224: Use of uninitialized value in =TablePlugin/Core.pm= ]
http://foswiki.org/Tasks/Item11506 [ Item11506: Value of GMTIME{"$week"} is not ISO8601 conform ]
[01:00]
SvenDowideitpharvey, imo if its in a skip, then that should be because something totally optional to a tested commit has happened [01:01]
gac410I thought we included Lexicon stuff in our local CPAN lib. [01:01]
SvenDowideitie, dev didn't install a module that is required is not optional [01:01]
pharveywell, the point is that MAKETEXT *isn't* using a skip at all
but I guess we're saying this should be a hard dep
so it's not appropriate to skip anyway
[01:01]
SvenDowideitelse we'll get lots of commits where devs don't know what they don't have installed
yes, a hard dep - at minimum for UnitTestContrib
[01:01]
Babarok, gac410, I have nothing in the pipe [01:02]
pharveyIt should be *very* obvious by now - the skip summary says exactly that you're missing selenium, Apache::Htpasswd, Crypt::SHA, etc. [01:02]
SvenDowideitpharvey, yes, but if the tests pass
then i can commit
so skipped for missing modules _is_ dangerous
[01:03]
pharveyso does this mean we want everybody to run selenium? [01:03]
SvenDowideiti'd rather fail for Apache::Htpasswd, Crypt::SHA
as those are core bits
[01:03]
GithubBot[foswiki] foswiki pushed 1 new commit to Release01x01: http://git.io/wVPftg
[foswiki/Release01x01] Item11383: Release notes for 1.1.5 Beta 1 - GeorgeClark
[01:04]
***GithubBot has left [01:04]
FoswikiBothttp://foswiki.org/Tasks/Item11383 [ Item11383: General documentation task for 1.1.5 release ] [01:04]
SvenDowideitwhereas Selenium is not yet a fully er, doitorelse infra [01:04]
pharveyit should be
I don't release a new TinyMCE unless selenium passes
[01:04]
gac410Well ... I think we've pretty much deprecated Apache::Htpasswd - it really shouldn't be used. [01:04]
SvenDowideitie, the Selenium tests should really be in a seperate contrib [01:04]
pharveythey are [01:04]
SvenDowideitseparate from UnitTestContrib?
SvenDowideit has no idea
[01:05]
pharveyTinyMCE's Selenium tests are in WysiwygPlugin [01:05]
SvenDowideitthats not separate [01:05]
pharveyhow is it not separate?
Where else would you put them?
[01:05]
SvenDowideitits in the core package [01:05]
pharveyBecause it's a core test [01:05]
SvenDowideitargh!
clearly, its not a _real_ core test
else it would be run by all devs
SvenDowideit wonders if gac410 runs them for releases
[01:05]
pharveyNot all devs are uploading new TinyMCE releases [01:06]
Babarnot all devs have taken care to setup selenium tests [01:06]
SvenDowideitno, but all devs need to check if their shit breaks core stuff [01:06]
BabarBabar never finished [01:06]
gac410selenium nope. I've not been able to get it set up. [01:06]
SvenDowideiti've got 2 selenium setups, but i need to add it to the nightlys
inset more time for the OP part of devops
[01:06]
Babaranyway... there is no other logical explanation of the failure Sven [01:07]
SvenDowideitBabar, poop [01:07]
Babarso as I can't debug it myself... [01:07]
SvenDowideityup [01:07]
Babarthe code called Foswiki::Macros::MAKETEXT [01:07]
FoswikiBothttp://trunk.foswiki.org/System/PerlDoc?module=Foswiki::Macros::MAKETEXT [01:07]
Babarwhich called Foswiki::I18N::maketext [01:07]
FoswikiBothttp://trunk.foswiki.org/System/PerlDoc?module=Foswiki::I18N [01:07]
Babarwhich does: [01:07]
pharveyI've been trying to get trinbuild1-cdc nightlies going with selenium tests, but been fixing pseudo-install to be more windows-friendly instead [01:07]
Babarif ( $text =~ /^_/ && $text ne '_language_name' ) {
this test can't fail. So if it fails, it has to be because the I18N isn't initialized
[01:07]
pharveypharvey assumes gac410 assumes that pharvey is running selenium tests :P
(this is true)
[01:08]
gac410gac410 assumes someone is running them [01:08]
pharveydefinitely need a nightly selenium run though
thought I'd get it running last weekend, but (misc)
[01:08]
gac410I need a 2nd server to run selenium? [01:09]
pharveyno [01:09]
Babarah wait... [01:09]
pharveyjust start a java applet, configure UnitTestContrib with a valid user/pass
and ensure you have the selenium perl driver lib
[01:09]
BabarBabar has an idea! [01:10]
pharvey(and ensure your server env has firefox or some other supported browser, but other browsers need more config effort) [01:10]
BabarExpected:'<span class="foswikiAlert">Error: MAKETEXT argument's can't start with an underscore ("_").</span>' But got:'_edit'
VICTORY!
of course...
no languages are enabled by default, therefore the I18N isn't loaded
which explains the error
[01:10]
SvenDowideitmumble
so its missing some setcfg stuff
SvenDowideit giggles
[01:11]
pharveyactually, the new version of selenium (which requires a bit of refactoring to use the new API) emulates a totally fake browser/DOM so you don't need a real browser at all, but meh [01:11]
SvenDowideitpharvey, oooo, fun :[ [01:12]
pharveyI'd rather test with real browsers, namely the IEs which none of us devs seem to use willingly [01:12]
Babarok, got a fix
just need a task
[01:13]
pharveytheir SeleniumGrid thing can distribute tests across a multi-server-setup to run lots of browsers at the same time, but that's more config effort than anybody here has time for [01:13]
Babargac410: could you wait 5 minutes? :) [01:13]
gac410sure - I'm just testing a trial build now. [01:13]
SvenDowideitmmm [01:15]
BabarSvenDowideit: if I give you a fix, can you try it on the build server real quick?
in UnitTestContrib/test/unit/Fn_MAKETEXT.pm, just after $Foswiki::cfg{UserInterfaceInternationalisation} = 1;
put:
$Foswiki::cfg{Languages}{de}{Enabled} = 1;
[01:15]
SvenDowideityup
fuxed
so we need 2 tests
one with enabled
one without
and one with $Foswiki::cfg{UserInterfaceInternationalisation}=0
so that we cover the real possibilities
maddenin :)
i'm so glad i didn't tell Babar about just changing the pseudo-install default cfg generator :D
[01:17]
Babarnah. MAKETEXT doesn't make any sense unless all is enabled
it ought to just return the string
so it's kind of stupid to test that
[01:20]
SvenDowideity, and thats the point
the test checks to see that it actually does 'just return the string'
do you trust the code without testing it? (ok, so yes, we do :))
Can't use string ("%META:TOPICINFO{author="AlexBrat") as a HASH ref while "strict refs" in use at /home/sven/src/foswiki/core/lib/Foswiki/Func.pm line 2000.
at /home/sven/src/foswiki/core/lib/Foswiki/Func.pm line 2000
Foswiki::Func::saveTopic('Main', 'ABRLogBook', undef, '%META:TOPICINFO{author="AlexBratovic" date="1043799660" forma...', '%META:TOPICINFO{author="AlexBratovic" date="1043799660" forma...', 'HASH(0x20d0570)') c
[01:20]
gac410Whoops - found something that used the obsolete DEPENDENCIES file. Configure: Error: Failed to load DEPENDENCIES: No such file or directory [01:21]
SvenDowideiter, that makes no sense to me [01:21]
FoswikiBothttp://trunk.foswiki.org/System/PerlDoc?module=Foswiki::Func [01:21]
SvenDowideitSvenDowideit is pasting crazy and will stop now [01:23]
Babarand I'd write unit test to test all variables if we had a proper framework :) [01:23]
SvenDowideitdeclare something as proper
and pharvey will contempalte interrupting his stack and frying his brain :)
} catch Foswiki::AccessControlException with {
my $e = shift;
# see documentation on Foswiki::AccessControlException
push( @output, $e->stringify() );
[01:24]
FoswikiBothttp://trunk.foswiki.org/System/PerlDoc?module=Foswiki::AccessControlException [01:25]
SvenDowideitnope, X is not helping me c&p [01:25]
Babarok, committed the fix
funny. All languages where enabled by default in the Release branch
[01:25]
GithubBot[foswiki] foswiki pushed 1 new commit to master: http://git.io/iFjtmA
[foswiki/master] Item11584: Ensure we have at least one language enabled, otherwise Foswiki::I18N does nothing - OlivierRaginel
[01:32]
***GithubBot has left [01:32]
FoswikiBothttp://foswiki.org/Tasks/Item11584 [ Item11584: Fn_MAKETEXT::test_underscore fails unless at least one language is enabled ] [01:32]
gac410SvenDowideit: Any concern pushing @dir, 'Foswiki', 'Contrib', 'core' ... so configure can find DEPENDENCIES file?
ie, @dir points to lib, and I think you removed lib/DEPENDENCIES
[01:33]
Babarcrap
forgot to edit the message, I hope it went well anyway
[01:33]
SvenDowideiti'd need to know why we can't make it explisit
SvenDowideit wonders why configure should be looking somewhere different for core (other than historical)
[01:34]
yes, I am having brain trouble [01:41]
gac410SvenDowideit: It looks directly in lib. And IIRC, lib and core had duplicate DEPENDENCIES files [01:41]
SvenDowideitna, it was more crap than that [01:41]
gac410lib/DEPENDENCIES and lib/Foswiki/Contrib/core/DEPENDENCIES. In my 1.1.5 build lib/DEPENDENCIES is missing. [01:41]
SvenDowideitas they were out of sync and no-one ever knew which [01:41]
gac410right. [01:41]
SvenDowideityes, so the q i have
why would we want to look in lib at all
[01:42]
gac410no idea. We just need one. configure dynamically validates DEPENDENCIES [01:42]
SvenDowideitgiven that lib/Foswiki/(Contrib|Plugins)/contribname/DEPS is the place for everything else
i i wonder more - i presume the configure code is looking for the normal place, plus an extra bonus if for 'core'
which we can replace with 'if core, contrib called 'core'
[01:42]
gac410Configure/UIs/CGISetup.pm is hardcoded in sub _loadDEPENDENCIES { to look in lib.

