#foswiki 2013-10-23,Wed

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

WhoWhatWhen
***gac410 has left [04:36]
..................... (idle for 1h40mn)
MichaelDaumgmc, you can ignore these messages. you are not the first thinking they are misleading. wasn't me who worded them.
"missing ids" only flags constraints not being satisfied. it does NOT mean that they express a necessary requirement.
read them as: if you load jquery.migrate, then insert it BEFORE jquery.foswiki.
so the constraint is more about the linear order of javascript files, NOT about them being there at all.
^gmc^gac401 ... pinged the wrong person
somebody insisted on these debug messages to be produced. I objected for exactly this reason: it confuses more than it helps.
[06:16]
***ChanServ sets mode: +o CDot [06:21]
MichaelDaumah there is this somebody (!) ;) [06:21]
pharveyThey are useful when debugging.
In fact, debugging is almost impossible without them
[06:22]
..... (idle for 23mn)
MichaelDaumdebug what? the addToZone api? [06:45]
..... (idle for 22mn)
JulianLevensCDot, SvenDowideit: comments please on http://foswiki.org/Development/StoresShouldBePassedConfigHash [07:07]
CDotIIRC I wrote the messages. because without them, working out WTF the addToZone stuff is doing is a nightmare. Once you understand how it works it's easier, but you still need the messages sometimes. [07:09]
JulianLevensI'm working on enhancing change_store.pl as written by CDot to convert between stores. I've been using that to convert to Versatile. I have also copied and hacked this script as test_store.pl to do performance tests [07:09]
CDotIf you feel they are misleading, then feel free to unmislead them. [07:10]
JulianLevensI'm planning to create a StoreToolsContrib with these scripts in them. I'm also aware of a possible need to create derived stores. E.g. Fowsiki::Store::Versatile::Variant1 just to test alternate options out. The variants would inherit the original store and in principle implement minor differences
This also impacts $Foswiki::cfg{Store}{ImplementationClasses}, I need a way to change this per store *configuration* being tested (it's still one logical store in the end). So I need to be able to change this and instantiate a particular store config, then change the ImplementationClasses and instantiate another config to either convert to a new store or test multiple config options
[07:13]
CDotAh, ok. That isn't clear from your post.
Yes, the namespacing was snafu pre-1.2.0. It's better now (but may still be improved)
[07:18]
................... (idle for 1h31mn)
pharvey: can you mail me with your preferred email address please? i have your csiro addr in my contacts book..... [08:50]
MichaelDaumJulianLevens, any news on your Amsterdam idea? [08:51]
pharveyCDot: sent [08:57]
CDotthanks [08:57]
JulianLevensMichaelDaum: not yet. Thanks for the reminder. I'm working from home today, tomorrow at work I must formally propose this [08:58]
MichaelDaumokay let us know as it might be a vehicle to host the GA as well :) [09:00]
pharveyMichaelDaum: I guess CDot already answered, but I mean it's really tough for newcomers to use %ADDTOZONE and figure out why their ordering isn't happening the way they thought it should.
is it requires="JQUERYPLUGIN" or requires="JQueryPlugin" or requires="jQuery" or ...
[09:03]
MichaelDaumpharvey, but does that mean everybody in needs debug output in the markup? [09:05]
pharveyperhaps /requires/ should've been named /after/ or something
MichaelDaum, no I suppose not
but there needs to be an in-your-face way of knowing the feature is there
[09:06]
MichaelDaumthe ids are: uc(jqueryplugin::<name of your module>) [09:08]
JulianLevensMichaelDaum: I can be persuaded to host the GA, but I was hoping to keep them separate [09:08]
MichaelDaumJulianLevens, we might end up having them both separate, yes.
so that we can concentrate on hacking :)
we had a simple irc based ga last time.
[09:08]
JulianLevenshacking +1 [09:12]
CDotthe GA needn't take long; we combined it with our hacking session in CERN to considerable effect [09:14]
JulianLevensI'll leave the GA option open for now. First thing is to get the camp itself agreed [09:20]
...... (idle for 29mn)
***ChanServ sets mode: +o Lynnwood [09:49]
.................................. (idle for 2h46mn)
ChanServ sets mode: +o gac410 [12:35]
gac410G'Morning all (or afternoon / evening ...)
Thanks for the explanation about those debug messages, I was sure somehow I had broken 1.1.9.
CDot, the task you bumped to Urgent - upload failure w/ closed file handle. Does it need to block 1.1.9, are you making any progress?
