#foswiki 2016-12-19,Mon

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

WhoWhatWhen
***ChanServ sets mode: +o cdot [06:14]
.............. (idle for 1h8mn)
ChanServ sets mode: +o MichaelDaum [07:22]
..................... (idle for 1h44mn)
GithubBot[WorkflowPlugin] cdot pushed 3 new commits to master: https://git.io/v1drY
WorkflowPlugin/master 0bdaf4a cdot: Item14210: support notification when a topic is stuck in a state
WorkflowPlugin/master 97f7191 cdot: Item14210: regex fu deserted me
WorkflowPlugin/master 1a1fa4f cdot: Merge branch 'master' of github.com:foswiki/WorkflowPlugin
[09:06]
***GithubBot has left [09:06]
FoswikiBothttps://foswiki.org/Tasks/Item14210 [ Item14210: Support re-notification when a topic is stuck in a state ] [09:07]
.................................... (idle for 2h59mn)
MichaelDaumbefore 2.x, unicode chars have been stripped off from a formfield name. now we allow arbitrary unicode chars in there: see this code change from June 2015: https://github.com/foswiki/distro/commit/bfe3afe3e63c6ac86d18e93f2944914f370fe7da
problem is when people move content from an old to a new site: they run into potential data loss due to the new naming scheme
there is no way to switch back into the old way other than reverting this particular patch
[12:06]
https://foswiki.org/Tasks/Item14256 [12:13]
............ (idle for 57mn)
***ChanServ sets mode: +o Lynnwood [13:10]
ChanServ sets mode: +o gac410 [13:19]
gac410MichaelDaum: I have a fix for Item14253, in the patch. But Wysiwsg white space is such a minefield, I'd like cdot to confirm it. [13:22]
FoswikiBothttps://foswiki.org/Tasks/Item14253 [ Item14253: wysiwyg inserts extra spaces infront of a link ] [13:22]
gac410Regarding Item14256, I was going to build 2.1.3 tomorrow. :( [13:23]
FoswikiBothttps://foswiki.org/Tasks/Item14256 [ Item14256: data loss when moving content from 1.x to 2.x ] [13:23]
MichaelDaumHi gac410. k.
I know I know. Yet still this is serious.
I am bit of chocked nobody noticed earlier.
[13:23]
gac410On the white space, I can't figure out why the $WC::CHECKw is needed. But it's been in the code since tw* ... I can't see any impact of removing it other than fixing your bug.
White space is still preserved if present. Just not added. Doesn't make sense.
[13:24]
MichaelDaumy [13:25]
gac410MichaelDaum: Rather than a switch, what about a utility that to scan and "Change a Formfield name" [13:25]
MichaelDaumyea [13:25]
gac410we already have to scan to change deprecated ACLs so what's one more. [13:26]
MichaelDaumthe code change to get legacy formfield names back again is minimal though [13:26]
gac410Would you want it globally - a LocalSite.cfg change? [13:26]
MichaelDaumthat's what I was thinking of for this particular client: minimal change that solves the problem. I do see the flip side of the coin too. [13:27]
gac410we can treat it like the ACls as well. DisableUnicodeFormfieldnames
disabled by defaut, recommend scan/rename to resolve, or enable to use old operation.
[13:27]
MichaelDaumah good point
interestingly the feature request dates back a decade though nobody ever discussed data incompatibilities. shame on us all.
[13:28]
gac410We call the ACL one: $Foswiki::cfg{AccessControlACL}{EnableDeprecatedEmptyDeny} = $FALSE; [13:29]
MichaelDaumwe also have {LegacyRESTSecurity} [13:29]
gac410No {DataForms} key ... but we have {FormTypes} Maybe time to add {DataForms}{LegacyFormfieldNames} [13:31]
MichaelDaumI am a big fan of naming parameters in a way their name does not include a negation... such as DisableFooBar, or nonoise, etc
LegacyFormfieldNames is perfect imho
[13:31]
gac410yeah you've caught me on others :D
At some time {FormTypes} should be moved to {DataForms}{FormTypes} ... to keep the LSC more organized. But not really a big need.
[13:31]
MichaelDaumdon't get me started on FormTypes...I'd opt for removing this useless feature
there aint no code using it as far as I know
this was mid way Sven working on a data form wizard kinda stuff.
something that never materialized
[13:33]
gac410Anyway ... cdot. if you could just take 5 minutes to review my plan to remove the $WC::CHECKw from      return ( 0, $WC::CHECKw . '[' . $nop . '[' . $href . ']]' ); - when generating [[squab]] links in Node.pm It would be greatly appreciated.
