#foswiki 2017-09-25,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 Lynnwood__ [02:12]
........ (idle for 36mn)
ChanServ sets mode: +o Lynnwood [02:48]
........................................................... (idle for 4h54mn)
ChanServ sets mode: +o MichaelDaum
MichaelDaum changes topic to: Download: https://foswiki.org/Download - Logs: http://irclogs.foswiki.org/bin/irclogger_log/foswiki - Slack: https://foswiki-slackin.herokuapp.com/ - Bugs: https://foswiki.org/Tasks - Next Release meeting Monday 2 Oct 2017, 1300Z
[07:42]
............................. (idle for 2h20mn)
GithubBot[JQDataTablesPlugin] MichaelDaum pushed 1 new commit to master: https://git.io/vdktd
JQDataTablesPlugin/master dd7b0b9 MichaelDaum: Item14496: fixed html5 data attribs
[10:03]
***GithubBot has left [10:03]
FoswikiBothttps://foswiki.org/Tasks/Item14496 [ Item14496: error generating html5 data attributes ] [10:03]
..... (idle for 23mn)
FoswikiOnSlack<itaki> In order to make sure, I would like to ask if it is really possible to use the filtering function of JQDataTablesPlugin for the wiki tables, as we did in MS Excel. This is really what I am looking forward to seeing in Foswiki. [10:26]
..... (idle for 23mn)
<nuddlegg> @itaki there's no XSL or XLSX connector for JQDataTablesPlugin yet. Sponsors welcome :slightly_smiling_face: [10:49]
GithubBot[XSendFileContrib] MichaelDaum pushed 1 new commit to master: https://git.io/vdkON
XSendFileContrib/master 636e421 MichaelDaum: Item14497: show txt files inline by default
[11:01]
***GithubBot has left [11:01]
FoswikiBothttps://foswiki.org/Tasks/Item14497 [ Item14497: show txt files inline by default ] [11:01]
GithubBot[SendEmailPlugin] MichaelDaum pushed 1 new commit to master: https://git.io/vdk3u
SendEmailPlugin/master d743dea MichaelDaum: Item14498: use newer switchboard settings
[11:09]
***GithubBot has left [11:09]
FoswikiBothttps://foswiki.org/Tasks/Item14498 [ Item14498: use newer switchboard settings for scripts ] [11:09]
GithubBot[MultiSaveContrib] MichaelDaum pushed 1 new commit to master: https://git.io/vdksm
MultiSaveContrib/master 6fae466 MichaelDaum: Item14499: make proper use of $request->param()...
[11:16]
***GithubBot has left [11:16]
FoswikiBothttps://foswiki.org/Tasks/Item14499 [ Item14499: make proper use of $request->param() ] [11:16]
GithubBot[OEmbedPlugin] MichaelDaum pushed 1 new commit to master: https://git.io/vdksg
OEmbedPlugin/master 5b3537f MichaelDaum: Item14500: add REST handler for cache maintenance
[11:21]
***GithubBot has left [11:21]
FoswikiBothttps://foswiki.org/Tasks/Item14500 [ Item14500: add REST interface to clear the local cache ] [11:21]
GithubBot[PiwikPlugin] MichaelDaum pushed 1 new commit to master: https://git.io/vdksb
PiwikPlugin/master f7a8d09 MichaelDaum: Item14501: fixed broken init scripts
[11:26]
***GithubBot has left [11:26]
FoswikiBothttps://foswiki.org/Tasks/Item14501 [ Item14501: broken init scripts ] [11:26]
GithubBot[GoogleMapsPlugin] MichaelDaum pushed 1 new commit to master: https://git.io/vdkGo
GoogleMapsPlugin/master e7d2459 MichaelDaum: Item14502: fixed sending the API to google
[11:34]
***GithubBot has left [11:34]
FoswikiBothttps://foswiki.org/Tasks/Item14502 [ Item14502: error sending the API key to google ] [11:34]
GithubBot[GridLayoutPlugin] MichaelDaum pushed 1 new commit to master: https://git.io/vdkZc
GridLayoutPlugin/master 03f9312 MichaelDaum: Item14503: improve mobile device layout...