my fix. adds a push to push Foswiki, Contrib, core onto the path for DEPENDENCIES and it works.
[01:44]
SvenDowideityou mean that we don't add all the DEPS from all the contribs? [01:45]
gac410nope. [01:45]
SvenDowideitoh, sorry, i presumed we were not-dumb :)
push away and add 1.2 task
SvenDowideit needs bot:addtask
[01:45]
gac410We don't even ship non-core DEPENDENCIES [01:46]
SvenDowideitSvenDowideit kills someone
cdot always insisted that we did - in the installer
but you're right, its insaneaboblle
[01:46]
GithubBot[foswiki] foswiki pushed 1 new commit to Release01x01: http://git.io/ZsZt9g
[foswiki/Release01x01] Item11584: Ensure we have at least one language enabled, otherwise Foswiki::I18N does nothing - OlivierRaginel
[01:46]
***GithubBot has left [01:46]
FoswikiBothttp://trunk.foswiki.org/System/PerlDoc?module=Foswiki::I18N [01:46]
SvenDowideitso, um [01:47]
gac410For extensions installed by configure, they are stashed as part of the _installer package into working/configure/pkgdata/blah_installer [01:47]
SvenDowideitthats a blocker for 1.1.5
i need to build lib/DEPs from all the lib/F/(C|P)/*/DEPs
[01:47]
gac410what, not shipping DEPENDENCIES files? [01:47]
SvenDowideityea
do the push for this build
[01:47]
gac410:p Extensions save DEPENDENCIES into the pkgdata [01:48]
SvenDowideitthen i have to update the build scrip
y, and that won't work for !included things
this include MANIFEST stuff is a little dodgey
[01:48]
gac410right. There is a proposal to restructure the DEP and MANIFEST info Babar and I are committed iirc. [01:49]
SvenDowideitsssshhhh
cos i was the one that broked it :)
cos i need DEPs and MANIFEST stuff to do packaging
[01:49]
gac410extension installer aint done until it completely blocks those pesky package manager stuff.
:-D
[01:50]
SvenDowideitSvenDowideit shoots self in foot
chapter1 of book: fast installations of foswiki
[01:50]
gac410tar -zxf ... happiness [01:51]
SvenDowideiti wish that were good enough for people [01:52]
gac410I'm pullin your leg you know, right. [01:52]
SvenDowideiti'm not
i really really wish that tar was good for people
ok,
i think we broke Func::SaveTopic*
it says in there
*Deprecated* 6 Aug 2009 - use saveTopic instead.
=saveTopic= supports embedded meta-data in the saved text, and also
supports the full set of save options.
[01:52]
gac410anyway. I don't think these issues are enough to stop me from going ahead w/ 1.1.5 beta on f.o ... however that would be. [01:53]
SvenDowideitgac410, exactly
(this is on trunk right now, but i will be running same code on 1.1.4 soon
[01:54]
gac410SaveTopicText is deprecated, SaveTopic is the replacement. righ? [01:54]
SvenDowideity, but it looks like both don't do what they should with leading %META:TOPICINFO [01:55]
gac410Bad mojo - not allowed. [01:55]
SvenDowideitas i quoted
*Deprecated* 6 Aug 2009 - use saveTopic instead.
=saveTopic= supports embedded meta-data in the saved text, and also
supports the full set of save options.
its _supposed to be allowed
and the deprecated one certainly did do it
[01:55]
gac410IIRC that was really breaking stuff. I'm pretty sure TOPICINFO must not be embedded int eh text [01:56]
SvenDowideitand that would be a backwards compatibilty bug [01:56]
gac4101.1.4 is restricted as well [01:56]
SvenDowideitat least one of these needs to support it [01:56]
gac410Part of CDot's fixes for TOPICINFO issues. It must be the first line, and only the first line. [01:57]
SvenDowideitna, that was mine
but this _is_ the first line
ie, it looks more like saveTopcText isn't extracting the meta from it anymore
[01:57]
gac410Hm. saveTopic separates out Meta though. I assumed that the supports embedded was a typo. [01:58]
SvenDowideitSvenDowideit hopes this checkout isn't mongo polluted :/
docco is docco :)
given the docco example for saveTopicText too
[01:58]
gac410docco can be wrong as easily as code .. .sometimes more easily [01:58]
SvenDowideityup, a bugisabug
but i'm pretty tooting sure that the original Func impl saved with embedded meta
mmm, it occurs to me that saveTopicText does extract meta correcty on 1.1.4
but that saevTopic didn't
mmm, backpedaling - it seems to have done FIELDS ok
[01:59]
gac410writeatest [02:04]
SvenDowideityup :/ [02:04]
GithubBot[foswiki] foswiki pushed 1 new commit to Release01x01: http://git.io/SllFMA
[foswiki/Release01x01] Item11585: Load DEPENDENCIES from correct location - GeorgeClark
[02:17]
***GithubBot has left [02:17]
FoswikiBothttp://foswiki.org/Tasks/Item11585 [ Item11585: Configure loads =DEPENDENCIES= from lib, but file is in =lib/Foswiki/Contrib/core= ] [02:17]
GithubBot[foswiki] foswiki pushed 1 new commit to master: http://git.io/ieh0fw
[foswiki/master] Item11585: Load DEPENDENCIES from correct location - GeorgeClark
[02:17]
***GithubBot has left [02:17]
pharveySvenDowideit: what thing did Babar want to do that our test infra currently can't? (btw I'd love to use Test::More or Test::Class, but we need more devs :) [02:18]
SvenDowideitaward for frustration: assert_matches has a ! version
what is it called?
assert_not_matches would be my first thought, but nooooooo
[02:19]
pharveyyeah, I also like str_equals... should be string_equals or str_eq, damn schizophrenic code :) [02:20]
SvenDowideiti blame people [02:20]
pharveywait.. I thought we only had assert_matches. We have a negatory one too? [02:21]
SvenDowideityes
assert_does_not_match
[02:21]
pharveyohhh that rings a bell [02:21]
SvenDowideitdesperatly important to test both what you want and what you really don't want [02:21]
pharveywell, I guess you could /(?!....)/
or is that (?!: - I forget
[02:21]
SvenDowideitwhich would make me cry more [02:22]
GithubBot[foswiki] foswiki pushed 1 new commit to Release01x01: http://git.io/BL1avg
[foswiki/Release01x01] Item000: BUILD Foswiki-1.1.5-Beta1 at Wed Feb 29 02:22:32 2012 GMT - GeorgeClark
[02:33]
***GithubBot has left [02:33]
FoswikiBothttp://foswiki.org/Tasks/Item000 [ Item000 ] [02:33]
...... (idle for 25mn)
gac410foo - Someone has been updating System topics on f.o - bunch of conflicts [02:58]
SvenDowideitcrap. 1.1.5 seems to also kill saveTopicText :(
i'm getting nothing done today - i think i hate wednesdays
SvenDowideit declares: WackyWednesday
[03:11]
gac410thankfully it's tuesday here [03:12]
SvenDowideithttp://www.robbyonrails.com/articles/2008/04/10/git-svn-is-a-gateway-drug
SvenDowideit runs git svn fetch for the first time on a > year old repo
[03:19]
gac410Looks like CDot updated the System.DevelopingPlugins topic to add in his new DISPLAYIF stuff. [03:23]
GithubBot[foswiki] foswiki pushed 1 new commit to Release01x01: http://git.io/GNOSXQ
[foswiki/Release01x01] Item11383: Sync doc changes edited on foswiki.org - GeorgeClark
[03:30]
***GithubBot has left [03:30]
FoswikiBothttp://foswiki.org/Tasks/Item11383 [ Item11383: General documentation task for 1.1.5 release ] [03:30]
SvenDowideityup, confirmed
f 1.0 saveTopicText does what i expect
now to figure out when we busted it :(
[03:39]
gac410or if we did :) [03:40]
SvenDowideitwe did [03:43]
gac410open an urgent if you need it in 1.1.5 [03:43]
SvenDowideiti need it in 1.1.4, and 1.1.5 will probly be out after i need it >:}
but ya
[03:44]
gac410well ... if you invent the wayback machine, we'll be sure to get it into 1.1.4 before it releases last December.
