#foswiki 2012-11-25,Sun

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

WhoWhatWhen
gac410pharvey1: On the issue you found in Net.pm ... I think a better solution is to set $old_locale = POSIX::setlocale( LC_TIME(), 'en_US.ISO8859-1' ) || "";
The locale of "" will use the system default locale if I understand setlocale correctly. So it avoids the undef and sets a valid locale if the original locale was not set.
I think I'll check that in
[00:02]
pharvey1gac410, good to know [00:18]
gac410I'm still not convinced 100% that the code is correct. I guess Timothe's away. [00:19]
pharveyI've almost got a new pseudo-install using Foswiki::Configure exclusively [00:19]
FoswikiBothttp://trunk.foswiki.org/System/PerlDoc?module=Foswiki::Configure [00:19]
gac410cool [00:19]
pharveyhaven't tried it out on Release01x00 branch :P
pharvey guesses he'll have to un-delete a bunch of code to support it as a fall-back
[00:19]
gac410I think the Load Value Valuer, and FoswikCfg classes are about the same between Release and trunk. Need to compare though :( [00:21]
pharveyyeah, I'm vaguely hopeful [00:21]
gac410hm... So Timothe definitely got the code wrong. :( The 2-operand setlocale does NOT query, it returns the set locale, or undef if the local requested is not available. [00:24]
....... (idle for 30mn)
GithubBot[foswiki] FoswikiBot pushed 1 new commit to master: http://git.io/Dm8s_w
foswiki/master 472196e GeorgeClark: Item12264: Fix undefined variable in Net.pm...
[00:54]
***GithubBot has left [00:54]
FoswikiBothttp://foswiki.org/Tasks/Item12264 [ Item12264: AccessControlTests fail with Use of uninitialized value in subroutine entry at lib/Foswiki/Net.pm line 435. ] [00:55]
....................... (idle for 1h52mn)
gac410howdy pharvey... i've been playing around with a cli lsc updater. Issue is going to be the use of shell specific metacharacters in regexes.
Simple things - setting a path or wikiname, etc. seem to be fine with Getopt::Long, but not things like NameFilter
[02:47]
...... (idle for 28mn)
pharveyah. Shell escaping madness.
FoswikiBot: corelist Getopt::Long
[03:15]
FoswikiBotpharvey: Getopt::Long was first released with perl 5 (released on 1994-10-17) [03:15]
pharveyw00t [03:15]
gac410yeah it's safe [03:15]
pharveypharvey wonders why we wrote our own opt parser for pseudo [03:16]
gac410NIH? [03:16]
pharveyno idea. I just converted it to a hashtable rather than cascading if-else :P
speaking of, seems twiki finally fixing the SSP code in that respect
[03:16]
gac410Getopt::Long can return a hash or array, seems pretty flexible. Yeah I saw that.
Also a %CALCULATE that's a registered tag too. Though CDot had a version of that in our code, but commented out.
[03:17]
pharveyoh, I thought I saw those checkins on our SSP [03:19]
gac410I think Sven was looking at porting some of the work too.
He checked in the test cases.
[03:19]
pharveynice
I've always tried to use QUERY rather than calc, haha.
[03:19]
....... (idle for 34mn)
SvenDowideitfunnily enough
all i did was uncomment cdot's
and change its name to be compratible
[03:54]
gac410Ah... what is naf mean... [03:55]
SvenDowideitand then added the test case peter made, and was suprised how many passed
yup
naf means that i agree, waht a naf bug
tis british - i blame cdot for reminding me of it :)
[03:55]
gac410Ah... [03:56]
SvenDowideiti recal that cdot found some issue that made him change what it woed for
but its vague, very vague
pharvey, just in case you missed ti
yes, i'm still going to osdc
i'm not sure if i'm going to the cmsexpo nite tho
[03:56]
gac410Item12263 - is the fix micha posted correct, and does it go in Users.pm or UserMapping.pm or ... passwordError is defined in four places [03:57]
FoswikiBothttp://foswiki.org/Tasks/Item12263 [ Item12263: using wrong mapper handling password errors ] [03:57]
gac410We have no tests for that at all I could find. [03:57]
SvenDowideityes, um, i dunno [03:57]
gac410He made it a blocker for 1.1.6 :( [03:58]
SvenDowideitbut users.pm should make sure it gets the right mapper
yes, in Users.pm
and yes, its important enough to just use
[03:58]
gac410His fix would have an undefined variable. Since nobody calls it with the login name [03:59]
SvenDowideitas without the error will never be accessible
eh?
