#foswiki 2012-11-14,Wed

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

WhoWhatWhen
***FoswikiBot has quit IRC (Ping timeout: 245 seconds) [00:03]
gac410hm... foswiki.org seems to be down ... maybe routing? tracepath fails [00:10]
pharveyoooh... no ping from me
what's the difference between tracepath vs traceroute?
[00:11]
gac410probably some internals. more modern, not completely sure. [00:12]
pharveyouch. So do we ping gmc && ping foswiki.org? :) [00:13]
gac410tracepath doesn't need superuser auth to run.
for me, tracepath stops at 194.109.5.1 ... Unless that's part of gmc's network, it's probably out of his hands.
[00:15]
pharveythat's where it stops for me too
perhaps xs4all has been shut down again
I say "again", but... it was many years ago wasn't it
[00:20]
gac410no idea.
xs4all appears up.
[00:22]
pharveyindeed [00:24]
gac410http://www.isitdownrightnow.com/foswiki.org.html claims down for everyone. [00:26]
kip3fdown for me too [00:26]
gac410kip3f: Ah ha... you are back. After all day spent trying to come up with a version string solution that doesn't need version 0.77 .... we are back to needing it. [00:27]
kip3fgac410 - i have more version fun for you :-(
in my development foswiki on my mac
[00:27]
gac410Our conclusion is that there really is no good way to get to real version strings without reqiring version. [00:28]
kip3fit stopped working when I set FOSWIKI_ASSERT [00:28]
gac410What was the ASSERT [00:28]
pharvey.77 is in perl 5.10 I think? [00:28]
gac4105.10.1 [00:29]
kip3fkip3f is checking [00:29]
gac410I run with that enabled all the time. [00:29]
SvenDowideitmmm, configure should now error for Class::Load
SvenDowideit used it in Foswiki.pm
guess thats my faule
[00:30]
gac410I don't think Configure actually loads Foswiki.pm any more. [00:30]
SvenDowideitconfigure says nothing, but going to bin/view dies
SvenDowideit recons its sven's fault
[00:31]
gac410Ah hm... Oh... Needs to be in DEPENDENCIES file. Hm. and we ship it in lib/CPAN/lib [00:31]
SvenDowideity, i intentionally don't use it
that way i know when i stuffed up
and omg, the deps tree for Class::Load is craped too
[00:31]
kip3fFoswiki.pm: v-string in use/require non-portable at /Users/kip/dev/www/src/foswiki/trunk/core/lib/Foswiki.pm line 55
perl 5.10.0
what's this, then?
perl -V gives:
Locally applied patches:
/Library/Perl/Updates/<version> comes before system perl directories
installprivlib and installarchlib points to the Updates directory
[00:32]
gac410I'm trying to find out. It seems to be a warning that has been eliminated in some perls. I've changed the file so many times, what exactly is in your Foswiki.pm line 55 [00:35]
kip3foh, nothing to do with version.pm
n/m
ah sorry
require v5.8.8; # see http://foswiki.org/Development/RequirePerl588
[00:35]
gac410hm. cool. Now THAT's strange. [00:36]
SvenDowideitoh crud, dreamhost's not using bash anymore? /me needs a new brain, this one's full of old unused data [00:38]
kip3fError goes away if I change line 55 to
require 5.008_008;
see http://www.nntp.perl.org/group/perl.perl5.porters/2007/10/msg130062.html
[00:39]
gac4105.10.0 is a rather unusual perl version. does that ship standard on your mac, or something local. [00:41]
kip3fstandard. I have Snow Leopard
I found a nicer solution that doesn't give a warning:
no warnings;
require v5.8.8; # see http://foswiki.org/Development/RequirePerl588
use warnings;
[00:41]
gac410Did you install version 0.77? [00:42]
kip3fno, I got the latest from CPAN [00:43]
gac410okay. good. or not. That warning comes from version, not from perl. [00:43]
kip3fdo you know how I can find out the installed version of a module from the command line? [00:43]
pharveyI think I usually do something like perl -MFoo -e 'print $Foo::VERSION'
there's probably a better way
perl -Mversion -e 'print $version::VERSION' shows 0.88 on my PC
[00:45]
kip3f0.9901
for me
[00:46]
gac410Supposedly the sequence:
use 5.006; # First version to accept v-numbers.
require v5.8.8; # see http://foswiki.org/Development/RequirePerl588
[00:47]
SvenDowideitoh crap. [00:47]
gac410should prevent the problem. [00:47]
SvenDowideitClass::Load is a problem on this hoster [00:47]
kip3fpharvey: what do you get for
perl -w -Mversion -e 'print "$version::VERSION\n"; require v5.8.8;'
[00:47]
pharveykip3f: just 0.88 on my deb testing PC
0.9 on Ubuntu 10.04.4 LTS where I have perl -v = 5.12.4 (hot-wired)...