:D
[03:44]
SvenDowideitexcellent - thanks >grin<
this is the point where i wish that the tags had the unit test contrib in them
[03:45]
gac410Maybe Babar can get some of those superluminal neutrinos for the time machine [03:47]
SvenDowideityeah, but they're italian
so probly break if they travel too far
[03:47]
GithubBot[foswiki] foswiki pushed 1 new commit to Release01x01: http://git.io/Fds00w
[foswiki/Release01x01] Item11383: Fix typo - GeorgeClark
[03:48]
***GithubBot has left [03:48]
SvenDowideitmmm, do we have tags in git
i mean git-tags
not svn tag-branch things
[03:48]
GithubBot[foswiki] foswiki pushed 1 new commit to master: http://git.io/A-PIvg
[foswiki/master] Item11383: Sync doc changes edited on foswiki.org - GeorgeClark
[03:49]
***GithubBot has left [03:49]
gac410I don't think we use them [03:49]
SvenDowideitwe do for 1.0*
it seems
ah, no, they're shite
this'll be 'fun'
FoswikiBot, help
[03:49]
FoswikiBotSvenDowideit: Ask me for help about: google, dns, foswiki, vars, join, karma, auth, assemble, seen, infobot, corelist, weather, loader (say 'help <modulename>'). [03:53]
SvenDowideitFoswikiBot, help foswiki
FoswikiBot, foswiki 1.1.0
WikiRingBot, help
[03:53]
WikiRingBotWikiRingBot v0.3 ($Rev$), a logger bot
(c)opyright 2007 MichaelDaum@WikiRing.com
Type 'WikiRingBot, <command>' or /msg WikiRingBot <command>, commands are:
hello: say hello
help: this text
info: display the current bot status (logging, seen nicks, etc) and some more info
join <channel>[:<password>] : join a channel, optional password key (admins only)
leave [<channel>] : part from a channel, defaults to current
quit : terminate bot (admins only)
reload: reloads its configuration (admins only)
replay [<date>]: replay the logs of this channel, date might be the date in yyymmdd format of an earlier date, defaults to today's logs
seen <nick> : report when nick was last seen
[03:53]
FoswikiBotSorry, I haven't seen <nick> : report when nick was last seen. [03:54]
WikiRingBotstart [<channel>]: starts logging a channel, defaults to the one the call came from
stop [<channel>]: stop logging a channel, defaults to the one the call came from
time: report current GMT time
[03:54]
gac410what are you looking for SvenDowideit [03:55]
SvenDowideithow to checkout each release
but the release with dev tools
atm, i'm trying
git reset --hard ; git clean -fd ; git checkout `git rev-list -n 1 --before="2010-10-04" fix1x1` ; ./pseudo-install.pl -A developer
[03:56]
gac410Which release? [03:57]
SvenDowideiteach [03:57]
gac410oh. [03:57]
SvenDowideitie, loop through all git tags for the dev version of a release would be ideal :)
so.
it looks like it was ok in 1.0.10
and busted in 1.1.0
(the legacy saveTopicText
i'm much less concerned about the non-legacy, as that functionality is shite anyway
[03:57]
gac410git log --oneline | grep BUILD will give you the abbreviated hash for each build. [03:59]
SvenDowideitooo, very nice
thankyou
those are hashes in the release branches?
mmm, an i get nothing
[04:00]
gac410yeah. However if there were multiple attempts at a release, you'll need to grab the latest.
git log --oneline | grep BUILD
773761e Item000: BUILD Foswiki-1.1.5-Beta1 at Wed Feb 29 02:22:32 2012 GMT
4cf81f3 Item000: BUILD Foswiki-1.1.4 at Tue Dec 20 15:32:06 2011 GMT
[04:01]
FoswikiBothttp://foswiki.org/Tasks/Item000 [ Item000 ] [04:01]
SvenDowideitmmm, an i get nothing cos i'm in a weird place
ie, it only works if you have a HEAD checkout of the rel branch
but that should let me make git tags for it
[04:01]
gac410yeah. i do everything from a full git svn checkout of trunk and release. [04:02]
SvenDowideitlalala :)
actully
1.1.0 might be ok
just different
[04:02]
gac4107a5d10a Item000: BUILD Foswiki-1.1.0 at Mon Oct 4 15:25:59 2010 GMT [04:02]
SvenDowideitSvenDowideit crys a little more
1.1.0 works ok
[04:03]
gac410c2a1f3d Item000: BUILD Foswiki-1.1.2 at Wed Nov 10 00:28:35 2010 GMT
0f5e535 Item000: BUILD Foswiki-1.1.1 at Mon Oct 25 21:57:27 2010 GMT
[04:04]
SvenDowideitiHATEcdot's logs dir crash thing
why can't it just friggen create the dir
(yes, i grok why, but argh!)
oho.
i was right
1.1.4 is ok
so it is a 1.1.5 tissue
[04:06]
gac410You know about git bisect, right? [04:07]
SvenDowideitonly 4 hours more wasted on it
y
but i wanted to know about specific releases
i know i should use bisect more tho
[04:07]
gac410makes isolating things to a specific commit so much easier. The instructions on HowToUseGit are a bit more involved than really required. [04:08]
SvenDowideitmmm, ok, so we're weird.
SvenDowideit goes back to 1.0
which he does not have
on dear
the 1.0 branch is dead :(
how did i get it tested a moment ago
ah, forgot to git svn rebase this one
[04:08]
oh dear
so - saveTopicText in 1.0 with a META:TOPICINFO{author="something" gets saved as though the line didn't exist
[04:22]
gac410I believe that Store is supposed to create a new TOPICINFO. The user / plugin / whatever should never be able to control it directly. [04:24]
SvenDowideityes - ish - to a point :)
but in this case, i agree
the thing here, is that my test call of saveTopicText intentionally uses an author that isn't there
so i expect that to dissappear, as the TOPICINFO is created by the system
1.1.4 on the other hand
[04:24]
gac410The TOPICINFO should set the author to the author doing the save. [04:26]
SvenDowideitsaves the TOPICINFO as is
y, and 1.0.10 is does
[04:26]
gac410I think that sounds broken. [04:26]
SvenDowideitin 1.1.4, it does a shufty
but the topic text is ok, as there is no meta duplication
[04:26]
gac410Does 1.1.5 get it right (Create TOPICINFO from the current user saving the topic) [04:26]
SvenDowideitin 1.1.5pre, we get 2 lines of TOPIC:INFO
so it stuffs up in a magically more obvious way
neat huh
[04:27]
gac410Hm. I worked on that a bit too. one of the extensions was doubling up the META [04:27]
SvenDowideitlets play with 1.1
1.1.1 i mean
mmm, checking out 1.1.4 again to be sure
[04:27]
gac410crap Can't search for TOPICINFO ... it finds every topic. [04:30]
SvenDowideitoooo, maybe
i can't trust that i'm seeing anyway
cos it looks like readTopicText might have been bastardised the same way - but i can't tell without reading the file :(
mmm, maybe not, but i'm going to need to be sure
ok, so 1.1.5 looks to have the same bug as 1.1.0
[04:31]
gac410I think it was NewUserPlugin - was doubling up the META::TOPICINFO [04:36]
SvenDowideitcompared to 1.0.10 which i think is correct vs historical
SvenDowideit goes look
i'm not using it on my trunk
[04:36]
gac410no- not saying it's related to your issue. It was a support q that reported doubling of the TOPICINFO
Trying to find what it did if I did anyting.
[04:38]
SvenDowideity - mongo did seem like it had the issue too
but it might just be that trunk has a porblem
[04:38]
gac410Sven - Item11334 had to do with the dup topicinfo [04:40]
FoswikiBothttp://foswiki.org/Tasks/Item11334 [ Item11334: NewLdapUserTemplate displays =%META:TOPICINFO= at top of each new user topic ] [04:40]
SvenDowideitaha.
so fixed the symptom
but the cause might be the issue that i'm chasing
darnit
[04:41]
gac410Yeah that's why I was searching for it - I knew it sounded familiar.
Foswikirev:11894 ... you made a change to preserve TOPICINFO in the topic text?
[04:43]
FoswikiBothttp://trac.foswiki.org/changeset/11894 [ Changeset 11894 – Foswiki ] [04:44]
SvenDowideitmmm
y, thing is, this _is_ the first line
but lets see
[04:49]
gac410gac410 feels funny about any api accepting a topicinfo from unknown sources. ... If we are to truly TRUST topicinfo to reflect the RCS rev info, then we have to be the controller of it. [04:50]
SvenDowideitthis is a legacy API
the API trusts embedded META up to a point
it does not seem to use the META:TOPICINFO directly
[04:51]
gac410ah saveTopicText ... the deprecated one? [04:51]
SvenDowideitbut if there are additional attrs, they seem to follow
yes.
i'm not looking deeply at the new one
though i probly should
sadly, waht this really has shown, is that there is a small breakage between 1.0 and 1.1 and a huge breakage in 1.1.5
er
sadly, waht this really has shown, is that there is a small breakage between 1.0 and 1.1 and a huge breakage in trunk
i'm still confised wrt 1.1.5
cos, um, i'm getting zero done today :(
oh argh!
