#foswiki 2015-01-14,Wed

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

WhoWhatWhen
foswiki_irc3is it possible to add icons for new file types in the attachments table
because I was able to get .tar attachment icon to work with %ICON{tar}%
but it doesn't resolve in the attachment table
[00:04]
gac410hm. I don't think I've been asked that one before. I have to look at how that works. [00:05]
well I'm reading the code ... the row is built and icon expanded down in Foswiki::UI::Attach ... but so far I'm confused. [00:11]
foswiki_irc3or how about removing the attachtable
I know how to do it with a new skin
but I am hoping that there is a preference setting for the pattern skin
[00:12]
gac410It looks like it expands %ICON{"ext"}% where ext is the suffix of the filename. [00:13]
foswiki_irc3right
and i got it working for %ICON{"ext"}%
but it isn't working in the attach tables
[00:13]
gac410hm. I'll have to poke a bit more [00:14]
foswiki_irc3its not a big deal
don't worry about it
[00:23]
gac410now I'm really curious. It has to be an attachment of the DocumentGraphics topic, I/m trying to add a "vim" type on my system, still not getting it picked up. hmmm [00:24]
foswiki_irc3yea
i have no problem with getting the %ICON% Macro to work
and the automatic icon even worked with the AttchmentList Plugin
just not for the attachtables
[00:26]
gac410I'll try to figure it out. Might be a bug. I'll have to put some print statements inline to figure out how it's working. [00:28]
foswiki_irc3in 'attach.pm': elsif ( $attr eq 'ICON' ) { return '%ICON{"' . $file . '" default="else"}%';
shouldn't $file be $ext
elsif ( $attr eq 'ICON' ) { $file =~ m/\.([^.]*)$/; return '%ICON{"' . $1 . '" default="else"}%'; }
that fixed it for me
thats line 169 in lib/Foswiki/Attach.pm
[00:28]
gac410cool. If you have an account on foswiki.org please open a task and I'll get your fix applied. If not I can open it. We'll get the fix into Foswiki .1.2
Thanks for reporting and fixing this :)
[00:31]
foswiki_irc3I'll try and make a ticket real quick [00:31]
gac410Your fix looks good. works fine here. I'll see if I can find if we have a unit test for this and will check it out. We probably ought to have this documented somewhere ... how to add a filetype icon :)
interesting that it works for .txt and .zip files without your fix. There must be something else happening there as well.
[00:35]
foswiki_irc3http://foswiki.org/Tasks/Item13212
yeah, i was thinking that too
[00:41]
gac410Thanks for the task. This code is probably way more complicated than it needs to be. Too many layers of modifications over time I suspect. The old DocumentGraphics got supplanted by the FamFamFam graphics, and the icons.tmpl template based rendering.
Not anything we can touch for 1.2. ... small bugfixes only, we hope to start the alpha / beta process in a couple of weeks.
[00:49]
foswiki_irc3it seems like I need to put someing in pub/System/DocumentGraphics/_filetypes.txt
but my first guess at it didn't really work
i think i figured it out actually
[00:50]
gac410hm.. I was able to get a vim.png / vim.gif displayed for a blah.vim file with your fix and without touching _filetypes. [00:52]
foswiki_irc3i think that they have to be title [ext]/png
or [ext].gif
while i have [ext]icon.gif
and then I think I have to add the icon I want to the pub/System/DocumentGraphics/_filetypes.txt list
[00:52]
gac410ah. yeah. I created vim.gif and vim.png
mapping must be needed if the icon name is not identical to the extension.
though using file extensions is not great. we really ought to do mime magic to examine the contents. though I suppose the overhead of doing mime magic against each file would be excessive
so it looks like there are two paths. (with your fix). 1) Add an icon named with the extension. vim.png for blah.vim files. 2) Add a mapping in _filetypes that maps file suffix to icon.