[00:49]
kip3fI bet that the require directive does something like parse the v string directly [00:49]
pharvey0.77 on a 10.04.4 LTS where I have perl -v = 5.10.1 [00:50]
kip3faha
yes so it must be limited to 5.10.0
[00:50]
gac410works for me on 5.8.4, 5.8.8, 5.10.1 [00:50]
pharveyI don't have a machine with 5.10.0 AFAICT
oh, I do
perl -v = v5.10.0
it shows 0.74
no warnings
[00:50]
kip3fum [00:51]
pharveyI used perl -w -Mversion -e 'print "$version::VERSION\n"; require v5.8.8;' [00:51]
kip3flet me try again on mine, without the newer version.pm in the path [00:52]
SvenDowideitgac410,
Could not find a suitable Class::Load implementation: Can't locate Class/Load/XS.pm in @INC (@INC contains: /home/uwiki/foswiki/core/lib . /etc/perl /usr/local/lib/perl/5.10.1 /usr/local/share/perl/5.10.1 /usr/lib/perl5 /usr/share/perl5 /usr/lib/perl/5.10 /usr/share/perl/5.10 /usr/local/lib/site_perl /home/uwiki/foswiki/core/lib/CPAN/lib/arch /home/uwiki/foswiki/core/lib/CPAN/lib/5.10.1/x86_64-linux-gnu-thread-multi /home/uwiki/
foswiki/core/lib/CPAN/lib/5.10.1 /home/uwiki/foswiki/core/lib/CPAN/lib) at /home/uwiki/foswiki/core/lib/CPAN/lib/Module/Runtime.pm line 317.
Package::Stash version 0.14 required--this is only version 0.05 at /home/uwiki/foswiki/core/lib/CPAN/lib/Class/Load/PP.pm line 9.
BEGIN failed--compilation aborted at /home/uwiki/foswiki/core/lib/CPAN/lib/Class/Load/PP.pm line 9.
Compilation failed in require at /home/uwiki/foswiki/core/lib/CPAN/lib/Module/Runtime.pm line 317.
at /home/uwiki/foswiki/core/lib/CPAN/lib/Class/Load.pm line 24
and annoyingly, i can install a local Class::Load, but not Class:load:XS - as that tries to write to the wrong dir
[00:52]
gac410in a word .. crap [00:52]
SvenDowideiti'll update Package::Stash, but this is complicated [00:53]
kip3f0.74
v-string in use/require non-portable at -e line 1.
[00:53]
pharveyso "require v5.8.8" works for me on 5.14.2, 5.12.4, 5.10.1 & 5.10.0 [00:53]
SvenDowideitmumble - and that says its up to date too
grrr
[00:53]
pharveymy perl 5.10.0 is $version::VERSION = 0.74 as well
so kip3f something is weird :)
[00:54]
SvenDowideit$CLANBASE - not quite what i had in mind [00:54]
pharveymy 5.10.0 is Ubuntu 9.04 [00:54]
kip3fhttp://pastebin.com/wqyMtNGn
how does your perl -V compare
[00:55]
SvenDowideityou have a 5.10.0? wow [00:55]
pharveynot me, just an old VM
oh right, kip3f's mac
[00:55]
kip3fSven - it is shipped with OSX snow leopard [00:55]
pharveythat is saddening [00:55]
SvenDowideitkip3f, oh that terror
not only ancient and buggy, but compiled with incorrect compile options
apple do not love you
[00:55]
gac410Well the safest would be to use 5.008_008
Though I suspect there are "use v5.8.8" elsewhere in the code.
[00:56]
kip3fyeah, it had the architecture wrong
FML
[00:56]
SvenDowideithow many hidden dependencies are there in this CLass::Load mess :( [00:56]
pharveySnow Leopard released Aug 28, 2009 [00:56]
gac410when I installed Log::Disppatch with cpanm, it loaded a bunch of stuff,. [00:57]
SvenDowideiti'm doing this on dreamhost [00:57]
pharveyperl 5.10.0 released 2007-Dec-18 [00:57]
SvenDowideitusing a local user cpan
not cpanm
and i'm getting lots of failures
Can't locate Module/Runtime.pm
the latest
[00:57]
pharveyTo be fair perl 5.10.1 was only released 6 days prior to Snow Leopard, 2009-Aug-22 [00:58]
SvenDowideitand to be fair to users, they send updates magically more than once a month [00:58]
kip3fseriously, are there any foswikis running on OSX? [00:59]
SvenDowideityes, there were
i used to build installers
but OSX is so poorly done
the official Perl advice is to install a real perl
same advice that was always given to Solaris users
[00:59]
kip3fI wish that I had done that
right away
[00:59]
gac410Sven... if configure had a button that could run "cpanm -L lib/CPAN/lib Some::Dependency" I wonder if that would help [01:00]
kip3fI had some good times, compiling mysql libs [01:00]
gac410I'm very impressed with perlbrew [01:00]
SvenDowideityeah - i stopped building and supporting osx pretty much because apple isn't interested
gac410, no, it would not
as i, and others have existing infrastructure
[01:00]
gac410So is LogDispatch DOA for a core logger? [01:01]
SvenDowideiton the other hand, maybe it would help alot
its much worse than that
Foswiki.pm uses Class::Load
[01:01]
gac410yeah, but only because LogDispatch added it. :P [01:02]
SvenDowideitas i jumped at the chance to use something less crap than eval 'require...
SvenDowideit is going to ignore this a little til we have more ideas
might try cpanm :/
no, even that makes no sense yet
as just using lib/CPAN/lib falls over on that host
for some reason it needs a newer Package::Stash, and then can't find the XS version
so the PP isn't used
[01:02]
gac410Do you have lib/cpan/lib at the end of libpath or the start? [01:05]
SvenDowideittbh, i can't work that shit out
there is no docco telling me about where it goes in LocalLib.cfg
[01:05]
gac410So the end. hang on, it's a simple tweak to LocalLIb.cfg [01:06]
SvenDowideitthere are thousands of words, and none tell me that
(yes, it always seems to end up at the end)
[01:06]
gac410$CPANBASE = ''; # Uncommented: Default path prepended
Just uncomment that line, and it becomes prepended instead of appended.