SvenDowideit just asked git to checkout an svn rev number
[04:51]
gac410did it work? [04:55]
SvenDowideitnope [04:55]
gac410well that's good I think. [04:55]
SvenDowideitmore not-useful-ness
i wonder which numpty thinks
Timestamp:
06/12/11 12:14:41
is actually useful
like - is that december, june, and what tz are we talking about
oh dear
sven@quiet:~/src/foswiki_rel1.1/core$ git rev-list -n 1 --before="2011-12-07" master
40ad0fbeb1fe4c3bdc291d99cb4d9906b97f9e57
sven@quiet:~/src/foswiki_rel1.1/core$ git rev-list -n 1 --before="2011-12-06" master
40ad0fbeb1fe4c3bdc291d99cb4d9906b97f9e57
so its telling me there are no commits there
and that id is for something way off - in sept
2010!
for want of an git svn rebase in master?
SvenDowideit hasn't used master in a while,
nope, that wasn't it
[04:55]
OMG trac is in US mode? [05:08]
ok gac410 i think that means that
i fixed the 1.1.1 introduced issue where we just used the TOPICINFO coming in blindly
but then in that commit you linked
that i then broke things that use saveTopicText
subtle as a brick in the face that one :(
[05:20]
gac410wow - that was the breakage? I was just poking around and the comment looked interesting. lucky guess [05:21]
SvenDowideitvery very lucky
and well done
[05:21]
gac410thats back in last june, so 1.1.4 has to be busted as well then. [05:22]
SvenDowideitthat commit is june for trunk [05:23]
gac410ah. oops. [05:23]
SvenDowideitand i have a bad habit of fixing things in trunk and forgetting to merge
mmmm
mm, nope
this is in 1.1 branch too
and now i have no idea why 1.1.4 seems to work ok
argh
ok, i know we have a problem
but i can't look at this anymore, my week is over :(
[05:23]
gac410so bottom line - is this a 115 blocker? [05:27]
SvenDowideitdouble up of TOPICINFO - probly [05:27]
gac410and any impact to my installing 115 on f.o - probably 8-9 hours from now. [05:27]
SvenDowideitnone [05:27]
gac410okay great. [05:27]
SvenDowideitsleep well :) [05:30]
***gac410 has left [05:30]
..... (idle for 20mn)
pharveySvenDowideit: when's your blog 'deadline' [05:50]
SvenDowideittomorrow :)
i've given up on today totally
as i've wasted it all on a bug that i could have trivially worked around and ignored like everyone else has :)
worse
its due to a fix i made for something Temi found
but i only unit tested
i didn't write an integration test for the Func::save... calls
[05:52]
.... (idle for 17mn)
FFS>
so, i have a checkout that i'm trying to navigate lots of versions on
I try to run
git reset --hard ; git clean -fd ../ ; git checkout 1d28438 ; ./pseudo-install.pl -A developer ; mkdir working/logs
but the git checkout fails.
error: Untracked working tree file 'JQueryPlugin/pub/System/JQueryPlugin/themes/flickr/jquery-ui.css' would be overwritten by merge.
which means that the clean _isn't_
and yet, if i call git clean -df ../ manually a few times
eventually it does
and now i'm getting a totally different result from my previous testing
[06:10]
OK this is mad.
Deprecation policy: 2 further major releases
um, saveTopicText and readTopicText may have been broken for a very long time
but we have to decide how to maintain them for the next how many years?
Functions documented in Foswiki::Func and Foswiki::Meta have the same deprecation policy (retained for at least 2 major releases after deprecating release)
erm, so my store2 work can't happen until foswiki 4.0?
[06:28]
FoswikiBothttp://trunk.foswiki.org/System/PerlDoc?module=Foswiki::Func http://trunk.foswiki.org/System/PerlDoc?module=Foswiki::Meta [06:32]
SvenDowideitSvenDowideit gives up on a really bad day [06:32]
Babar, i need a script :(
something that will take a unit test that is in the current checkout
and run it against each _release_
i'm clearly too stupid to consistently get git to definitily give me what i want
SvenDowideit suspects theres quite a number of things wrong now :(
as i just noticed that 1.0 seems to be removing the META:FILEATTACHMENT :(
[06:38]
pharveySvenDowideit: you need the x option, that's git clean -0fdx
git clean -fdx
[06:42]
SvenDowideitbecause?
SvenDowideit coud try reading but
[06:42]
pharveybecause pseudo-install generates a .gitignore file of all generated files [06:43]
SvenDowideitoh, its you!
you did it to me :)
[06:43]
pharveypharvey hides [06:43]
SvenDowideitgrin [06:43]
pharveyI thought it would be a nice thing, so that 'git status' actually is useful [06:43]
SvenDowideitmmm
i never call git status
cos its so messy
[06:44]
pharveybut with .gitignore files working, it's not messy [06:44]
SvenDowideiti hate being old, i remember bugs that were fixed ages ago :( and so i use the workaround eventhough they're not needed [06:44]
pharveythough, you need to be ignoring working/tmp as well, otherwise git status shows you all the cgisess* [06:45]
SvenDowideitSvenDowideit tasks pharvey with fixing all the bugs [06:45]
pharveyare they fun, interesting bugs? [06:46]
SvenDowideitno
they're all sucky shitty ones
if i'm actually running on a 1.0.10 branch checkout like i think i am
mmm, i have to go pick up the kids in 30mins time
[06:46]
pharveypharvey wonders how we get unpaid volunteers to work on the important boring stuff :) [06:48]
SvenDowideitdeprecate faster
by our deprecation policy
i can't release store2 until foswiki 4.0
[06:48]
pharveyso there's no way we can support Foswiki::Meta in store2? [06:49]
SvenDowideitunless by some fluke i can make a Foswiki::Meta shim that is bug compatible [06:49]
pharveyah [06:49]
SvenDowideitand i consider that a flukifitworks [06:49]
pharveySvenDowideit, is bug-compatible necessary? Just documented-compatibleness is surely acceptable to the community, I guess that means we'd need to test a bunch of plugins :(
CDot knows what to do
[06:50]
SvenDowideitpharvey, i was reading the deprecation doc :) [06:50]
pharveydeprecation of a published API, sure, but I feel as if you're alluding to undocumented 'features' or bugs, is that also required ? [06:51]
CDotpharvey: CDot knows nothing. CDot only just got up. [06:51]
SvenDowideitbut its worse, in that i thinkg Foswiki::Func::saveTopicText and readTopicText might have changined how they work dramatically
without anyone knowing or using it
[06:51]
pharveythat's super-mega-lame [06:51]
SvenDowideitand yet we have to keep them for another 2 major releases [06:52]
pharveylet's make Foswiki 2.0 Foswiki 4.0 instead [06:52]
SvenDowideit2 major releases
not 2 major release numbers
[06:52]
pharveyok ok [06:52]
SvenDowideitmore like, lets make 1.1.5 2.0 [06:52]
pharveyI was being facetious [06:52]
SvenDowideitand 1.2.0 3.0
and thus...
i'm rather tempted you see
or better, make 1.1.5 into 1.2.0 plus we add lots of deprecation notices
then release 1.2.0 from trunk next month as 2.0
[06:52]
CDotCDot can't tell from the logs what this is all about [06:53]
SvenDowideitthen the first patch release to it we call 3.0 [06:53]
pharveySvenDowideit is panicking about keeping Foswiki::Meta bug-compatible with Foswiki 1.1.x in a post-store2 world [06:53]
SvenDowideitoh, it seem that Func::saveTopicText hasn't done what it said ever
but that what it has done has changed several times in foswiki
wrt META:TOPICINFO embedded in the text
[06:54]
pharveywere bugs raised? [06:54]
CDotwrt what? [06:54]
SvenDowideitpharvey, ish and no
i'm kinda discovering this in detail today
[06:54]
pharveyso, part of the pain is validating our understanding of what Func and Meta used to do in 1.0, then what it does in 1.1, and then trying to measure how we break it post-store2
or even in FW 1.2
[06:55]
SvenDowideitin 1.0.10 it looks like saveTopicText with embedded META::TOPICINFO worked ok [06:56]
CDotyes, it should [06:56]
SvenDowideit(but i just noticed that it might be stripping out the attachment meta [06:56]
CDotdefine "ok" though
erm, that doesn;t sound right
[06:56]
SvenDowideitas in, saved essentially ignoring the topicinfo meta
CDot, exactly
[06:56]
CDotit may *correct* the attachment meta-data
as in, "do these attachments really exist"
[06:56]
SvenDowideity, that i have to test further atm
but
in 1.1.0 the TOPICINFO in the text is written verbatim to the file
[06:57]
CDotthough I would not expect that in saveTopicText; i don;t recall any code to that effect
y, that was a bug
[06:57]
SvenDowideitwhich is not right [06:57]
CDotwriting TOPICINFO verbatim - or writing FILEATTACHMENT, if I'm honest - is bug material [06:58]
SvenDowideitthen, in trunk, we have that plus a double up of the TOPICINFO meta
y
[06:58]
CDotcos file metadata should not be driven principally from what's on disc, *not* what's in the file. [06:58]
SvenDowideitbut the core issue is that no-one seems to have noticed - other than me writing new code to import stuff
the key issue for me
is taht saveTopictext only_ exists to support legacy use
so either it works like in twiki 2.3 (or whatever)
or it has no value
[06:58]
CDotwhat do you mean by "noticed"? As in, no unit tests? Or as in "no-one has complained"? [06:59]
SvenDowideitboth [06:59]
CDotcos at the end of the day, the litmus test is "does anyone care"? [06:59]
SvenDowideitexactly
i wasted a day thinking i should care
[06:59]
CDotok, yeah, i do that every so often [07:00]
SvenDowideitbut then as i got pissy about a wasted day, asked myself - what is the point of the deprecation policy
if the code doesn't actually comply with the policy anyway
[07:00]
CDotI think that was the first thing I ever deprecated - and I had a fight about it at the time. [07:00]
SvenDowideiti, and thus any other coder in this posistion
woudl be better off having it not exist
[07:00]
CDotCDot has always seen saveTopicText as evil [07:01]
SvenDowideitthus we crash and re-write with a proper
it is, but its damned useful (in theory) for importing from an old twiki
[07:01]
CDotI don;t think it was ever used for that [07:01]
SvenDowideiti've used it for that many times [07:01]
CDotthe justification for keeping it was plugin calls [07:02]
SvenDowideitmany many times [07:02]
CDotok, i didn;t know that. Why? [07:02]
SvenDowideiti use it to use a func API without having twiki/foswiki clean up the content [07:02]
CDothmmm, /me really isn't sure that's a good thing [07:03]
SvenDowideitand then often do a proper readTopic and meta based saveTopic with a force rev
the point is to get the first rev seeded with what is ni the old one, and then thesecond rev being the fixed one
[07:03]
CDotwhy not just copy the file and doa readTopic/saveTopic? [07:03]
SvenDowideitthus you get a history
copy the file - there's an API for that?