[00:53]
foswiki_irc3i had to do both [00:59]
gac410interesting ... .vim worked fine without my touching _filetypes. (though I'm running 1.2 from git checkout, so might be differences) [01:00]
foswiki_irc3its only working for me if I name it [ext].png(/gif) AND add to the _filetypes.txt [01:00]
gac410Did you also add it to System/DocumentGraphics as an attachment? [01:01]
foswiki_irc3yes
well, whatever, I got it working
thanks (again) for your help
[01:03]
gac410interestingly, with your fix applied, the _filetypes stopped working for a somefile.ebuild. [01:06]
foswiki_irc3yeah, mine kind of breaks some of the functionality
if you have an icon that is realted to the filename instead of the extension
the example given in the Attach.pm is "build" i think
[01:07]
gac410By passing in the filename, then ICON processor maps the suffix by the filetype table. By passing the extension only, then the ICON doesn't process the fileytype [01:07]
foswiki_irc3yeah [01:08]
gac410I'll poke at it some more, but I don't think we can take your fix as is. Something more is needed. [01:08]
foswiki_irc3i agree [01:09]
gac410The type to icon mapping is done it ICON not Attach. So attach calls ICON with a full filename, and ICON does the magic. So the bug is on the ICON side. [01:09]
foswiki_irc3yes
thats how i found the _filetypes file
I think maybe there should be an attempt at resolving %ICON{"[ext]"}% before opening the _filetypes.txt
because the way it works now is that if there is no %ICON{"someString.something"}%
it tries first to resolve .something extension with the _filetypes.txt
[01:09]
gac410right. [01:11]
foswiki_irc3and if it doesn't find a match, it give it the 'else' icon
but it never tries to simply use the $ICON{".something"}%
unless there is a match in _filetypes.txt
which seems unnnecessary to me
[01:11]
gac410right. it should also try "something" as an icon directly [01:12]
foswiki_irc3thats what I am thinking
but I also am not sure where the design ideas came from
originally
[01:12]
gac410A lot of this has evolved. in some cases the original design is long lost. foswiki evolved from twiki which evolved from JOSWiki [01:14]
foswiki_irc3Interesting.
I don't see how it would break anything
it should probably try to use the _filetypes.txt before trying to strip out the extension
[01:16]
gac410With your patch removed the _filetypes worked fine. My "vim" and "ebuild" types both worked.
vim being an icon I added, and ebuild I mapped to zip icon.
[01:17]
.................................. (idle for 2h48mn)
GithubBot[distro] gac410 pushed 1 new commit to master: http://git.io/u6U6Wg
distro/master 345c6f0 George Clark: Item13212: Document file type icon processing...
[04:06]
***GithubBot has left [04:06]
.................................... (idle for 2h57mn)
ChanServ sets mode: +o MichaelDaum [07:03]
.... (idle for 18mn)
GithubBot[distro] MichaelDaum pushed 1 new commit to master: http://git.io/LqeC3A
distro/master 70061ba MichaelDaum: Item13213: fixed nonce recycling in save+continue...
[07:21]
***GithubBot has left [07:21]
.... (idle for 19mn)
ChanServ sets mode: +o CDot [07:40]
....... (idle for 33mn)
ChanServ sets mode: +o Babar
ChanServ sets mode: +o Babar
[08:13]
................ (idle for 1h17mn)
ChanServ sets mode: +o MichaelDaum [09:30]
....... (idle for 32mn)
andrei_tHi, is there a way to enable the original rdiff compare on user-specifically?
As in, is there a configuration a user can set to see the wikitext compared instead of the html
[10:02]
GithubBot[FastCGIEngineContrib] MichaelDaum pushed 1 new commit to Item13010: http://git.io/tJ1ToA
FastCGIEngineContrib/Item13010 9d9a8b0 MichaelDaum: Item13010: make LocalSite.cfg checks optional
[10:06]
***GithubBot has left [10:06]
MichaelDaumandrei_t, try appending render=debug
to the url
[10:07]
andrei_tMichaelDaum: thanks, but now it compares it side by side
still html
?rev1=18;rev2=19&render=debug
like this?
[10:08]
MichaelDaumand rdiff instead of compare [10:10]
andrei_tyup, that did it
thank you
[10:12]
***MichaelDaum changes topic to: Next Release Meeting, Monday 26 Jan, 1300Z #foswiki-release. Channel logs: http://tinyurl.com/7atzrpz [10:15]
........ (idle for 37mn)
MichaelDaumThere's a google calendar now for events, meetings, milestones and release planning: http://foswiki.org/Development/FoswikiCalendar [10:52]
CDot:-)# [10:58]
Nice one; imported. Now I have no excuse for missing another release meeting. [11:06]
..................... (idle for 1h41mn)
GithubBot[distro] MichaelDaum pushed 1 new commit to master: http://git.io/gALVOg
distro/master fe71c45 MichaelDaum: Item13207: add whitespaces in TML-sensitive areas...