Confirmed MichaelDaum - FormTypes is not used in core except for the Config Checker, Unit Tests, and Documentation ( VarQUERY
I'll open a task to remove it from Foswiki 2.2 We really need to strip this stuff out when we find it.
Yup. documented %QUERY{"{FormTypes}[].types"}%  does not do anything. Damn.
[13:34]
cdotgac410: would that be a cunning plan? [13:42]
gac410I don't know. White Space and Wysiwyg scares the crap out of me :D No good deed goes unpunished in this area of the code.
It seems like such a simple fix. And I can't find any downside. But someone put it there for a reason I'd gather.
Unless maybe older versions of TMCE were eating whitespace. But so far it seems clean.
[13:43]
jastthe markup doesn't seem to require a check there
so my first impulse would be to agree with you
[13:46]
cdotI think the CHECKw was there for compatibility with old regexes.
AFAICT it's not required.
i.e. your fix is correct.
[13:46]
gac410Ah. okay. excellent. Great minds all converge. I'll strip it out and hopefully MichaelDaum will be happy :D
We'll see how badly I get punished for this one
[13:47]
jastwell, given your past with WysiwygPlugin... :D [13:48]
gac410I did check for growing / disappearing white space, and it does seem stable without the $WC::CHECKw
I think there are 3-4 Wysiwyg releases in my past that were solely due to whitespace debacles
[13:49]
cdotit should be. The CHECKw just checks that there is delimiting space there.
y, handling WS is very hard, due to the regexes that "parse" TML.
[13:50]
jastplus all the encoding things into HTML that HTML isn't supposed to care about [13:51]
gac410So a squab is valid pretty much anywhere with or without whitespace. So link[[me]]here is perfectly legal, as well as Michaels "[[LinkMe]]" and that $WC::CHECKw was preventing it. [13:52]
cdotas is [[[Link me]] - but does it link "[Link me" or "Link me"? [13:53]
gac410ugh. cdot mercy please. No puzzles like that one [13:54]
jastI believe most of the regexes are greedy? so, the former, at least with the patch (assuming the URL parser isn't strict enough to catch '[' as invalid)
actually it's not invalid in a relative URL, so disregard that last part
[13:54]
cdotno, it's not invalid; and it's doesn't really matter, so long as it is handled consistently. [13:57]
GithubBot[distro] gac410 pushed 1 new commit to Release02x01: https://git.io/v1Feb
distro/Release02x01 ec987f2 George Clark: Item14253: Don't insert extra white space before [[links]]
[14:00]
***GithubBot has left [14:00]
FoswikiBothttps://foswiki.org/Tasks/Item14253 [ Item14253: wysiwyg inserts extra spaces infront of a link ] [14:00]
gac410Okay ... that's fixed. MichaelDaum Can you have a fix for your Formfield name issues sometime today so I can work on 2.1.3 tomorrow?
Opened Item14257 for the {FormTypes}
[14:00]
FoswikiBothttps://foswiki.org/Tasks/Item14257 [ Item14257: VarQUERY and Configure document ={FormTypes}= which is not implemented. ] [14:02]
gac410Thinking maybe I should fix the documentation in 2.1.3, but remove the configure keys in 2.2 ... just in case. [14:03]
.... (idle for 15mn)
GithubBot[distro] MichaelDaum pushed 1 new commit to Release02x01: https://github.com/foswiki/distro/commit/47238cf68eef3af218d0b99c02895aafcd45626b
distro/Release02x01 47238cf MichaelDaum: Item14256: implemented {LegacyFormfieldNames} flag...
[14:18]
***GithubBot has left [14:18]
FoswikiBothttps://foswiki.org/Tasks/Item14256 [ Item14256: data loss when moving content from 1.x to 2.x ] [14:18]
GithubBot[distro] MichaelDaum pushed 1 new commit to master: https://github.com/foswiki/distro/commit/c5d298f0e728fd58f002a0757b6c1a2124ce810b
distro/master c5d298f MichaelDaum: Merge remote-tracking branch 'origin/Release02x01'
[14:19]
***GithubBot has left [14:19]
............................ (idle for 2h17mn)
ChanServ sets mode: +o cdot [16:36]
........................ (idle for 1h57mn)
gac410Darn.. TestCases - TestCaseAmISane is failing [18:33]
............................................................ (idle for 4h56mn)
Lynnwood: I noticed that https://github.com/hhurz/tableExport.jquery.plugin/network is a much newer and actively maintained version of the export code. [23:29]

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