and one that works for non file based stores?
didn't know that
SvenDowideit still doesn't in fact.
[07:03]
CDotwe've never *had* non-file-based stores before now [07:04]
SvenDowideiti had on in 2007 [07:04]
CDotok, never had one that was public [07:04]
SvenDowideitbut then some **** changed the code on me so its dead
but that was before func too
[07:04]
CDota while ago, I decided that RCS is evil, and wrote a non-RCS, just-files store [07:05]
SvenDowideitbasically, i use saveTopicText to create a pre-history
please tell me you put the versions into a separate dir
[07:05]
CDoty, I see why you use it. And also why that depends on it being able to write incorrect TOPICINFO
I never finished it, but no, I used the old VAX/VMS version naming
[07:06]
SvenDowideitoh, i don't mind the TOPICINFO being corrected
its the rest of the problems that hurt
but what hurts more is that we have the API, but that its untested and broken
if it was gone, i would not have spent the day
and instead would have..... been ok with losing pre-history
and when i sit down to think about it, i think that would be better
[07:06]
CDotCDot sees unit tests using saveTopicText [07:08]
SvenDowideitcos anyone else that used the old API would waste time on the broken API
not enough to actually test things it seems
[07:08]
CDotalso MailerContrib! yikes [07:09]
SvenDowideitreadTopicText is just as bad [07:09]
CDotMsOfficeAttachmentsAsHTML too
SignaturePlugin
ChecklistTablePlugin
newUserPlugin
GlobalReplacePlugin
*lots* more
[07:09]
SvenDowideity, micha fixed NewUserPlugin when i broke saveTopicText in 1.1.4
but I'm trying to work out how much more it was broken before that
(as in 1.1.4 i quietly introduced the double TOPICINFO issue
and no-one has kicked up the root cause
[07:11]
CDotI know the root cause - or at least, I found it, though i can't remember what it was
IIRC I added a unit test at the time
when I was working on resolving the "latest is .txt" issues
[07:13]
SvenDowideitthen they should be failing now [07:14]
CDothmm, ok, so we're back to double TOPICINFO? With a plugin handler, or without?
CDot has a vague recollection it was due to a preSave/postEdit handler
[07:15]
SvenDowideitno
without
and i know the commit that caused that one
[07:15]
CDotok, useful [07:15]
SvenDowideitand that is also the task that fixed the save the TOPICINFO verbatim issue
but now i'm looking at what else gets quietly messed with
ffs, who chose the naf names for asserts in the test suite
they're never what i expect
i see assert_equal, and expect assert_not_equal
but no
assert_matches vs assert_does_not_match
[07:16]
CDotdunno. /me copied them from JUnit/PerlUnit [07:18]
SvenDowideitoh god [07:19]
CDot:-)
CDot sees 14 extensions using saveTopicText outside of a test/unit context
and a further 10 using it in unit tests
sorry, not a further 10; just "10"
[07:19]
SvenDowideitSvenDowideit just ported a plugin for someone that er, used alot of direct file access [07:21]
CDotmost seem to be using it because of a perception that "ot's faster" [07:21]
***SvenDowideit changes topic to: Foswiki 1.1.4 is released: http://foswiki.org/Download/FoswikiRelease01x01x04 -- Today's IRC Logs: http://ioh dearrclogs.foswiki.org/bin/irclogger_log/foswiki [07:22]
SvenDowideithuh/
gottago
[07:22]
CDotwhy did you change the topic to that useless link?
CDot isn't a channel op so can't fix it
pharvey: can you comment on the differences in WysiwygPlugin between trunk and Release01x01?
CDot was getting very confused yesterday
[07:22]
pharveyCDot: I saw your task, and added some comment [07:24]
CDotpharvey: Item2516 is a problem; we can't merge TinyMCEPlugin or WysiwygPlugin from trunk, because the changes are incompatible
but I was expecting that
[07:30]
FoswikiBothttp://foswiki.org/Tasks/Item2516 [ Item2516: Add new TML syntax to support indented paragraphs ] [07:30]
pharveyOk; I wasn't expecting that
I hoped it might be possible to make it 1.1 compatible
[07:30]
CDot, is your time spent on WYSIWYG urgent? I want to ask if you can wait till the weekend when I can resume that work
if not, I understand
gotta run, will check the logs
[07:38]
CDotpharvey: yes, it's a release blocker for 1.1.5 [07:39]
..... (idle for 24mn)
SvenDowideitsorry, had to run, was late to pick up girls [08:03]
***SvenDowideit changes topic to: Foswiki 1.1.4 is released: http://foswiki.org/Download/FoswikiRelease01x01x04 -- Today's IRC Logs: http://irclogs.foswiki.org/bin/irclogger_log/foswiki [08:03]
SvenDowideitCDot, paul is expecting to work on 1.1.5 wysiwyg bugs on the w/e [08:04]
CDotSvenDowideit: thx, I got that. [08:04]
....... (idle for 30mn)
SvenDowideitoh ffs.
http://foswiki.org/Tasks/Item11586
had 2 %META: in the verbatim
SvenDowideit attaches latest version of unit test as a patch
and then goes for girls
[08:34]
.......... (idle for 46mn)
***ChanServ sets mode: +o MichaelDaum [09:20]
............... (idle for 1h11mn)
pharveysilly aarnet's debian mirror seems to be constantly b0rked [10:31]
kip3fHi pharvey, I posted an update to http://foswiki.org/Development/CreateISODateFormat I defer to your experience in writing wiki apps with date/times
The thing that I really am concerned with is to clean up storage for 'pure date' fields
[10:38]
.... (idle for 16mn)
pharveyhi kip3f, I must confess that when my users sort on a date/time string field, they're almost happy enough to get them grouped by day. Preserving the GMT offset is an important signal useful for data validation and gives extra confidence as to the data quality (not to mention it's easier to judge the context for the time: an hour gap between records might make more sense if it was around...
...lunchtime, and a 3am record might be absurd if we know they could only be operating during the day).
It can also help explain out-by-one date errors, if the system was rolling over midnight during the day
In my previous job I also had the job of sometimes correcting operator errors whereby the operator had 'corrected' the time of some equipment transported to a new TZ by simply adjusting the clock, instead of the GMT offset in the software
much easier to recover bad data if the GMT offset assumed at the time is known
[10:55]
CDotpharvey: you should move a more sensible timezone ;-) [10:59]
pharveypharvey dreads the day GMT gets DST :P [10:59]
CDotGMT can't, by definition; it is Zulu time [10:59]
pharveymy attempt at a lame joke
lameness achieved!
[11:00]
kip3fmore precisely, we should be all talking about times in UTC not in GMT
http://geography.about.com/od/timeandtimezones/a/gmtutc.htm
[11:00]
pharveyI know, I know
old habits die hard
[11:01]
CDoterm, UTC *is* GMT, just more accurate [11:01]
pharveykip3f: I think it's better to truncate, rather than zero-out the 'unwanted' parts of a date-only time
(if there is an option to do so)
[11:01]
kip3fI agree. I just what to know if it will be enforced in code
want
[11:02]
CDotpharvey: did you conclude what to do about iso8601 "simple form" i.e. 20120229?
CDot can't see how epoch times can be supported if you support iso8601 simple form
[11:03]
SvenDowideitthere is a way [11:06]
pharveyCDot: my only idea is to change it from 4 consecutive digits to 9+ :( [11:06]
SvenDowideitput epochtimes into a different attr [11:06]
pharvey99999999 is circa 1973
(I think)
[11:06]
SvenDowideitand if the form is defined to take them as user input, then it won't take others [11:07]
CDoty, epochtimes in another attr is fine, so long as it's clear what has precedence [11:07]
SvenDowideitit does?
why?