[12:47]
***GithubBot has left [12:47]
.... (idle for 16mn)
ChanServ sets mode: +o Lynnwood [13:03]
....... (idle for 34mn)
ChanServ sets mode: +o gac410 [13:37]
.... (idle for 18mn)
gac410Howdy CDot ... I've been pondering the issue of .spec and {Module} not being applied when pseudo-install of extensions. It could probably be done a few different ways.
pseudo-install invoke some sort of merge spec wizard, Add a "bootstrap extension" that merges the spec, or maybe that's the same thing,
or expand Configure::Package to be able to symlink in addition to copy during install, ... it adds the {Module} already
I'm concluding that my problems with incomplete config after psuedo-install is really an issue with pseudo-install
gac410 really needs to build an alpha build so we can try all this in a "real" installation.
[13:55]
GithubBot[NatSkin] MichaelDaum pushed 1 new commit to master: http://git.io/mYBTzw
NatSkin/master 9e2c41c MichaelDaum: Item12481: added screenshots and docu
[13:57]
***GithubBot has left [13:57]
gac410I changed pseudo-install -A to call tools/configure to bootstrap. We need the equivalent to set just a single extensions's defaults.
Howdy MichaelDaum ... saw you deleted my two little thumbnails of tmce vs. nat. They were lame I guess. What I was really trying to accomplish was to give the new user some way to know which editor they were currently running.
[13:57]
MichaelDaumyea wel
l
[13:59]
gac410What might be better is if we could render on the top of the help. "You are currently running the _blah_ editor" [13:59]
MichaelDaumI th-th-think they figure that out right away [14:00]
gac410hm. yeah I supposed. I've found never underestimate the ability to confuse users. ... cdrom cupholder analogy comes to mind. :D [14:01]
MichaelDaumand me failing on apple designed user experience
was trying to forward 3 images received via sms using email ... the hell thats complicated
[14:02]
gac410what a tangled mess trying to figure out how to simply add a new filetype icon to the attachments table. Nothing like having a hidden .ext -> icon map that user needed to update. DocumentGraphics/_filetypes.txt
Unless you know to update that file, all new icons and types are just NOP in the attachment table.
[14:03]
MichaelDaumMy intention behind the help-text simplifications was not to push down the real help - the edit shorthands - too far down the line
gac410, yea. thats why I've made MimeIconsPlugin ... which contains all the magic required
[14:04]
gac410I agree with you. As I said, they were a bit lame ... Don't want too much boilerplate at the top for sure. Maybe even want to twistyize it. [14:04]
MichaelDaumtwisty -> gone
wrt NatEdit: I was planning to abstract out the editing engine so that different ones can be added using the same facade (toolbars, dialogs, etc)
[14:05]
gac410yeah that would be nice.
Not for 1.2 :D
[14:07]
MichaelDaumRawEditEngine.js, TinyMCEEngine.js, CodeMirrowEngine.js, AceEditorEngine.js [14:07]
gac410ping gmc ... are you around. Your pita RM is knocking ...
CKEditorEngine ..
yeah the framework around tmce has gotten really complicated. I think cdot was lamenting the hacking done over time with the foswiki extensions to the interface
[14:07]
jastand let's not even mention WysiwygPlugin [14:09]
MichaelDaumsame toolbars, menus and popups, different callbacks ... [14:09]
jastspoiler: using the same WysiwygPlugin for TMCE and for CKE will make users of one of the two extremely unhappy [14:10]
gac410Considering the task wysiwygplugin attempts to accomplish, it's amazing it does as well as it does. [14:11]
jastthe only problem is that it's a huge ball of sticky tape :) [14:11]
MichaelDaumgac410, I've built a new FastCGIEngineContrib at Extensions/Testing including a {CheckLocalSiteCfg} switch [14:13]
gac410Ah... great thanks. I'll install it on Foswiki.org .... {puts fingers in ears}
right now we are a bit behind ... running 0.9.5, extensions has 0.96
I can't believe how far configure has come .... hits you when you have to go back and run 1.1.9 configure
here it goes...