[01:07]
SvenDowideitthat comment is extremely opaque
which 'it'?
prepended to what
[01:07]
gac410The libpath [01:07]
SvenDowideitso you say if i set CPANBASE to ''
then '' is prepended to the PERL5LIB?
[01:08]
gac410yes [01:08]
SvenDowideitthat makes no sense to me [01:08]
gac410# if the setting is uncommented but left empty (undefined) then the default path
# will be prepended to the "@INC library path" while by default it is appended.
[01:08]
SvenDowideitwhy would i want to prepend '' to the path?
and how does that get lib/CPAN/lib into PERL5LIB
[01:08]
gac410So it uses your local version rather than trying the hosts version when the module is in two places [01:08]
SvenDowideitwhat local version? [01:09]
gac410If you have blah v2 installed in lib/cpan/lib, and site has blah v1 in @INC somewhere. By uncommenting that line, blah v2 will be found first. [01:10]
SvenDowideithow? [01:10]
gac410magic [01:10]
SvenDowideithow does it know where my lib/cpan/lib is?
let alone that i don't understand what you're saying
SvenDowideit may give up on using LocalLib again, as i've never understood it, and the whole 'magic' thing isn't what i want :/
[01:10]
gac410Foswiki lib/cpan/lib, The lib/cpan/lib we ship. setlib.cfg appends it. If you uncomment the line in LocalLib.cfg, setlib prepends it. [01:11]
SvenDowideitso why are there so many examples in LocalLib.cfg where i need to tell it where lib/CPAN/lib is
if it can work it out by itself
[01:12]
gac410It's telling you how to alter what it does automagically. [01:12]
SvenDowideitseriously, i am failing to understand what it does all over :/
so much so that i can't even try to re-write it
looking in setlib.cfg
it looks like if i set @CPANBASE=''
then it _won't_ magically find lib/CPAN/lib
$CPANBASE = "$foswikiLibPath/CPAN/lib" unless $CPANBASE;
[01:13]
gac410It works fine. It uses the standard perl @INC. It appends Foswiki/lib/CPAN/lib to the end. All you should ever have to touch is in LocalLib.cfg [01:14]
SvenDowideitoh '' is false
i have no idea if 'it works fine' as i can't understand what its supposed to do
whereas if i replace setlib.cfg with some hardcoded paths
then wouldn't i know what it does?
[01:14]
gac410Yes, but the rest of us might not want your hardcoded paths. [01:16]
SvenDowideit(i think magic is useful, but when there is no non-magic route, and i, as a developer have no idea what that magic does
then i'm very much in the dark
and seriously, i'm not being intentionally painful - h have zero idea what each combination of each of the things in LocalLib.cfg result in
but it looks like the different combinations are very deeply magical
[01:16]
gac410actually they are just too wordy. There are only 3 options. [01:18]
SvenDowideitits almost like the CPANBASE setting coule be a bool - prepend/append
and then we could tell users to add paths to the localLib array
yes, it is very wordy, but never actually tells me what will happen
except to someone that already knows
[01:18]
gac410CPANBASE = /svens/lib it gets appended CPANBASE = '', Foswiki/lib/cpan/lib gets prepended [01:19]
SvenDowideity - that has insuffient information
and is confused by the text that preceeds it
for eg
# $CPANBASE = '/path/to/dir'; # Keep commented: Default path appended
the comment reads like an instruction
telling me that i need to keep that commented
[01:19]
gac410It should probably say "IF commented - default Foswiki/lib/CPAN/lib appended [01:21]
SvenDowideitthat is a little better [01:21]
gac410If you needs something completely different, @localPerlLibPath is the way. [01:21]
SvenDowideiti don't have the first clue if i need something different
as i don't undestand what this stuff does
or (more importantly) what its supposed to do
(practically - being told it magically XYZ isn't enough for me)
[01:21]
gac410Sven, I've never used any of it. I'm reading. The magic comment was a joke. [01:22]
SvenDowideitargh!
who the **** has used it!
[01:22]
gac410The only thing I ALWAYS do is un-comment the last line, so my testing is done with lib/cpan/lib versions we ship' [01:23]
SvenDowideitcos its written by the implementor, with implementor assumptions
so the default is very busted now
setting to '' stops it crashing
[01:23]
gac410no... The default is to prefer the System provided @INC libs over the stuff we ship, which is often quite old. We never update them. [01:24]
SvenDowideityes
and that is now broken
as using an old Package::Stash with the shipped Class::Load
[01:24]
gac410right, because we ship a version newer than what your host provides. [01:25]
SvenDowideitmakes it fail to load the PP.pm version
wich then makes it die loading the missing XS
because we magically use out Class::Load
[01:25]
gac410We use ours only because the local site doesn't provide it. [01:26]
SvenDowideiti grok that mech, but i'm basically saying that what we've been doing probly won't be a good idea
and having a cpanm -L in configure won't fix that
[01:26]
gac410It's very difficult to be best for all uses. For non-hosted sites, our default is probably the better, so the local site can install updated versions of packages. [01:27]
SvenDowideitunless we prepend lib/CPAN too
i recon that might be wrong
as people install foswiki, but don't update theur load packages
[01:27]
kip3fit would be nice to give the ability to prepend packages [01:28]
gac410So another option is to not ship CPAN libs at all ... except maybe to make configure run. [01:28]
kip3fI mean, add lib/CPAN/lib to the end of the path [01:28]
SvenDowideitits there - but the docco didn't make sense to me
kip3f, it defaults to that
[01:28]
kip3fand then give the ability to selectively prepend lib/CPAN/Class-Load [01:28]
SvenDowideitthat won't help [01:29]
kip3fi.e. put each module in a subdir of CPAN [01:29]
SvenDowideitas there are Class::Load deps that are in system, but too old for Class::Load
fundamnetally, the problem is that Class::Load is a heck of a lot more complicated than what we previsouly put into lib/CPAN/li
[01:29]
gac410Shipping our own CPAN is really a bad idea. [01:30]
SvenDowideitand so the assumptions we've made over the years gets much more complicagted
yes, it is
but forcing manager downloaders to know about installing cpan packages is considered worse
i'm not sure what the right answer is
[01:30]
gac4101) sites using deb / rpm don't want it. 2) We are distributing someone elses code, often outdated. [01:31]
SvenDowideiti gork that, always have
the pain is that new users trying it out generally want php-simplisity
so unzip, try
and for some reason they don't just use the foswiki on a usb stick
course, i don't know how valid this assumption is anymore
[01:31]
gac410So we either start to reign in dependencies ... or figure out a better solution. Koen's dependency_installer was a good idea, but way too big. [01:32]
SvenDowideityour configure cpanm idea needs the lib/CPAN infront?