[11:07]
CDotit does what? [11:07]
SvenDowideitaka - precendence for what [11:07]
pharveyfor *TIME macros, I wrote that we should emit an foswikiAlert if the user provides both a stringy and an epoch time param [11:07]
SvenDowideit<CDot> y, epochtimes in another attr is fine, so long as it's clear what has precedence [11:07]
CDotright; if I defined "20120229" epoch="123454678910" it has to be clear what has precedence; _DEFAULT or epoch (in this case)
CDot was simply stating the bleeding obvious
[11:08]
SvenDowideitbasically, i think we should store the ecpoch milliseconds for internal use - and the user input is justa record of the user input [11:08]
pharveyif you're referring to the lively discussion over at http://foswiki.org/Development/CreateISODateFormat though, I don't think we're advocating two different storage formats, but rather to pick one or the other (but maybe as in kip3f 's case for timeless-dates, YYYYMMDD is fine) [11:08]
SvenDowideitok, but doing so in an obtuse way :p [11:09]
CDotepoch *milliseconds*? Hmmm, interesting [11:09]
SvenDowideiti've needed more than milliseconds before too, but sssshhhh
once we introduce 123412341234.1234124123
[11:09]
pharveypharvey wonders which proposal CDot is talking about :) [11:09]
CDotpharvey: we were talking about auto-recognising epoch times in date macros such as %UTCTIME%. I forget where. [11:10]
pharveyyes, good
I wrote in the proposal that it should emit an foswikiAlert
I'm sick of wasting time helping users track down silent-failures of this nature
they need to know they did something dumb, and why
[11:10]
CDotfairy snuff. [11:11]
SvenDowideitwe need to do alot more of that :/ [11:11]
kip3fSo a wiki app can set a datetime to '2012-02-26T07:00-0800' and a property will be automagically set, epochseconds='1330268400' [11:11]
CDota lot of the time error reporting isn't "obvious" so people don't bother with it. Even foswikiAlert isn't pertickly well supported
kip3f: is '2012-02-26T07:00-0800' == '2012-02-26T07:00-0800Z' ?
[11:11]
pharveyno, but it's better than zero. Actually, I've wondered if a ADDTOZONE{"foswikiAlerts" text="You numpty, pick an epoch OR a date string, not both!"} would be interesting [11:12]
CDotoops, sorry, misread the string
-0800 is fine
[11:12]
SvenDowideitpharvey, yes, it would [11:13]
CDottext= or error= ? [11:13]
pharveyI guess really, macros should return Foswiki::Func::foswikiAlert('id10t error') and we can figure out niceness later [11:13]
FoswikiBothttp://trunk.foswiki.org/System/PerlDoc?module=Foswiki::Func [11:13]
SvenDowideitADDTOZONE
basically
[11:13]
CDotCDot ponders making error= a 'standard' [11:14]
SvenDowideitmacros should be able to return stuff __and__ alert [11:14]
CDotSvenDowideit: YES!
I would like a flash alert
[11:14]
SvenDowideitelse we're numpties all over again
as there are lots of things that work, but should complain
like %SCRIPTURL%/view%SCRIPTSUFFIX%/%WEB%/%TOPIC
[11:14]
CDotarrrrgggghhhhh
CDot *hates* that
[11:15]
pharveyhaha! [11:15]
SvenDowideitya, me too
i re-write it when i have time in my importers
[11:15]
pharveyI'm 99% sure I deleted one of those from the System web only last year [11:15]
SvenDowideitbut thats also why i use saveTpicText first
SvenDowideit depressed
i wrote a test on 1.0.10, including FILEATTACH
in 1.0 the ATTACH is removed
in trunk, it stays
[11:15]
pharveykip3f: I love the idea of extra pseudo-properties (like an epoch version of the iso-string) popping into existence automagically, but I worry about the mechanics of achieving that in the vanilla store..> CDot? [11:16]
SvenDowideitso now i have to go find 'ancient history' make it work again and run there [11:16]
CDotdon't be depressed! It's spring! The birds are singing, the flowers are out, it's not raining (for a change)! [11:16]
SvenDowideitCDot, argh wat? [11:16]
pharveyit's raining cats & dogs here for 5 days :P [11:17]
SvenDowideitIt's SUMMER! The birds are singing, the flowers are out, it's not raining (for a change)!
you guys have stupid weather.
and you're on the wrong side of the world
[11:17]
CDothey, GMT == UTC == 00:00. We're on the *right* side of the world ;-) [11:17]
SvenDowideitnope, thats just silly
wow.
i crashed my ntebook for the first time in a long time
[11:18]
CDotwhat's more, I'm to the right of the Greenwich meridian [11:18]
SvenDowideiti decided to try gnu screen
and split screens
totally boooom
back to X for me :(
[11:18]
pharveycan't you use vim's split screens? :) [11:19]
SvenDowideitfor normal use, probly
but er lalala
[11:19]
pharveykip3f: FWIW we already have META:CREATEINFO which isn't-really-similar but an example of a META:DATUM that pops itself into existence without actually being in the topic.txt [11:20]
kip3fIIRC I have tried stuffing a value into a select field which was not specified in the list of valid values. it didn't get saved. So there already must be some validation checking going on somewhere ... [11:24]
pharveyyes
which sucks, we should go to an oops instead
(another silent failure!)
[11:28]
SvenDowideitwe should never goto oops :p
oh wow.
i restarted chrome, and one of the tabs was an edit url to a tmpl auth foswiki
it also has require auth
i now have a tinyMCE with .... the tmpl login UI in it
so its possible, but don't ask me how my 1.1.4 non-dev install managed it
[11:31]
...... (idle for 27mn)
pharveybtw: tiddlywiki, I love this way of editing pages [12:03]
SvenDowideiti don't like its way of viewing tho [12:04]
pharveyagree, viewing is odd
ah, that's right, SvenDowideit is forcing me to reveal my CodeMirror TML-mode scratchings
pharvey sees if it even runs
:P
[12:08]
SvenDowideitskewses
dammit, the commit that i thought i made that causes the double mess, isn't :(
[12:10]
pharveySvenDowideit: I wonder if the 401 error code will accidentally fix TinyMCE trying to make you edit the templatelogin screen
as it's really just editing what the rest handler gave it
[12:12]
SvenDowideitgobble.
so there's the porblem
i added rest to auhscripts
sadenin
though, edit is in authscripts too
so pass
[12:13]
pharveyyour browser opened up a cached copy of the edit screen
but really, what this actually points to is that I really need to stop talking about folding the unnecessary TML2HTML rest request into the edit page itself, and actually do it
[12:13]
SvenDowideitooo, maybe..
that might just be useful for what i wanted to play with one day :)
[12:15]
pharveydiesel powered leprechauns? [12:16]
SvenDowideitinline sectional edit of all 'blobs' that are identical between view html and edit html
presuming that there are no tinymce hints added for simple prose
[12:20]
pharveytinymce doesn't pollute much, but often 'cleans' it (in a mod_proxy_html kinda way) [12:22]
SvenDowideitsorry, i means wysiwyg rest hander hints put in
oh dear, so if my theory is right, saveTopic
the one that has a separate meta
also will extract META from text
[12:25]
pharveyit's not tinymce hints, but hints which preserve TML things which wouldn't otherwise survive in HTML
like newlines
[12:26]
SvenDowideitbut again, only in a half arsed way
SvenDowideit wonders how this ball of sand is held together
y, that i'd have to detect :)
but then, if we're sane, we'd encode those hints in a viewable way, and could have them in view html too
SvenDowideit is insane
[12:26]
***kip3f has left [12:31]
SvenDowideitso 1.0 removes FILEATTACHMENT meta that was in the topic text
1.1 leaves it in
i'm guessing leaving it in is actually more correct, as someone might be doing this before cp -r-ing the pub dir :/
SvenDowideit has a hack that fixes the duplicate TOPICINFO for saveTopicText, but
its rubbish
its very weird - $meta->text($text)
will just set the meta text to anything
but if you save, then read the text() again, any misplaced META will not have been escaped
[12:33]
pharvey:( [12:36]
SvenDowideiti don't really grok why CDot kept this weirdness for the newer API [12:36]
pharveypharvey wonders if this stuff is why renames are broken in mongo [12:36]
SvenDowideitSvenDowideit presumes that it was a choice :/
it could well be related yes
[12:36]
CDotSvenDowideit: I kept the misplaced meta because of demands from the choir [12:39]
SvenDowideiti'm thinking Func::saveTopic
but its all related to meta->text()
SvenDowideit resolves to ask CDot and pharvey to remind me to break it all in Meta2
[12:39]
CDotmisplaced META is *not* meta, it's text [12:39]
SvenDowideitno, its not [12:40]
CDotyes, it is [12:40]
SvenDowideitif i write
meta->text('%META:TOPICINFO.....
[12:40]
CDotthat's misplaced [12:40]
SvenDowideitthen onsave, that gets saved as META [12:40]
CDotthat is wrong, then [12:40]
SvenDowideitand when reloaded is not in meta->text() [12:40]
CDotthat is a bug [12:40]
SvenDowideitthen you need more unit tests [12:41]
CDotindeed
and a lot more time to write and run them. i know.