[12:38]
.... (idle for 19mn)
Anyone heard from Babar lately? [12:59]
MichaelDaumhe's been moving as far as I remember [13:00]
CDotgac410: I am only just back from travelling, and haven't had a chance to look at it yet. I shall do so now. [13:09]
gac410Okay thanks CDot .. And welcome back. :) [13:10]
.......... (idle for 46mn)
***ChanServ sets mode: +o CDot [13:56]
......... (idle for 41mn)
ChanServ sets mode: +o CDot [14:37]
.......... (idle for 47mn)
gac410CDot, another when-you-have-a-chance. Your commit http://trac.foswiki.org/changeset/16869 reverted at least parts of changes I had made
And removed them from the plugin release history.
In particular http://foswiki.org/Tasks/Item11918
[15:24]
CDotyay! I managed to reproduce 12048 - at long last!
gac410: what task was that?
(trac links are broken)
[15:25]
gac410Task was a catch-all Item12525: bring Release01x01 codebase up to trunk level but it reverted stuff as well.
gac410 wondering if you found a different fix for the stuff you reverted.
[15:27]
CDotI see; OK, then why weren't those changes in trunk? That is the master codebase for all extensions.... [15:28]
gac410They were [15:28]
CDotI was working from trunk [15:28]
gac410Item 11918 has the commit to both trunk and release branch [15:28]
CDotyes; I always commit to trunk, and then copy the same code to Release01x01
but if your changes weren't in trunk, they would get blown away....
[15:29]
gac410Well a subset of commit to Item11918 was removed. The unit test stayed, the Foswiki.spec was reverted. I COMMITED TO TRUNKI! [15:29]
CDothmmm, ok. That shouldn't have happened, then [15:30]
gac410I probably missed the change log, but thats NBD, The bigger concern is other stuff that *was* synced.
And if one piece was accidentally reverted ... now I need to review every commit :( yech.
[15:30]
CDotwhat tasks? [15:32]
gac410Not sure yet. Other than the obvious ones missing from the change log: See http://trac.foswiki.org/changeset/16869
They are mostly inconsequential. .... but it leaves me in doubt :(
[15:33]
CDotsure. I would not normally have committed without an svn up immediately before the checkin (standard procedure)
of course, if I had done a perltidy locally first, that might have caused a problem
why not svn diff the rev before my checkin with the "after"?
CDot will look after he has nailed 12048
[15:34]
gac410Trunk has been so questionable to me, I never copy en-mass
Yeah 12048 is much more important. I think it's more the log. But the Foswiki.spec change is also concerning.
[15:36]
CDotnever copy core code en masse - but contribs are not core, they are extensions [15:36]
gac410Default contribs are part of core.
At least that was always my interpretation. Some stuff is way different ... SlideShow for ex.
[15:36]
CDotwe could argue about that until the cows come home. I have always battled to maintain a common codebase on all 'core' extensions [15:37]
gac410And the mass copy of TMCE for 1.1. brought a lot of 1.2 madness into 1.1. Can't recall which release, but the whitespace changes on trunk were TOTALLY broken
er... Wysiwyg.
Common code base only works if trunk changes are stable and well tested. The wysiwyg whitespace was a disaster to pull into 1.1.
I know you've always battled to maintain them as common. But mass-sync-up of *others* changes (not yours) ... anyway ... water under the bridge.
Okay, other than http://trac.foswiki.org/changeset/16869/branches/Release01x01/MailerContrib/lib/Foswiki/Contrib/MailerContrib/Config.spec which reverts part of Item11918, I don't think anything else was missed.
And since the unit test I added remained, and passes Oops... nope
http://trac.foswiki.org/changeset/15940/trunk/MailerContrib/test/unit/MailerContrib/MailerContribSuite.pm still tests the example you reverted.
Nope, you reverted to a BAD example. MailerContribSuite::testExcludeAll
*** $+ matches null string many times in regex; marked by <-- HERE in m/
^\s+\*\s(?-xism:(?:TemporaryMailerContribTestsUsersWeb\.)?)
The example ends in $, and gets wrapped with a +, which breaks the regex.