[14:15]
jastI haven't used the new configure yet
I fear that when I do, the pain when working on all the legacy installs will only increase... :}
[14:19]
MichaelDaumwe still have "Study Webserver" [14:20]
gac410yeah... That needs work
I want to move it to a restricted System web topic.
Well the site is still up. Maybe I should reload apache just to be sure all the new code gets in there.
Okay. apache is restarted. So far so good.
[14:20]
MichaelDaummy configure has got FindBin problems
calling jsonrpc/configure
[14:24]
gac410hm. New issue? [14:24]
MichaelDaumrelated to virtualhosting I guess [14:25]
gac410ah ... I never touch that. It did too many unnatural acts in manipulating the config for my comfort :D [14:26]
MichaelDaumFindBin croaks with "Cannot find current script"
well
[14:27]
gac410Ooo MichaelDaum I see you get the 2014/2015 letter up on blog.foswiki.org. Figured out the issues with the blog? [14:27]
MichaelDaumyea
and deleted 32k zombie users
and changed commenting to use disqus
[14:27]
gac410wow. [14:28]
MichaelDaumto avoid local registration [14:28]
jastit was disqus before, too [14:28]
MichaelDaumand made it look a bit more Foswiki-ish [14:28]
jastwonder what changed there
before = several months ago when I looked at that
[14:28]
gac410I'm afraid I have made foswiki.org registration a bit too restrictive. ... but not gonna discuss the restrictions here or the bots will figure it out
New users still do get through but the bot registrations have been very effectively killed.
[14:28]
MichaelDaumnot sure disqus was in place as it required a new plugin and blog.foswiki.org to be registered at disqus.com [14:29]
gac410blog.foswiki.org .... needs a link back to foswiki.org. Can the logo in the top pop you back to f.o? [14:29]
MichaelDaumI disabled registration for now. New users only via invite.
gac410, all of the horiz menu "About, .... ,Download" link back to f.o.
[14:30]
gac410Ah... okay
okay.... didn't think to clicky them. as I said, never underestimate user getting confused.
gac410 wonders if we should remove some of the "Why the fork" dirty laundry. Enough time is gone, time to let our roots become more obscure. New users should be "who cares"
[14:30]
MichaelDaumI could imagine you didnt require "Download"
gac410, +1
[14:33]
gac410require Download? ... referring to registration? [14:35]
MichaelDaumyou are on git [14:37]
gac410er... yes?
gac410 has completely lost this thread
need more coffee
[14:37]
MichaelDaumbad configure option 'M' in lots of plugins [14:41]
gac410Oh... yeah I meant to mention that to Crawford. We can't just invalidate old syntax and crash
That one needs an urgent task.
[14:41]
MichaelDaumI had to disable FindBin in lib/Foswiki/Configure/Pluggables/SCRIPTHASH.pm to make it work under virtualhosting+fcgi [14:42]
gac410And add to that. CHECK="emptyok" needs to be the default. Otherwise lots of old config variabled error with "must not be empty" [14:42]
MichaelDaumstill lots of duplicate error messages [14:42]
gac410Why isn't $Foswiki::cfg{ScriptDir} set?
Findbin is only needed if config is mis-configured I think. SCRIPTHASH needs to find the list of installed scripts.
Even my old 1.1.9 configs have a ScriptDir configured.
"Study web server" crashes for me. 500 error - you must have permission to execute. Somehow session data is not passed along so the request fails. hm....
Study Web Server should be a blocker too.
[14:44]
Michael, Item13117 is for the Spec file issues. Extensions with incorrect spec files cause configure issues I'll make urgent. and reword to foswiki 1.1. spec files have issues.
http://foswiki.org/Tasks/Item13117
[14:55]
MichaelDaumwhat about http://foswiki.org/Tasks/Item13214 [14:55]
gac410I can live with that. The form works. May not be beautiful, but it is functional.
Task has a screenshot but nothing specific about what needs to change.
[14:56]
Okay Item13117 has been escalated to urgent with a proposed solution.