[12:41]
SvenDowideitas it appears that has been the case forever
ie, we can't call it a bug
but a miss-implementation of your prefered theory of what you thought you wrote :p
[12:41]
CDotnot sure what the difference between a "misimplemention" and a "bug" is, but I have a theory [12:42]
SvenDowideitsomething ala 'spec, by tch' [12:43]
CDotI am sure I did not want meta to be extracted from anything set via ->text() [12:43]
SvenDowideityeah, me too
sadly, doing so would require escaping ->text() onsave
or
some new guards
which is also a file format change
[12:43]
CDoty, I know. And that was a bridge too far. [12:44]
SvenDowideitso we were stuffed every which way
so atm, i'm going for adding a s/%META::TOPICINFO to Func::saveTopicText
[12:44]
CDotTalking of being stuffed, I need to get some veg planted or I won't have any for the summer. [12:45]
SvenDowideitas we throw away the value
ooo, nice
avagood day
[12:45]
CDoterm, META:TOPICINFO only works at the head of the topic (1st line) [12:45]
SvenDowideityes [12:45]
CDotat least, that was the intention [12:45]
SvenDowideitthe problem is because i added that [12:46]
CDotso %META:TOPICINFO is (in theory) safe embedded in text [12:46]
SvenDowideitthe text(%META:TOPICINFO is left in [12:46]
CDotMETA:FIELD, on the other hand.... [12:46]
SvenDowideiteven though it was coming in to a saveTopicText legacy [12:46]
CDotok, I see what's happening then [12:46]
SvenDowideity
its not the entire story
but its enough to fix _that_
the real story is that its all shocking 1999 hack-code
[12:46]
CDotthe entire story is "keep %META out of band" but that's another story [12:47]
SvenDowideitna, i don't beleive in meta oob [12:47]
CDoty. I must get round to either rewriting RcsLite or releasing VaxVMSStore [12:48]
SvenDowideitunless oob includes in the same file with guards
but its a pile of meh all round :)
[12:48]
CDotthe core *must not know* about META in text. Guards is fine, it just should be separate. [12:48]
SvenDowideitstore2 with JSON serialisation + woteva
y - except for small serialisation bits
enjoy your day :)
[12:49]
CDotenjoy your night :-) [12:50]
SvenDowideit:) [12:50]
pharvey is there any chance you can add a save button to the full screen tinymoce for 1.1.5?
SvenDowideit really wants his wysiwyg edits to be just that.
SvenDowideit goes to bed
[12:59]
.... (idle for 16mn)
pharveySvenDowideit: maybe not for 1.1.5, but some other patch release? We'll have to ask the RM :) (and get it working on trunk first)
g'night
[13:15]
........ (idle for 37mn)
MichaelDaumMichaelDaum reading apache manifold cf's handling of grant / deny tokens stored in the search index of solr ... https://cwiki.apache.org/confluence/display/CONNECTORS/FAQ [13:52]
.................... (idle for 1h39mn)
gac410ArthurClemens: Any idea why we don't get a subject Select box when editing an attached form to Support questions? See Foswiki:Support/Question1051 - I want to change the "Subject". [15:31]
FoswikiBothttp://foswiki.org/Support/Question1051 [ Displaying form at beginning of the topic ] [15:32]
ArthurClemenshttp://foswiki.org/bin/edit/Support/Question1052?t=1330529594;action=form shows a select box
so something has happened to the topic
[15:33]
gac410I don't get one. Subject ExtensionPublishPlugin
All I get is the | Subject | Using an Extension | And the selected subject is highlighted - (reverse vid). If I type into the field it jumps to other values
Ah... works on chrome, fails on Firefox
both on trunk and f.o
Running Firefox 9.0
And Konqueor shows a drop box - but notihing drops down, and the selected value is blank
[15:34]
.... (idle for 15mn)
Arthur - any idea if the broken drop-down on the support forms is a bug in DataForms, or just the particular form on our site? If it's in DataForms, it should probably be a 1.1.5 release blocker.
CDot - on Foswiki.org, it looks like your trunk documentation for DISPLAYIF made it onto Foswiki:System.DevelopingPlugins. It's showing up as a merge conflict when I try to upgrade to 1.1.5
[15:55]
FoswikiBothttp://foswiki.org/System.DevelopingPlugins. [15:57]
CDotgac410: ? DISPLAYIF is only on trunk [15:58]
gac410Do you want the trunk docs to remain on f.o, or should I just overwrite. [15:58]
CDotf.o docs should *not* be trunk
t.f.o should, however
[15:58]
gac410Right. But you edited the topic on f.o [15:58]
CDotdid I? Are you sure? [15:58]
gac410gac410 checks again. [15:58]
CDotCDot usually only edits in the checkout [15:59]
gac410Topic revision: r2 - 16 Feb 2012 - 09:38:56 - CrawfordCurrie [15:59]
CDotgood heavens
I may have been experimenting. Sorry about that.
please blow away my changes.
[15:59]
gac410yeah - I run a modified colas script to apply new versions, and it reported a local conflict. No problem I'll just say "Use new" when it updates.
ArthurClemens edited a couple as well, and I merged those into PatternSkin since they looked valid. and found some spam edits - Sandbox.WebSearch :-(
[16:00]
........ (idle for 38mn)
Any objections ... I'm going to install 1.1.5-Beta1 on foswiki.org ? [16:39]
.... (idle for 19mn)
okay - f.o is now running 1.1.5-Beta1 let me know if anyone sees any issues! [16:58]
.... (idle for 16mn)
SvenDowideit: I added a blog post about 1.1.5-Beta1 [17:14]
ArthurClemens: The issue with the Select ... needs a size > 1 or you can't navigate to select multiple. Doh... [17:28]
.... (idle for 19mn)
GithubBot[foswiki] foswiki pushed 1 new commit to Release01x01: http://git.io/iI8QyQ
[foswiki/Release01x01] Item11383: Note incompatibility - GeorgeClark
[17:47]
***GithubBot has left [17:47]
FoswikiBothttp://foswiki.org/Tasks/Item11383 [ Item11383: General documentation task for 1.1.5 release ] [17:47]
.... (idle for 18mn)
MichaelDaumgac410, we need a decision about blog.foswiki.org versions foswiki.org/Blog
^version^versus
[18:05]
gac410I do think it's better to use our own code for our blog. in meantime I should post to both I suppose. tbh forgot about it, and Blog is now on our top menu for f.o [18:06]
MichaelDaumhonestly, this BlogAddOn doesnt look professional enough. besides some main features of wordpress are desperately missing on foswiki.
these are essential to be visible with your blog out there
that aside I find it quite discouraging that people run in different directions without coordination
[18:07]
gac410gac410 doesn't have a strong opinion as I do very little with blogs. So I'm feeling more caught in the middle of other strong opinions. [18:08]
gac410 lost in wordpress. How do I make links [[System.ReleaseNotes01x01][Beta release notes]] doesn't work. Just type in raw html? [18:13]
MichaelDaumI gave Sven and Arthur a detailed summary why the current blog doesnt cut it: http://foswiki.org/Tasks/Item11285
they decided to ignore it
links: there's a wysiwyg dialog for it: mark some text - click on the link icon - sort of
[18:13]
gac410okay - added my links and published on blog.foswiki.org. BTW it appears that some of the issues in Item11285 have been addressed. [18:20]
FoswikiBothttp://foswiki.org/Tasks/Item11285 [ Item11285: Set up blog web ] [18:20]
MichaelDaumI've added some stronger words to this task item to make clear how I feel about this. [18:22]
gac410ArthurClemens: Found the multi+select bug. Firefox and Konqueor require the size be > 1, or they won't render a multi-select. Updated the forms and things are working. Added a note to the DataForms topic.
MichaelDaum: I think that the blog conversion falls into category of needing a Proposal. With a task list of what still needs to be done, and a plan to get it there.
The public commenting could be handled by the trunk CommentPlugin ... though we do need spam controls. trunk version can allow guest to comment.
Spam - need to get AntiWikiSpam plugin refreshed. The whole site could probably benefit from that anyway.
Need to get your PingBack plugin updated I assume.
[18:26]
MichaelDaumthe site needs a blog-like skin
the sidebar needs be released from non-blog related stuff
[18:30]
gac410gac410 has to defer to others on that comment - don't do enough to understand blog-like
Sidebar - that's easy enough I assume
[18:30]
MichaelDaumreadability is very poor on all of the site, even more when it comes to blogging
and then this is dog slow
[18:31]
gac410Hopefully your rewritten cache for 1.2 will fix the speed issues. Though without a trunk cross-connect we'll have issues. [18:33]
MichaelDaumgac410, I've replied to that on the page cache task [18:33]
gac410okay great I'll look [18:33]
MichaelDaumdid you read that? [18:33]
gac410not yet [18:33]
MichaelDaumthen there's rss feeds listing the full blog posting
the default foswiki rss feed simply doesnt cut it
the sidebar needs something like "Recent comments"
see also the "Archives" section on the wordpress blog.
these two arent doable easily on foswiki without serious performance problems
moderation of comments ... not there
[18:34]
gac410Archives? Add a formfield for "month" or issue? Then search? [18:36]
MichaelDaumerm, each topic has got a create date hasnt it
the date of rev1
[18:36]
ArthurClemenscan't we use BlogPlugin? [18:37]
gac410yes. Can we access it without performance hit? ... there might be a task to move it to %META
Okay - to help with a solution. Unless better ideas, I'll create a proposal to Convert public blog to Foswiki, and try to capture this stuff with details on how to solve it. Then we just need to pick through them.