and how does cpanm get into the system
[01:33]
gac410We would have to ship it. It is very small. [01:33]
SvenDowideitand iirc is getting deprecated ? [01:33]
gac410cpanm? not that I've heard. cpanplus I think is getting deprecated [01:34]
SvenDowideity, it is
hehe
[01:34]
gac410curl -L http://cpanmin.us | perl - App::cpanminus -L blah4 will bootstrap cpanm into the blah4 cdirectory. [01:36]
SvenDowideitrighto
still - the main users of lib/CPAN/lib that we're hurting
are those with no internet access on their foswiki server
like the person sitting next to me
[01:37]
kip3f^^ [01:37]
SvenDowideityou know - enterprises that restrict internet access to mission critical infrastructure
i'm glad i have one of those users locally, to remind me what horrible setups are like
[01:38]
gac410Right. So we either distribute a prebuilt lib/cpan/lib which has the XS issues, or ship a tarball archive ... or a script to build a lib/CPAN/lib on a system that has access. [01:39]
SvenDowideitthe one we have in trunk does not have XS issues [01:39]
gac410Well. I had to play games to avoid the XS stuff. [01:39]
SvenDowideitit has a major crash becuase the CPAN/lib is appended
and the symptom is the XS thing
but the cause is that we're only using missing modules from lib/CPAN/lib, not any that is more up to date
(which is how the PERL5LIB works, but thats the cause
[01:39]
gac410But at the same time, I intentionally did not include the XS version of the package. I had to delete the xs version and manually copy over the pure perl version.
cpanm built xs, and omitted the pureperl from the lib.
[01:40]
SvenDowideitso one thing we can do, is argh. [01:41]
gac410It's preferable for performance purposes is to build the xs. [01:41]
SvenDowideiti do not like the deffered checkers in configure
i see zero errors and warngins, because i've not clicked some random number of green buttons manually yet
even with the xs version
it would die
as it picks up the other system modules that are older
[01:41]
gac410Right. I understand.
Lib order is one issue. Optimized is another.
[01:42]
SvenDowideithow do i 'run all checkers'
tbh, that change has made configure _less_ useful to me
[01:43]
gac410hm. I guesw we need to add it. [01:43]
SvenDowideitas i can't goto configure and see whats wrong with my setup
without doing some other random thing i don't know i need to do
and i can't tell the difference between a green button that means 'run again' and one that means 'you have to click me to run it'
SvenDowideit is feeling the complexity today, is isn't coping
[01:43]
kip3fspeaking about configure [01:45]
SvenDowideitas i sat down to write slides for me talk [01:45]
kip3fas foswiki is still down [01:45]
SvenDowideitgmc, ping? [01:46]
kip3fhere is my update to ExposeDeveloperSettings
http://pastebin.com/eMKyHj3a
The new bits are at line 61
---+++ Proposal - scan entire lib directory, following System.PerlDoc approach
[01:46]
gac410Back to version string briefly. So Kip3f, require 5.008_008; would fix up your issues. since I can't check anything in. :( [01:51]
SvenDowideitOK, yay [01:52]
gac410Disabling warnings is probably not the right solution. The typ. user though will not get a failure. [01:52]
SvenDowideitso if i set CPANPATH='' and then use my local cpan dir, it actually works as i want
progress
[01:52]
kip3fno warnings; use v5.8.8; use warnings;
s/use v/require v/
gac410 - why can't you check in?
ah
n/m
[01:53]
SvenDowideitmmm, thats worrying
X-Foswiki-Monitor-Rendertime: 0.095858
with page cache off
X-Foswiki-Monitor-Rendertime: 0.052652
with it on, but the html is not displayed in chrome
plus the other js magic this site has works when the cache is off
argh!
what fun
if i set the scripturl{view} incorrectly - to '/'
then eventually rather than going to http://example.com/System
it goes to http://example/System
and i thought i'd eliminated that bug
[02:07]
gac410Sven, most of the checkers that Timothe ajaxed were very long running, The file path checkers run forever on some systems. I've seen 30 seconds or longer.
Basic errors are all checked every run and were not converted to ajax. You would not want to run ALL the ajax checkers, as they will generate spam email for no reason, and eventually might include "fixup" buttons.