oh gawd
[04:00]
gac410So it's not just Users.pm, every caller needs to add the $loginname [04:00]
SvenDowideitah, puce [04:00]
pharveyI see I see [04:00]
SvenDowideitin that case [04:00]
gac410Which is an api change. [04:00]
SvenDowideitits busted [04:00]
gac410Been busted a LONG time I suspect [04:01]
SvenDowideitand the real fix is to iterate through all mappers
and return the errors from all
and given that atm, we have no pwd errors in the base mapper?
then i'd fix it in 1.1.6
by unconditionaly returning the error from the non-basemapper
one mo, lets check
[04:01]
gac410I'm not thrilled about changes before the RC. So we need to delay 1.1.6, and do another beta. [04:02]
SvenDowideitoh crud, that isn't enough [04:02]
gac410I don't know the code well enough to risk the "quick fix" [04:02]
SvenDowideitna
push back :)
its been like this for >5 years?
[04:02]
gac410Same with Item12262 - I cannot recreate his bug. But I've found a different bug and fixed that. [04:03]
FoswikiBothttp://foswiki.org/Tasks/Item12262 [ Item12262: calling getWikiName with a WikiName pollutes the internal caches ] [04:03]
SvenDowideitwell, the pain is that he'll have found it as part of making ldapcontrib work properly
and um, so it matters more now :/
well, as there's no chance of 1.2.0 this year
[04:03]
gac410Yeah. But if he can't take the time to explain how to recreate them on the eve of the RC. I really wanted to build the RC Sunday [04:04]
SvenDowideitya, thats the other side of it
and then he can rm a 1.1.7 >:}
[04:04]
gac410I can't find any way to trigger that cache corruption. I've been dumping all the caches before/after the call he claims breaks them but they are fine. I wonder if it's an ldap cache that's getting corrupted. [04:05]
SvenDowideitno idea
SvenDowideit recons that it can't be a blocker if the report's not given enough info
[04:06]
gac410Well I'll build a private _005 beta on Sunday if I've not heard from him, and delay RC 'til Monday or Tuesday. [04:07]
SvenDowideit:)
acutaly
wrt the passwrd error
[04:07]
gac410Though if the patches are "iffy" then I'll do another public beta and delay a few days. [04:07]
SvenDowideityou _can_ just change F::Users;:getPasswordError to use a $CUID
and for the core to leave it undef
then the result will be the same for core
but anyone that calls it in a sane way will get a sander result
saner
[04:08]
gac410okay. Maybe I'll poke at it. [04:09]
Hm. Can we assume that the passwordError is for whatever the session use is?
session user
[04:20]
SvenDowideitmmm, no, i think it could happen for a user who's email we looking for [04:20]
gac410ah. :( [04:20]
SvenDowideit(which a plugin might do differently [04:20]
gac410Maybe make it, if passed a cUID, use it, otherwise default to the current session cUID which is probably better than no cUID [04:21]
SvenDowideity, thats probly very true [04:22]
gac410hm... Might be more complex. The user established by default is the guest user, but if I try to login, the password error would be generated by the mapper that handles the login request, not the base mapper. [04:31]
SvenDowideiti still recon just use a cuid if its there
and change nothing else
that way we have an improvement for 1.1.6
and can fix for real in 1.2
[04:32]
gac410yeah, adding cUID doesn't seem to break anything. [04:34]
BTW... do you have any clues about Item12084? I'm suspecting a store issue. [04:39]
FoswikiBothttp://foswiki.org/Tasks/Item12084 [ Item12084: Corrupted WikiUsers topic combined with AllowLoginName enabled breaks Upload. ] [04:39]
gac410Hm.... do we need to increment the 1.1.6 API version, since we are adding the cUID to the API? [04:44]
SvenDowideitno
because its a very inconsequential additive change :)
[04:44]
gac410yeah.. I guess it can't break regardless of how it's called. [04:45]
SvenDowideitor, yes - becauseyou want to [04:45]
gac410No... bumping the API to 2.3, means we'll have to bump trunk to 2.4, and any extensions that assume 2.3 is trunk. [04:46]
SvenDowideiti don't know when i last used the api version [04:46]
gac410We need to increment major if there is any chance of a minor bump. [04:46]
SvenDowideitprobly some time when it was 1.0 [04:46]
gac410I know I've used it in plugins. [04:46]
SvenDowideitas its alot sander to test for ->can{the func you need)
saner :/
[04:47]
gac410good point [04:48]
SvenDowideitna, to quote pharvey
SvenDowideit go fall down now
[04:48]
GithubBot[foswiki] FoswikiBot pushed 1 new commit to master: http://git.io/T3bV0A
foswiki/master 4b36222 GeorgeClark: Item12263: Pass the cUID in passwordError...