For Item13214, my sense of web design aesthetics is utterly worthless. If you can tweak the layout a bit that works for me ;)
[15:04]
hm. I have no idea how to deal with "M" -> negate CHECK="emptyok" We inverted our default in 1.2. 1.1, empty values are okay unless M specified. 1.2, empty is error unless CHECK="emptyok"
This change is really ugly.
Extensions .spec files need to exist on 1.1 and 1.2, we can't have different defaults.
we really need a big change here. CHECK="notempty" and assume "emptyok" Then we can have legacy "M" option set "notempty" CDot. .... this one needs your guidance.
probably a simple change. But needs a theological discussion :D
[15:16]
I think we leave "emptyok" as a valid option even if a nop. And the "notempty" is the active value. [15:32]
........ (idle for 36mn)
GithubBot[distro] MichaelDaum pushed 1 new commit to master: http://git.io/rFt8Pg
distro/master 2aac251 MichaelDaum: Item13214:Item13084 ease node layout in form...
[16:08]
***GithubBot has left [16:08]
...... (idle for 26mn)
gac410gac410 has a patch ready to go, changes default to emptyok, and activates an option "notempty" to replace the old "M" option
But still has to figure out parsing order to translate "M" option into "notempty" check.
[16:34]
.... (idle for 16mn)
CDotWysiwygPlugin is f***ing wonnerful. I challenge anyone else out there to write a better one. [16:51]
gac410I absolutely agree with that! [16:51]
CDoterm, gac410, all options to check are intended to be negatable
the inverse of emptyok is noemptyok
*not* notemptyok
*not* notempty
[16:51]
gac410Ah... okay. we have an utter disaster currently dealing with 1.1 era spec files.
M option needs to == noemptyok And we need the default to be emptyok Or this is going to be a real mess.
many many existing spec files use M to indicate mandatory. And spec files need to still work on 1.1
I agree what you did was Utopian but we gotta keep the 1.1. sites functional, and not re-release all the old plugins.
[16:52]
CDotok, that makes sense. I wasn't sure what to do with "M". Doesn't it have to be noundefok as well? [16:57]
gac410I'll quit while I'm ahead ( == didn't check in anything :) ) I don't understand all of this without making a real mess of it. [16:57]
CDotyour simplest interpretation is probably correct. The code is all fairly simple.
(he lied)
I inverted the default because there were so many hidden errors resulting from the previours default
[16:58]
gac410yes I think so. But so far have not had issues with undefok. That was new in 1.2, so we can make it anything we want I think
that is undefok
yeah.. no string UNDEFINEDOK found in 1.1. So invent away ... Undefined was more just happenstance in 1.1,
We only found issues with it when Timothe forced everything to be defined. which caused errors on some stuff that treated undefined special
I *like* the new default, but I don't think we can live with it due to migration issues.
Oh.. and through my testing, we can't assume '' + ! emptyok == omit from LSC. That triggers errors in checkers with undefined values. The JQuery DefaultPlugins is one instance.
gac410 is stashing his changes ... aka staged for bitbucket :)
[16:59]
MichaelDaumtoo many undef not okay [17:07]
CDotI didn't think it would be totally plain sailing; I was aiming for "consistent and documented" [17:07]
gac410But my changes were: add CHECK="noemptyok " for any element that had "M" in 1.1 spec plus the hacking to add notempty which is now all bogus [17:07]
MichaelDaumCHECK="emptynotok" sounds better [17:07]
CDotMichaelDaum can boild his head. It's noemtpyok.
noemptyok
[17:07]
MichaelDaumwhat about noemptynotok [17:08]
CDotactually, now you come to mention it, that's pretty confusing
fullok?
[17:08]
MichaelDaumabsolutelyok [17:08]
gac410I think that for best documented compatibility with 1.1 emptyok and noundefok needs to be default. ie, allow empty values, but don't undefine them unless explicit [17:09]
CDotnoemptyok needs a comma :-(
if someone can give me better terms than "emptyok" and "undefok" I'd be happy to change
bearing in mind they have to have inverses though having "no" prepended
[17:09]
gac410how many angels dance on the head of that pin ??? [17:10]
MichaelDaumnodonttouchnotokayonmonday [17:10]
CDotso "emptyok" could be "novaluerequired"
except it doesn;t mean that
"nonullstring"
[17:10]
gac410empty really is a "string" thing. but in layman terms. [17:11]
MichaelDaumI better leave now. see you tomorrow. [17:11]
CDotciao [17:11]
gac410I'd say live with empty and undef
see ya MichaelDaum have a good evening
:P
[17:11]
CDoty, empty works for me, but "noemtpyok" could mean "noempty" "ok" or it could mean "no" "emptyok"
ah, buggerit, they can read the doc if they are unsure
[17:12]
gac410Y. This is developer docs, not users. The important part is that the messages in configure UI are meaningful. [17:13]
CDotaye
ok, so you are saying we need to support "emptyok" "noemptyok" "undefok" "nuundefok" ?