[18:37]
ArthurClemensthere IS a proposal to convert blog to foswiki. It is just that MichaelDaum has specific needs that are currently not met. [18:40]
MichaelDaumArthurClemens, I'd love to. But once you give me that site, I wont stop at BlogPlugin and add more plugins that meet requirements better, imho. [18:41]
gac410Hm. MichaelDaum can't find the cache topic: Not listed under Foswiki:Tasks/FoswikiCache [18:41]
FoswikiBothttp://foswiki.org/Tasks/FoswikiCache [ FoswikiCache ] [18:41]
MichaelDaumeven the Tasks web stops working now [18:41]
ArthurClemensthat site = die Seite, or that website? [18:41]
MichaelDaumhttp://foswiki.org/Tasks/Item11549
web-site
ArthurClemens, it is not _me_ having specific needs. it is the project's blog that has specific needs.
any blog has got these needs.
[18:42]
gac410MichaelDaum: So if I understood your comments on 11549 - it sounds like it is doable. Excellent - we really need the boost on f.o - the search engines kill us. [18:44]
ArthurClemensit is not that foswiki is a blog or has a blog, rather we have a communication need [18:44]
MichaelDaumand? [18:44]
ArthurClemensif you throw Wordpress against it, we loose other capabilities, hindering other needs
like editing blog topics
[18:45]
MichaelDaumcould you elaborate, please [18:45]
gac410i think MichaelDaum has listed some very specific requirements for blogs (independent of f.o) which need to be addressed. Some sound pretty easy - others more complex. [18:45]
ArthurClemensthe best would be a foswiki blog, not a wordpress blog [18:46]
MichaelDaumArthurClemens, what do you mean by "like editing blog topics"? [18:46]
ArthurClemensI and other experience blog.foswiki.org as another site behind a wall [18:46]
MichaelDaumwhy? [18:46]
gac410ArthurClemens: where is the feature proposal. FeatureProposals page doesn't have the word blog found anywhere that I can find. [18:47]
MichaelDaumfor the records: I don't like wordpress either, nor would I use it for my blogging requirements. point is the current alternatives simply arent there [18:47]
ArthurClemensjust look at http://blog.foswiki.org/2012/02/foswiki-org-running-1-1-5-beta1/
it doesn't feel like foswiki at all
it is separate from the rest of the site
[18:48]
MichaelDaumerm, it isnt foswiki [18:48]
ArthurClemenshow do I edit a post? [18:49]
MichaelDaumbasically I totally agree with that. yet it is still a "tool" that we use right now.
ArthurClemens, log in and go to postings
[18:49]
gac410Lets keep "feel" and "features" separate. wordpress is what it is. How do we make foswiki.org/Blog acceptable. ignore the wordpress implementation. That will just "be there" until the blog requirements are addressed. [18:50]
MichaelDaumthe alternatives are still under construction: these are BlogAddOn vs BlogPlugin; then ClassificationPlugin vs Tags/TagMe/TopicClassificationPlugin; then PingBackPlugin; then FoswikiSiteSkin vs NatSkin;
next; KinoSearch vs SolrPlugin
[18:51]
gac410vs Mongo [18:52]
ArthurClemensI cannot log in there. it is a login loop [18:52]
MichaelDaumuser + passwd on private channel
Mongo isn't a fulltext search engine
is it?
[18:53]
gac410Don't know. Do the Kino & Solr address Meta acceleration? If we used MetaCommentPlugin ... as metadata, would that help implement the "recent comments" type features? [18:55]
MichaelDaumyes
it also allows to have an rss feed for all recent comments per topic and all of the web
not there with CommentsPlugin
DBCachePlugin will enable navigation to the next blog posting
SolrPlugin + ClassificationPlugin will do "Similar postings" for the sidebar
[18:55]
gac410trunk CommentsPlugin seems much closer to MetaComments plugin ... wondering if possible merge should be considered. [18:56]
MichaelDaumDBCachePlugin or SolrPlugin will enable a well performing "Archives" sidebar section
NatSkin comes with a decent blog-like css
[18:56]
gac410So it sounds like we have semi-agreement actually. It's a matter of figuring out how. We have the technology :-) (Maybe too many options) [18:58]
ArthurClemensour real currency is time [19:05]
gac410Anyway ArthurClemens - I can't find a proposal where features and design and requirements of a blog are covered. actually the word Blog doesn't appear in the Development web [19:14]
ArthurClemensnot a proposal. http://foswiki.org/Tasks/Item11285 [19:16]
MichaelDaumdid anybody test the new page cache on trunk? [19:18]
GithubBot[foswiki] foswiki pushed 6 new commits to master: http://git.io/cCXaDg
[foswiki/master] Item8402: recent changes - MichaelDaum
[foswiki/master] Item11583: modernizing - MichaelDaum
[foswiki/master] Item11583: ported to foswiki, some modernizing - MichaelDaum
[19:18]
***GithubBot has left [19:18]
MichaelDaumbesides there are some monster topics in System which are able to bring down f.o all by their own [19:18]
FoswikiBothttp://foswiki.org/Tasks/Item8402 [ Item8402: port BlogPlugin ]
http://foswiki.org/Tasks/Item11583 [ Item11583: modernize xmlrpc and pingback ]
[19:19]
gac410MichaelDaum: in System web? Regarding new cache - I'm not sure the server has the database prereq. Probably need Babar's help.
I know that trac has had all sorts of issues with insufficient connections to the sql backend - so have to proceed with caution.
[19:20]
MichaelDauminstalling sqlite or even mysql should be doable
gac410, CompleteDocumentation is a monster topic
[19:20]
gac410It is *BSD ... free or net, running in a jail I can't recall. but my linux skills are not particularly transferable.
yeah - that one is indeed a monster.
[19:21]
MichaelDaumit should be hidden from crawlers [19:23]
gac410Don't know how to do that. [19:23]
MichaelDaumin robots.txt [19:23]
gac410Ah. [19:23]
MichaelDaumsimilarly all WebIndex should be listed in robots. they are expensive as well.
infact all WebIndexes should go away in favor of a proper WebTopicList
[19:24]
gac410Disallow: /System/CompleteDocumentation
Disallow: /bin/view/System/CompleteDocumentation
[19:25]
MichaelDaumnuf for today. I think I need a rest. [19:26]
..... (idle for 21mn)
Babarwe have pgsql running on the host
no mysl iirc
and sqlite should be there too
[19:47]
gac410: how could the log rotation bug be fixed?
we didn't code anything, right? :)
[19:55]
..... (idle for 20mn)
Rich_MorinI went to Jeffrey Thalheimer's talk last night (on Pinto). He seems to have put a lot of good thought and effort into the project and it might well be handy for FW development.
He says he would be happy to have folks try it out, send feedback, etc.
Basically, it's a system for managing multiple local PANs (eg, DarkPANs).
[20:15]
.... (idle for 16mn)
gac410Babar - It was a pending fix. already committed after the Jan 1st rollover
gac410 can't remember when he fixed something and is doomed to open duplicate tasks.
[20:32]
In FeatureProposals - anyone know what "| WhatDoesItAffect: | %WHATDOESITAFFECT% |" is all about. Does it actually work? [20:43]
.............. (idle for 1h5mn)
I've started Foswiki:Development/MakeOfficialFoswikiBlogOnFoswiki intended to set the threshold for what needs to be done before the Foswiki blog is transferred from WP to Foswiki. [21:48]
FoswikiBothttp://foswiki.org/Development/MakeOfficialFoswikiBlogOnFoswiki [ MakeOfficialFoswikiBlogOnFoswiki ] [21:48]
gac410Mostly organizing and re-hashing MichaelDaums comments on Item11285 ... We need to identify what are the must-have features and how they can be gotten. [21:50]
FoswikiBothttp://foswiki.org/Tasks/Item11285 [ Item11285: Set up blog web ] [21:50]
........ (idle for 39mn)
BabarBabar could flame on this post, but not in the mood [22:29]
gac410which post? mine? [22:30]
Babarthe task
like... why do we care? Wordpress might be great, but to host one post a year, I'm not sure technology matters much
we already had this discussion several years ago
[22:30]
gac410Michael has some rather concrete objections. So lets identify which ones really are needed and move forward fixing / implementing them. He has a point esp. about the feed automation. If we can become better for blog usage we benefit as do users. [22:33]
Babarbloggers always argued that some things are mandatory for a blog: pingbacks, trackbacks and RSS
not sure we have any of those working properly for the tools out there :)
[22:37]
gac410right. So lets get them working, or at least evaluate what it requires. There are lots of pieces there, If nothing else it helps document to others how to use Foswiki in the blogging space. [22:43]
Babartrue [22:56]
.... (idle for 16mn)
gac410babar, is webmaster@foswiki.org a valid email address? Noticed in the error_log >>>> FAILURE Sending e-mail to Wiki Administrator <webmaster@foswiki.org> - ERROR: Exit code 4390912 (17152) from Foswiki::cfg{MailProgram}
I was able to successfully register though, so it's not too serious.
[23:12]

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