[02:21]
SvenDowideitum, yes, i do want to run all checkers
not necessarily by default
and i'm quite happy for that to result in a re-confirmation of the email system too
[02:22]
gac410I was thinking of adding an "Audit Mapper" button to verify the .htpasswd file against the WikiUsers topic. .. maybe tie it in with AllowLoginName. Since when that is enabled, a damaged mapper topic will really break things. [02:23]
SvenDowideita 'checker' should check things
not do stuff?
sounds like we need to separate checking from doing automation
_I_ want to run all the checkers - where checker is the set that mean what the name means
if that makes sense
[02:24]
gac410yes. [02:25]
SvenDowideitits not my fault that damned programmers don't use english :p
mmm, i get the feeling that i should just go for a bike ride for the rest of the day to get my smile back
[02:25]
gac410damn... I can't explain it without reading foswiki.org . but a Checker now has both check methods (runs everytime) and the ajax stuff that can run on demand.
We now have a balancing job to figure out which parts to run everytime and what to defer.
[02:26]
SvenDowideitand need a button for an admin that wants to know the answer to 'what might be busted'
ie, check everything on demand
[02:27]
gac410A lot is even now in the Config.spec. No need to even write a checker, vs. rules on the spec definition.
right. He is looking into that.
[02:28]
SvenDowideitare you confusing me with impolementation details? :) [02:28]
gac410Well it's the implementation that decides if it is every run, or on change, or on demand.
You have the power
[02:28]
SvenDowideitand i demand to have a button that runs all ondemand checkers
so ! :)
[02:30]
gac410In anatomy of a checker. sub check. Run on every load. sub provide_feedback. Run when button pushed. sub check can run the provide_feedback routine. so you have initial test, and button to re-test.
provide_feedback will run against UNSAVED settings. So you can check before save.
[02:36]
SvenDowideitso you're saying that when i load configure all checkers are already run? [02:37]
gac410Finally some fields can be set to auto, so just changing the field and moving focus re-checks it. Change a regex. [02:37]
SvenDowideiteven tho the UI tells me some have not
like the Environment one?
in 1.1, if my foswiki is broken
first thing i do is goto configure and see what its warning/erroring about
in 1.2, i can't do that, as it looks like some checkers are not run until i click on some random hidden stuff
[02:37]
gac410Okay, no. In his judgment, the "long running checkers" are deferred to on demand. [02:39]
SvenDowideitand as the UI hides the buttons from me [02:39]
gac410I needs an example [02:39]
SvenDowideiti don't have the first clue what, or how many to click
that was the eg
[02:39]
gac410AFAIK the ones he deferred are the env, and the file permissions checkers. [02:40]
SvenDowideitbut how the heck would a user know that
or even have confidence that some other checker isn't hidden
[02:40]
gac410relax.... jeeze... it's his first pass. [02:40]
SvenDowideitthis is my first pass trying to think through the UI
and its very much written for someone that knows how the code works
which by itself is worrying
given that the more you knwo about the internals, the less you use configure
[02:40]
gac410It's more written so it doesn't take 30-60 seconds to get a configure screen for every transaction.... which is what I see on my windows vm due to poor file system performance. [02:42]
SvenDowideitthats not a UX thing [02:43]
gac410It's the checking of permission on 20,000 files for every enter. [02:43]
SvenDowideitSvenDowideit gives up - you do actually grok what i'm saying [02:43]
gac410yes. I already suggested an Audit button and he is pondering it. [02:43]
SvenDowideitand everything else is _not the level of abstration i'm at right now_
worse, I'm now re-writing the slideshowplugin topic :/
[02:44]
gac410I thought Arthur rewrote it for 1.2 Lavr had reverted it on 1.1 at some point. [02:45]
SvenDowideitthe docco?
ha!
SvenDowideit can't cope with the 1999 style docco that says 'cut and paste XYZ' into a new topic
and is re-discovering the fascinating INCLUDE override= param
[02:45]
gac410So I've checked over 1.2 configure. There are deferred checks in 3 places. The evironment, The paths (which are very long running), and the send email [02:56]
SvenDowideitplus any plugins that do it in future [02:56]
gac410Send email is a good example of check AND feedback. The checker validates that the address is a properly formed email address. That happens on every run. The deferred check sends an email. And thankfully that should not run on every enter. [02:57]
SvenDowideitso a button that says 'run N defered checkers now' would give me the confidence that someone had thought it through
and i'd be ok for that to send me an email
especially if (as a bonus) the email checker actually mailed me about the state of my foswiki :)
[02:57]
gac410:P [02:58]
SvenDowideitmanoman, we really could do with a release where we all just work on the distributed topics
we have so many cool features we don't use in them
[02:58]
gac410We could use another 20 deveopers [02:59]
SvenDowideitna, we could do with cleaning up what we have
which is happening in configure right now
but the topics..... gads
its like smilies plugin - it aught to use the new ICON code, but doesn't
in 1.2.0, its increadibly trivial to make a create new topic UI
but we still have mostly docco that says 'cut&paste into a new topic'
[02:59]
gac410pharvey... did you reach anyone ? I have not seen anything in the infra mailing list. [03:05]
kip3fand the distributed topics are the handiest examples [03:05]
SvenDowideity
so for eg
%INCLUDE{ "%SYSTEMWEB%.WebCreateNewTopicComponents" section="newwebtopic"}%
is the easiest way to get a UI that will let the user select a web and type a topic anme
and then click 'create'
then with a few extra params, set the form and topic template to create that topic
(ok, so section="newtopic" exists, i've just create newwebtopic
oh kip3f thankyou for the bug reports
they are appreciated, tho atm, i'm poking other things
including poor gac410 :)
[03:05]
gac410need to poke gmc ... still down.. :( [03:11]
SvenDowideithe'll likely be alseep
and worse, probly has his email served byt the same datacenter :)
[03:11]
gac410Are we officially hosted. .. ie could open a support ticket? [03:12]
SvenDowideitSvenDowideit knows nothing
mmm, ouch, AutoINC does not play nicely with natedit
mmm, i may have reported that release blocker before
[03:12]
gac410title AUTOINC0000 [03:13]
SvenDowideitnot just title
mmm, but yes, i did
now how to gel the slideshow docco that says
---++
headings are used to delineate slides
when the user get wysiwyg edit :/
[03:13]
pharveygac410: just sent it, sorry. got a phone call. [03:19]
gac410np. I just got it as well. Was more concerned that it was moderated ... or running on the same server [03:19]
SvenDowideitoh.