[04:53]
***GithubBot has left [04:53]
FoswikiBothttp://foswiki.org/Tasks/Item12263 [ Item12263: using wrong mapper handling password errors ] [04:53]
***gac410 has left [05:05]
GithubBot[foswiki] FoswikiBot pushed 3 new commits to Release01x01: http://git.io/LDX29g
foswiki/Release01x01 6aae1ef GeorgeClark: Item12262: Unit tests that won't fail....
foswiki/Release01x01 dd1f294 GeorgeClark: Item12263: Pass the cUID in passwordError...
foswiki/Release01x01 d099862 GeorgeClark: Item12262: Test now fails...
[05:06]
***GithubBot has left [05:06]
FoswikiBothttp://foswiki.org/Tasks/Item12262 [ Item12262: calling getWikiName with a WikiName pollutes the internal caches ] [05:06]
.... (idle for 16mn)
pharveyso, it seems there's no way to remove a cfg entry via Foswiki::Configure::Foo. [05:22]
FoswikiBothttp://trunk.foswiki.org/System/PerlDoc?module=Foswiki::Configure::Foo [05:22]
pharveyA part of me wonders why we aren't using Data::Dumper, another YAML [05:23]
......... (idle for 44mn)
***sendak.freenode.net sets mode: +o ChanServ
sendak.freenode.net sets mode: +o SvenDowideit
[06:07]
......................................................................................................... (idle for 8h44mn)
gac410Ugh... the fix for Item12263 isn't right. Though this time I think it's in TopicUserMapping. It uses the cUID field in handlesUser to pass along the mappingID. It doesn't use it as a cUID. [14:54]
FoswikiBothttp://foswiki.org/Tasks/Item12263 [ Item12263: using wrong mapper handling password errors ] [14:54]
................ (idle for 1h19mn)
gac410SvenDowideit: Ran git bisect. Foswikirev:16040 is what broke TableParser tests [16:13]
FoswikiBothttp://trac.foswiki.org/changeset/16040 [ Changeset 16040 – Foswiki ] [16:13]
gac410gac410 really wishes that test writers would include at least a *hint* of a comment on wtf their code is doing. [16:24]
.................................................................................. (idle for 6h48mn)
SvenDowideitgac410, figures
i started thinking to revert those changes anyway
[23:12]
....... (idle for 32mn)
foswiki_irc4Good Evening [23:44]
gac410good evening foswiki_irc4 [23:45]
foswiki_irc4I wondered if anyone else uses the AcronymDefinitionsPlugin [23:45]
gac410SvenDowideit: I think the fix to the table parse unit test is simple. But I just don't understand what the heck the test is doing. [23:46]
foswiki_irc4I can't seem to get it working and I wondered if anyone had any suggestions [23:46]
gac410flexibeast: are you around :) I have not tried that plugin. [23:47]
flexibeastfoswiki_irc4: i developed that plugin; what's the issue you're having? [23:49]
foswiki_irc4Oh cool, thanks nice to meet you [23:49]
flexibeastLikewise. :-) [23:50]
foswiki_irc4I can't seem to get it to display the popup [23:50]
flexibeastHm. What browser are you using? [23:51]
foswiki_irc4I have probably got something wrong. After its install should it just automatically start linking all references found in the acronymdefinitions topic?
IE9 & Firefox 16
[23:51]
flexibeast*nod* Well, it should be automatically wrapping candidate acronyms in an HTML <acroynym> tag.
Does the page source show that's been done?
[23:54]
foswiki_irc4I never checked that :)
yes its looks like it >>> <acronym title='Active Directory Controller'>ADC</acronym>
[23:57]
flexibeastHmm, odd.
i've only tested it in Firefox and Chromium on Linux; haven't tried it on a Win system.
But i would have thought FF would honour the acronym tag, even if IE doesn't.
[23:58]
foswiki_irc4Maybe its my Theme I'm using FatWilly's I'll try it on the default [23:59]

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