and "M" needs to auto-map to "noemptyok"?
[17:14]
gac410yes. [17:15]
CDotthe logic (which I inherited) for CHECK is quite nasty for doing that.
Why are there multiple CHECK fields, for e.g.?
CDot hasn't seen a valid usage yet
[17:15]
gac410No idea. Timothe was building this as a platform for his event scheduler tool iirc. [17:16]
CDotok. Do you know of any valid usage for multiple CHECK?
or multiple FEEDBACK?
e.g., inheritance?
cos if not, I'm inclined to massively simplify it
[17:17]
gac410I can't imagine any. at least for core and default. Especially if CHECK="undefok" CHECK="blahter" can be written CHECK="undefok blather" [17:18]
CDotinheritance was the only case I could think of for it [17:18]
gac410FEEDBACK paints a button, right, So I could imagine multiple buttons
FEEDBACK=validate file syste, FEEDBACK=fix file system ??
[17:18]
CDotyeah, fair point, there's one FEEDBACK per wizard button
CDot checks for shared code
[17:19]
foswiki_irc4GooooooooOooOoOoooOOooOod Morning [17:20]
gac410btw, you'll hate me, but occasionally I still encounter {plugin}{Module} = '\'...\'' ie double quoting. :P [17:20]
CDotLoo: good evening! [17:20]
gac410and Good Afternoon [17:20]
CDot"still encounter" means what? [17:21]
gac410Occasionally after updating config, I'll find a $Foswiki::cfg with escaped quotes for a module. and errors in InstalledPlugins
gac410 tries to recreate
[17:21]
CDotCDot needs to be able to reproduce anything like that, otherwise it's fishing for polar bears in a whiteout. [17:22]
gac410hang on... My config is totally f'd due to my hacking of nonudefnemptynuffin [17:22]
CDotno hurry; I'm looking at "M" [17:22]
gac410I coded up M, and it was a nop, so I'm clueless there. :) [17:23]
CDotyou did something like this, right? M => { parse_val => '_MANDATORY' } [17:24]
gac410oh... no... i just defined M as {} and then in the CHECK parser, I did if $this->{M} [17:25]
CDotok [17:25]
gac410Ah... got the Module issue. If you use the [<] button to "set to default" it sets it to '...' which results in '\'...\'' [17:27]
CDotum. "noemptyok" is the default, so M is a nop, isn't it?
or were you thinking of inverting the default again?
[17:27]
gac410gawd I hate inverted logic. Okay on 1.1, M is NOT the default, so 1.2 should default to non-mandatory or emptyok [17:28]
CDotso what do you have against my better default?
CDot thinks everything should be specified
[17:29]
gac410nothing. if 1.1 didn't exist. [17:29]
CDotok, so you are concerned about the config.specs coming from extensions, right? [17:30]
gac410The problme is that we have 1.1 where the default is to write out empty '' values, and 1.2 where we delete them, and also we error every damn extension with empty values. [17:30]
CDotif you have already been through and "fixed" Foswiki.spec for your "new" default, can you check that in please? [17:30]
gac410okay. ... I gotta get it back from stash.
hm... okay... so my notempty must change to "noemptyok" correct?
Spec file with M on 1.1. == CHECK="noemptyok" on 1.2
[17:31]
CDotcorrect
undefok=>0 is an OK default, is it?
[17:32]
gac410Yes. very few fields will need noundefok [17:33]
CDotright. Code's done. [17:34]
gac410Fixing under Item13117
want me push first?