natedit doesn't obey templatetopic?
mmm, or do i have a cache problem
yay, a cache somewhere
[03:24]
oh FK. and then it uses the word 'template' in the skin context, and i want to add a topic template
SvenDowideit kicks the table
[03:33]
gac410Well tis good you are finding this stuff before 1.2.0 rather than after. [03:34]
SvenDowideitits amongst the list of things i've had on the 'before release' for um, as long as 1.1.0 has been out
which was a reason i avoided being RM :)
[03:35]
argh!
copy the "Default Slide Template" section below (or from [[SlideShowPluginTableTemplate]]) into your own slide template topic
and worse, this is a totally non-foswiki use of the word tempalte
eh what? 'Just customize the attached =slideshow.css= file to improve the view.'
and guess what there isn't
[03:40]
gac410This is the 1.2 version? [03:42]
SvenDowideitcourse
i'm resigned to using today as 'i'm just trying to use your freaking product' testing
argh, yup, lots of hardcoded html and stuff in the perl module
so the use of 'just' means, yes, this is painful, we're telling you to modify a file attached to a System web topic
[03:42]
...... (idle for 28mn)
ok, next confusion - why does the System.IfStatements topic exist
why not docco that in VarIF ?
does it make sense to separate them?
[04:13]
gac410I never remember which one to use. [04:14]
SvenDowideiti'm pretty much trained to not look at the VarIF and VarQUERY and VarSEARCH topics [04:15]
gac410Really we need a good documentalist to go through and clean up the docs. [04:15]
SvenDowideitbecause the content i want is distributed in several other topics
really, we need to spend time not coding perl and js, and relish in how many features we have
[04:15]
gac410coders do what coders do [04:16]
SvenDowideitand then to use that to polish some features that already mostly exist
mm, thats fatalistic
booooooo !
oh huh?
i goto the plugin topic, and it magically moved my window below the TOC
thus making the foswiki look slightly odd
i wonder why and where that happens
[04:16]
.... (idle for 15mn)
really?
cool! it seemed to me pretty ok
makes me more glad to have added it :)
[04:35]
gac410yeah. Crashes all over with various uninit varialbes.
Did you ?
[04:35]
SvenDowideitneed to add it to suite tho :)
y, its in the TestCases web
[04:35]
gac410Ah.. not in my checkout [04:36]
SvenDowideittrunk :) [04:36]
gac410yup. I'm on trunk
Though I can't svn up. :)
[04:36]
SvenDowideitmmm, i was pretty sure i did it a while back [04:37]
gac410I just copied them over from my svn checkout of t* [04:37]
SvenDowideity, i may have changed a few things
oh, i did
i added the 'run test' link and stuff like that
so that it actually works as a TestCase
yup, i still dislike how we do TOC
mmm, next experiment i guess
[04:38]
.... (idle for 17mn)
gmc, Babar are you aware that f.o is down? [04:58]
gac410yup... your topics were there. I just overlayed them.... but they crash like crazy... nothing works. [04:58]
SvenDowideiti presume the rather mass down-ness needs gmc tho
odd
oh wait
you need the change i made
[04:58]
gac410SvenDowideit: pharvey emailed gmc. and also the infra list. [04:59]
SvenDowideitso that the ssp plugin is on when you run it
else, when the test case plugin runs, all other plugins are turned off
[04:59]
gac410ah... nvm then. [04:59]
pharveyI've sent many emails today [05:01]
SvenDowideitargh!
of course, STARTSECTION{} does not work over multiple lines
more stuff that pharvey was working on
[05:12]
pharvey:( [05:15]
SvenDowideitokay! i now have a more flexible topic creator INCLUDE
lets see :/
more more more
[05:18]
sounds about right
SvenDowideit loves how lc() == 'SS'
er
SvenDowideit loves how uc() == 'SS'
and lc(SS) == ss
like fc() man :/
[05:25]
pharveyhrm where I have I seen that before [05:26]
SvenDowideitsome brian de floil'd your plans guy's blog :) [05:26]
nite :) [05:40]
***gac410 has left [05:40]
.... (idle for 17mn)
ChanServ sets mode: +o MichaelDaum [05:57]
MichaelDaumfoswiki.org
down as can be
[05:59]
SvenDowideitthat would be correct
someone noted that the entire centre its in is down
[06:00]
pharveyhi MichaelDaum, I already E-mailed gmc CC foswiki-infra :) [06:02]
MichaelDaumhe'll up it as soon as his thc goes down bit [06:02]
pharveyI hate when my thermal conductivity is too high [06:03]
SvenDowideiti'm sure it could be worse [06:04]
MichaelDaumhere's some news for you: 1&1's intranet twiki is being replaced by a foswiki atm [06:04]
SvenDowideitcept i don't know 1&1 :) [06:05]
MichaelDaum1and1.com [06:05]
SvenDowideitaw shucks :) [06:05]
dj_segfault1&1 is very popular. That's big news! [06:09]
MichaelDaumsee you later [06:16]
SvenDowideitlaters :)
mmm, i maybe should start cooking dinner
[06:16]
............ (idle for 59mn)
***ChanServ sets mode: +o MichaelDaum [07:15]
....... (idle for 31mn)
BenjaminWilkhello
foswiki,org down?