[15:38]
....... (idle for 32mn)
PatternSkin is one that I don't think can be synced - Trunk has changes not acceptable for release11 [16:21]
orevis there a good way to replace the left navigation bar on all webs with a global preference? or does the webleftbar page need to be manually replaced in each web? using patternskin [16:21]
gac410Unfortunately the web left bar is unique per web. [16:23]
orevok, can I modify the default template used when making a new web? is that in System? [16:24]
gac410The default comes from the _default template web
When you create a new web you can specify the template to copy. If none is specified, _default is copied.
[16:24]
orevis that updated if a new release is applied? [16:25]
gac410Yes, But we now ship WebLeftBarExample, not WebLeftBar, If WebLeftBar topic is not found, then WebLeftBarExample is used.
darn... except for _default web
We should probably ship WebLeftBarExample in the _default web, so if you create a WebLeftBar, it doesn't get overwritten.
[16:26]
LynnwoodGreetings everyone. Does anyone know how VarQUERYPARAM actually handles multiple values (such as from checkboxes)? The doc says "Multi-valued parameters will have a "row" for each value" but I can't figure (or find examples) of what that's suppose to mean. AFAICT from experimentation, it only renders the first value. [16:31]
gac410Looking at the perl code, it doesn't seem to handle that. :( [16:34]
Lynnwoodthat's kind of what i figured...
".. a "row" for each value..." doesn't really mean anything
there are no "rows"
[16:35]
gac410Right. internally, a multi-valued query param, returns an array, with the same parameter occurring multiple times.
QUERYPARAM would need to flatten that somehow.
[16:36]
Lynnwoodrats... i just tried formatting QUERYPARAM to call VarURLPARAM (which does handle multiple values) but it doesn't render.
i tried this: %QUERYPARAMS{format=" * *$name* = $percntURLPARAM{\"$name\" multiple=\"on\" separator=\", \"}$percnt"}%
no such luck... guess i'll have to use FORMATLIST or VarFORMAT
[16:38]
i think i'll go ahead and post a tast related to QUERYPARAM, at least to correct the docs.
task
[16:50]
gac410Yes please do [16:50]
..... (idle for 23mn)
JulianLevensLynnwood: what's the benefit of QUERYPARAMS over URLPARAM. Sorry, to ask but as I read the docs they are pretty much the same beast. If they are not then the docs should explain the reasons for choosing one over the other [17:13]
LynnwoodHey JulianLevens - QUERYPARAMS list all the URLPARAMS. URLPARAMS only renders a single specified param. [17:14]
gac410URLPARAMS though has code to join all elements of a multi-value parameter. [17:14]
Lynnwoodit joins all the values of a multi-instance parameter (such as checkboxes)
you understand that if you have 5 checkboxes (all checked) with the same name, the result is that the param is listed 5 times.
But I don't think URLPARAM will render multiple _different_ parameters
And you are right: URLPARAM _can_ render this kind of multi-value param.
[17:15]
JulianLevensOK, I've used URLPARAM many times so I get that, and re-created all the <input fields so QUERYPARAMS would be used to re-build this <input automatically for a subsequent POST [17:18]
Lynnwoodexactly
but unfortunately it breaks down in the case of a param that has multiple values.
[17:18]
JulianLevensQuite, that I see [17:19]
LynnwoodWhat i ended up doing was using VarFORMAT and used QUERYPARAMS to create the list of params to render and then in the format used URLPARAM to render each elements (and providing for multiple values).
does that make sense?
[17:20]
JulianLevensYes [17:20]
LynnwoodIn my case, i was passing on the values via hidden inputs so I didn't actually need to know what kind of elements created each param. [17:21]
JulianLevensPresumably QUERYSTRING is used in a GET situation [17:21]
Lynnwoodi guess... i am using GET here...
i wonder if QUERYSTRING would work with POST?
[17:21]
JulianLevensI suspect an experienced web-developer may understand this all easily from the existing docs. However, those with less experience would benefit from some extra notes in the docs [17:23]
Lynnwoody, it does seem to make some assumptions about understand url parameters.
It's hard to know what to assume. Perhaps a link to basic info about url parameters in wikipedia or such.
[17:25]
JulianLevensIt will need some thought, even something 'simple' like this can take more effort than you think. [17:32]
Lynnwoodah yes... :-) [17:32]
JulianLevensJust a quick look at wikipedia suggests that these topics would be overkill. However, I have other things to do right now, I'll try to raise a task though [17:35]
Lynnwoodthanks! [17:40]
...................................... (idle for 3h8mn)
***JulianLevens has left [20:48]
................................ (idle for 2h35mn)
ChanServ sets mode: +o pharvey [23:23]

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