[17:35]
CDotplease [17:37]
gac4108 elements get noemptyok, and ScriptUrlHash gains emptyok,
okay pull/push done
[17:37]
GithubBot[distro] gac410 pushed 1 new commit to master: http://git.io/5gsHuw
distro/master 98743e8 George Clark: Item13117: Set noemptyok for M 1.1 items...
[17:37]
***GithubBot has left [17:37]
gac410er.. ScriptUrlPath [17:37]
CDotanr the {view} one? [17:38]
gac410hm... I think that was already set
yes. that one is already undefok emptyok
[17:38]
CDotok [17:39]
gac410no, undefok emptyok :P
:D
Looks like the invert code doesn't work: CHECK parse failed: unrecognised option 'noemptyok'
[17:39]
CDotnot yet
working on it
patience
[17:40]
gac410oh ... okay.
I was wondering when you said that any option could be inverted, as I didn't see that.
[17:40]
CDotit's there, but the test was case-sensitive to it had to be NOemptyok
... which has been fixed
[17:41]
gac410ah.. hm... well that did make it clear that the NO applied to the "emptyok" as opposed to "noemptyOK"
but very unusual syntax, to be case sensitive. So noemptyok is fine
[17:43]
CDotok [17:43]
gac410this is all developer docs, so arcane is fine. :) [17:44]
....... (idle for 32mn)
Lynnwoodhere’s my next iteration of revising SupportQuestions: http://foswiki.org/Support/SupportQuestionsRedesign
It only does a single search per load now so loads fairly quickly.
[18:16]
gac410Nice... Much better to put the answers front & center, vs all the languishing unanswered questions. [18:17]
LynnwoodI’d welcome any suggestions about getting approval to replace the current SupportQuestions page with this version. At one time, this kind of change would inspire a lively debate and extensive reviews. Now… i don’t think so much so. [18:17]
gac410Any thought on a twisty per category vs. the fully expanded list?
Much easier to get forgiveness than approval :D
[18:17]
Lynnwoody, now it is. I don’t think many will notice.
When you say category, are you referring to the Subject Area list or the status categories?
[18:18]
gac410"This website" "Installation of Foswiki" ... Have them open/close? [18:19]
LynnwoodIf it’s Subject Area, i don’t think that would add much since there is a toc to jump to an area. [18:19]
gac410true. And hiding behind twisty makes stuff too obscure for some. [18:20]
LynnwoodFor some status queries, some of these list are pretty short.
(if there are none under a particular subject, it doesn’t show up at all)
[18:20]
gac410nice
Someday sometime we also need to scrub out some of the very old / obsolete questions. Some refer to issues long since fixed / rewritten.
[18:20]
Lynnwoodindeed.
i do think we could clear out a lot of the asked questions by moving them to “need more info” and such.
and then clear out some of the need more info ones after a period of time if no further info is provided.
[18:21]
gac410FrequentlyAskedQuestions needs some consideration too. Seems like we would want people to *start* there, then to answered questions, ... [18:23]
Lynnwoodright...
i could do that with a simple change on Support.WebHome.
done
[18:23]
I’m going to go ahead and give the new SupportQuestions page a try. If someone really cares, they are welcome to switch it back or revise it. [18:35]
gac410sounds good to me.
I've got the same thing Pending with ApacheConfigGenerator24 .... Looking for someone to validate the config with Apache 2.2, and 2.4 besides just me.
[18:36]
Lynnwoodinteresting… i will take a look at them although i frankly don’t remember what versions i’m running on various servers. [18:42]
gac410The big change is Apache 2.4 redesigned the access controls. mod_access_compat *should* let a 2.2 config run on 2.4, but I found one bug in our configurations
You are not allowed to code an AuthType if nothing requires authentication. So a Template auth config, will error out on 2.4, unless at least configure requires auth, or you comment out AuthType
[18:43]
............ (idle for 57mn)
Loo2is there a way to make the table of contents dynamic?
because I am using Jquery to remove some elements
but it breaks the TOC
[19:41]
gac410No, not that I'm aware of. The TOC is generated on the server in perl. Anything jquery changes is on the client. You'd need to remove it from the TOC as well on the client side. [19:42]
Loo2well, its breaking functionality completely
nvm
i'll look into
its pretty specific issue
for some reason the twisty I put around the TOC is breaking
not sure why
[19:43]
gac410Ah... okay. twisty can render as a <div> or as a <span> To put it around the toc, needs to render a <div> though the mode defaults to div so that should be okay
TWISTY around a TOC by default works fine. http://foswiki.org/Sandbox/TestTOCTwisty
If stuff breaks in HTML rendering, one thing to be very careful about is proper open/close of block elements Every <div> needs a </div> etc.