*foswiki.org
[07:46]
gmcit appears so
checking it out right now (i just woke up)
[07:50]
BenjaminWilkok :) [07:51]
gmcrebooting, it wasn't responding to anything anymore [07:52]
BenjaminWilk:D
okay
but its not windows or?
[07:53]
gmclol, no.. freebsd
and its back it seems
[07:56]
BenjaminWilk;)
hehe yes ...
[07:56]
***Babar sets mode: +v FoswikiBot [08:05]
BenjaminWilkhey i read the MailerContrib and other plugins ... but can i define, that foswiki send me an mail if only one specified topic was changed? [08:09]
PeterSvendsenHi, when I try to upload an attachment on 43Mb I get an error on my foswiki... I have changed the max size to 50Mb but still get an error.....Can&#39;t call method &#34;logger&#34; on an undefined value at D:/www/foswiki/lib/Foswiki/Store/VC/RcsLiteHandler.pm line 352. at D:/www/foswiki/lib/Foswiki/Store/VC/RcsLiteHandler.pm line 352... and a shitload of more errors [08:11]
.... (idle for 15mn)
BenjaminWilk* BenjaminWilk : + TestTopic1
means that i only geht a mail when the TestTopic1 was changed?
or how can i edit the auto generated mail vom wiki?
is there any template?
[08:26]
jastmailnotify.tmpl [08:31]
SvenDowideitgmc thankyou very much :) [08:41]
gmcnp :) [08:42]
SvenDowideitPeterSvendsen sorry, can you report a task with the full stack trace and info about version etc
can't call logger sounds like something that may be cfg related, but :/
so its worth checking your configure and seeing if the logger section reports anything
[08:43]
PeterSvendsenHi Svend I'll do that.. thanks [08:49]
GithubBot[foswiki] FoswikiBot pushed 1 new commit to master: http://git.io/4eeZXQ
foswiki/master 3c9fa45 SvenDowideit: Item11653: replace instructions that say 'just copy&paste into a new topic' with a simple webtopiccreator app...
[08:55]
***GithubBot has left [08:55]
FoswikiBothttp://foswiki.org/Tasks/Item11653 [ Item11653: convert 'create new' docco on SlideShowPlugin into a topic template app ] [08:55]
GithubBot[foswiki] FoswikiBot pushed 1 new commit to master: http://git.io/m4M1Sg
foswiki/master cffa0ca SvenDowideit: Item11916: remove the click on slide feature as it covers up links (and other things too?)...
[09:08]
***GithubBot has left [09:08]
FoswikiBothttp://foswiki.org/Tasks/Item11916 [ Item11916: Links go to next slide instead of to url ] [09:09]
........................... (idle for 2h12mn)
AlexanderStMichaelDaum: Hey Michael, we just have a problem with the LdapContrib regarding escapes within the dn. The original dn is dn=Mueller\, Franz,OU=Test,... . If i set up a test connection with testldap i get the correct dn. If i use the LdapContrib the escape gets escaped, what means i get dn=Mueller\\, Franz,OU=Test. Did you faced that problem somewhere else as well?
me and Jan reviewed the code, but i did not see any codeline where the dn is getting edited
For me it´s an absolutely strange behaviour within LdapContrib, since ldaptest is using the same mechanism to fetch ldap entries...
[11:21]
jastseems to be something Net::LDAP does, or maybe the server... hard to tell; at any rate I don't see any escaping mechanisms in either Net::LDAP or LdapContrib
ldaptest uses the dump method in Net::LDAP; the normal method used by LdapContrib doesn't do any extra processing, though, so I doubt that's the difference
[11:24]
AlexanderStthe only difference is that LdapContrib is getting the dn with the function entry->dn() instead of dumping the whole entry, but i tested it that way, that i changed ldaptest...and of course the dn is correct even by using entry->dn(); [11:26]
..... (idle for 22mn)
jastpossibly related to control extension (which LdapContrib itself uses for paging but ldaptest doesn't) [11:48]
..... (idle for 21mn)
GithubBot[foswiki] FoswikiBot pushed 1 new commit to master: http://git.io/adTegQ
foswiki/master 123d469 SvenDowideit: Item12244: its not just buttons that require change permission that should be styled consistently :/ any top tool bar button should look like a button. tbh, it looks like this css was generated by code that went overboard - so I'm expecting this commit to be fixed in a better way - but this one at least makes the subscribe line (and others that are not core) look consistent...
[12:09]
***GithubBot has left [12:09]
FoswikiBothttp://foswiki.org/Tasks/Item12244 [ Item12244: Subscribe link on top of page (next to Edit / Attach buttons) should be styled as a button instead of as a plain link. ] [12:09]
...... (idle for 29mn)
GithubBot[foswiki] FoswikiBot pushed 1 new commit to master: http://git.io/FN3y2g
foswiki/master a386f3c SvenDowideit: Item658: use html rather than a non-localised image to start the presentation - also means its styled reasonably...
[12:38]
***GithubBot has left [12:38]
FoswikiBothttp://foswiki.org/Tasks/Item658 [ Item658 ] [12:38]
MichaelDaumAlexanderSt, hi
could you write that down into a bug item describing the problem? I'll try to repro it then.