[19:45]
CDotgac410: done [19:51]
GithubBot[distro] cdot pushed 1 new commit to master: http://git.io/AF28Xg
distro/master 8edbfc5 Comment: Item13117: restore support for M (mandatory) and switch the default over to emptyok. Reduce complexity by only supporting one CHECK set
[19:51]
***GithubBot has left [19:51]
gac410wow... more involved tha I realized. [19:51]
Loo2$("body").html($("body").html().replace(/asdfasdf/g,"_some_UnIqUe_string_"));
this is breaking the TWISTY
and its not functionally doing anything...
(there is no asdfasdf string
[19:55]
gac410I have no idea, you just went beyond me...
MichaelDaum is our jquery expert, and he's in europe so not online currently.
[19:56]
Loo2its actually seems to be breaking jquery code elsewhere too
not sure why though
[19:57]
LynnwoodLoo2: what are you trying to do again? Delete items from the TOC? [19:58]
Loo2its kinda complicate
nothing about the js itself is too crazy
but when i run any .html().replace(//g,"")
it breaks the other javascript elements
i am running these from the developers console
$("#patternMainContents").html($("#patternMainContents").html().replace("/asdfasdf/g","_some_UnIqUe_string_"));
[20:04]
***ChanServ sets mode: +o MichaelDaum [20:06]
Loo2that is what I am working with now
and "asdfasdf" is not present in the HTML
yet all other js (that I can tell) stop working after I issue that command in the developers console
[20:06]
Lynnwoodthat is odd. the line you showed above looks fairly straight forward. Can’t see how it would impact of js. [20:10]
Loo2thats what I am thinking [20:12]
Lynnwood…although i’m far from a js wiz. [20:12]
Loo2well, here seems to be the issue
when i do a replace, these get removed: <link rel="stylesheet" href="/pnpWiki/pub/System/PatternSkinTheme/jquery-ui.css" type="text/css" media="all" /><!--JQUERYPLUGIN::THEME: requires= missing ids: JQUERYPLUGIN::UI--> <link rel='stylesheet' href='/pnpWiki/pub/System/TwistyPlugin/twisty.css?version=1.6.0' type='text/css' media='all' /><!--JQUERYPLUGIN::TWISTY: requires= missing ids: JavascriptFiles/foswikiPref-->
which seem pretty important to me
[20:13]
Lynnwoodyou might narrow the scope of the search to the topic body
but i can’t see why the string you presented should impact the text you mention above.
unless that’s not really the string you’re searching...
btw, in .replace(), you don’t need the quotes around the search string. .replace(/asdfasdf/g, "_some_UnIqUe_string_") should be ok.
i was working with .replace() recently and had some special characters in the search string and had to escape them.
[20:15]
Loo2ya, it doesn't make sense to me
im using the id for the topic text
so it shouldn't be altering the header
[20:20]
well F
the HTML is changing just as I expected
don't know why its breaking functionality
maybe if I try adding it to the actual javascript instead of doing it in the console...
[20:25]
GithubBot[distro] gac410 pushed 1 new commit to master: http://git.io/r8ZFEQ
distro/master 1fbf9c9 George Clark: Item13117: CHECK is no longer an array....
[20:32]
***GithubBot has left [20:32]
Loo2welp
it worked when i put it in the view.mySKin.tmpl under %ADDTOZONE%
I'm not entirely sure what thats about...
[20:32]
gac410yeesh another day another urgent configure task. All of the ValidatePermissions checks appear to be totally broken. Every file, every directory errors on my system. Somehow this got broken in the remodeling.
Well CDot ... I found where multiple CHECK clauses exist. Damn. File permission validations, One CHECK clause per filter.
[20:40]
................ (idle for 1h18mn)
GithubBot[distro] gac410 pushed 1 new commit to master: http://git.io/d6WH1Q
distro/master e11f80a George Clark: Item13117: CHECK is no longer an array....
[22:00]
***GithubBot has left [22:00]

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