[12:42]
AlexanderStMichaelDaum: I think Jan is still working on that issue..with a little hack (regexing \\ -> \) it´s working now, but we are still looking for the reason why the dn is getting changed within the LdapContrib (or Net::Ldap) [12:46]
jastI still suspect it's the server we need to blame [12:47]
MichaelDaumjast, ah ok. cool. [12:50]
jastyes, it's the server (vanilla MS AD; according to the internet, Apache's LDAP server behaves similarly). apparently happens whenever the control extension is used, e.g. for paging
I disabled paging in the code and the backslashes were okay then
(side note, paging can't currently be disabled in config since the value is initialized as $Foswiki::cfg{Ldap}{PageSize} || 200)
[12:57]
......... (idle for 41mn)
pharveyyay foswiki.org is up! gmc++ [13:41]
..... (idle for 23mn)
MichaelDaumjast, fixed in local version [14:04]
...... (idle for 27mn)
jastMichaelDaum: nice [14:31]
MichaelDaumthere are quite some enhancements in the upcoming ldapcontrib,i.e. support for multiple user and group base dns
as well as support for referrals and references as found in distributed ldap directories
and still: I am confronted with yet another directory layout that the current ldapconrtib cant cope with
even with those enhancements
[14:43]
gac410Babar: Versions again.... did you see my latest discussion with kip3f - Even with new version installed, if ASSERT is enabled, require v5.8.8; fails
Perl 5.10.0 on Mac OSX Snow Leopard fails v-string in use/require non-portable
[14:47]
jastMichaelDaum: gotta love LDAP, right? still, upcoming changes sound spectacular [14:48]
gac410The prior use 5.006; is supposed to prevent that, but it appears 5.10.0 only on osx fails. pharvey checked out another 5.10.0, no issues. [14:48]
........ (idle for 39mn)
Babarbut... should we require the v version? [15:27]
gac410Babar, I think that they are equivalent, right?
btw I think this has nothing to do with the version change. kip3f enabled FOSWIKI_ASSERT for the first time
which converted the warnings to errors. Previously he was using the WarningsAreErrors setting which was removed from 1.2
[15:29]
........ (idle for 38mn)
MichaelDaum: Babar: PatchItem12225Contrib upload is running now. I've tested on Foswiki 1.1.0 and beyond - it fixes the compare for 'HEAD' which breaks with version objects [16:08]
FoswikiBothttp://foswiki.org/Tasks/Item12225 [ Item12225: Plugins using new version system break older versions Foswiki ] [16:08]
gac410If you would all kick the tires on it. It hopefully should be safe. It only impacts configure. When we release extensions with the new "use version" string, this contrib should be a dependency so it is installed first. [16:10]
Now that I've tested from the web, one error is it's not logging the messages correctly, but patch did get applied. [16:20]
.......... (idle for 46mn)
Babargac410: with the parse, probably
without... no lue
*clue
[17:06]
gac410Babar, use 5.006_001; # ditto; preferred for backwards compatibility is what http://perldoc.perl.org/functions/use.html says
So use 5.008_008 should be okay.
decimal fraction such as 5.006, which will be compared to $] , or a v-string of the form v5.6.1, which will be compared to $^V
Or we can leave it as v5.8.8 and let OSX Snow Leopard users who enable FOSWIKI_ASSERT surround it with no warnings, which is what kip3f did
[17:11]
BabarI would never use v5.8.8; personnaly, but... as you wish [17:16]
gac410perl -e"use 5.008_008; print $];" works fine. I'm not arguing for the v5.8.8 That's been in Foswiki.pm for quite a while.
Foswikirev:14860
[17:17]
FoswikiBothttp://trac.foswiki.org/changeset/14860 [ Changeset 14860 – Foswiki ] [17:17]
gac410Blame SvenDowideit [17:18]
Babarwhat is wrong with those convicts? :p [17:21]
gac410Okay. I'm going to change it to 5.008_008 and add a SMELL that v5.8.8 fails on OSX Snow Leopard [17:23]
Babarwhy don't we just freaking: use 5.008_008; ??????
honestly. use 5.6; require... that's total bollocks!
[17:37]
gac410That's what I just did, didn't I
Oh... yeah I didn't delete that.
[17:37]
BabarSven: were you stoned again? [17:37]
GithubBot[foswiki] FoswikiBot pushed 1 new commit to master: http://git.io/4eNF5g
foswiki/master 4c04553 GeorgeClark: Item11878: v5.8.8 fails on OSX Snow Leopard...
[17:38]
***GithubBot has left [17:38]
FoswikiBothttp://foswiki.org/Tasks/Item11878 [ Item11878: require perl 5.8.8 for 1.2.0 ] [17:38]
Babarplus... I thought we said eons ago that we should require 5.8.8. Even for 1.1 [17:40]
gac410no, 5.8.4 for 1.1, 5.8.8 for 1.2 5.18.40 for Unicode :D [17:41]
Babar5.14 should be enough :)
but as unicode will be 5.6 (I'm talking Foswiki, not perl), that's fine :)
[17:42]
gac410I though the plan was to pull the unicode branch into trunk once 1.2 is branched.
Should I delete the use 5.6 before I close the task again?
[17:43]
Babarreplace the use 5.6; require 5.8; with a: use 5.8;
I really do not see any benefit to die later.
but... I guess Sven had a good reason for doing it that way
[17:50]
gac410I'll ask SvenDowideit later to see what is reasoning was. [17:59]
Babarthanks [18:02]

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