[11:43]
***GithubBot has left [11:43]
FoswikiBothttps://foswiki.org/Tasks/Item14503 [ Item14503: proper clearfix and border of columns all being laid out verticaly on mobile devices ] [11:43]
GithubBot[PageOptimizerPlugin] MichaelDaum pushed 1 new commit to master: https://git.io/vdknC
PageOptimizerPlugin/master 4c32e54 MichaelDaum: Item14504: added http/2 push headers
[11:53]
***GithubBot has left [11:53]
FoswikiBothttps://foswiki.org/Tasks/Item14504 [ Item14504: add http/2 push headers ] [11:53]
***ChanServ sets mode: +o Lynnwood [12:02]
.... (idle for 16mn)
zak256Can I use Fast CGI and mod_perl both at the same time?
I think I read that somewhere that only one is possible, but cannot remember where...
[12:18]
***ChanServ sets mode: +o gac410 [12:19]
MichaelDaumzak256, for foswiki? [12:19]
zak256Yes [12:19]
MichaelDaumwhy would you like to do this?
mod_perl is so 80ies ;)
[12:20]
zak256To get all possible performance gain I can get. [12:20]
MichaelDaumthen just use fcgi [12:20]
zak256Really? Is it so outdated? [12:20]
MichaelDaumit is
as well as conceptually flawed
[12:21]
zak256Okay, will do that then. [12:21]
MichaelDaumat the moment you will get best performance using nginx+fcgi
psgi will challenge that ... some day when foswiki 3.0 become a reality
[12:22]
zak256By the way, I opened Tasks.Item14495 . Can anybody say something about that? [12:23]
FoswikiBothttps://foswiki.org/Tasks/Item14495 [ Item14495: Including a table seems to break EDITTABLE quick edit feature ] [12:23]
MichaelDaumsorry, don't use EDITTABLE at all
^I don't use^
[12:24]
zak256hmm... anybody else? [12:25]
MichaelDaumperl 5.26.1 is out see changes at https://metacpan.org/changes/release/SHAY/perl-5.26.1
comes with 3 CVEs
[12:29]
gac410I ran unit tests with 5.27.4 yesterday. Saw very strange failures in RegisterTests. All of the Email headers had "Quoted Name" <quotedname@...> and test expected bare names.
5.26.0-RC1 I had hanging around - no quotes.
[12:32]
MichaelDaumlet's see whether/when distros will port these security fixes back to their outdated perls :>( [12:33]
gac410right... [12:35]
.......... (idle for 49mn)
vrurgMichaelDaum: to mod_perl/fcgi discussion: due to FCGI's stupid STDERR handling it doesn't work for me. Will have to use mod_perl instead. [13:24]
MichaelDaumstupid stderr handling? what daya mean? [13:27]
gac410Even I know that one. It prevents / blocks any ability to tie STDERR. :D [13:28]
MichaelDaum"tie stderr" whots dat? [13:29]
vrurgFCGI ties STDERR on its own. And doesn't implement 'OPEN'.
In my case Expat module fails because it's heavily based on duping filehandles.
s/Expat/Expect/
[13:30]
MichaelDaumah thats special
basically I don't care much about stderr other than looking at error.log
[13:31]
gac410We ran into this with the email config wizard. If you are under fcgi, configure is not able to report STDERR when you run the wizard. have to check the logs yourself. some issues really need stderr to diagnose. [13:32]
vrurgFirst time I got into this problem was related to autoinstallation of dependencies. Which is not that special. [13:32]
gac410You also tried to fix the email wizard for me. ;) [13:32]
MichaelDaumhow about forking a testing script rather than trying to use configure's perl ... and presumably breaking it [13:33]
vrurggac410: I just don't remember now which one was first. :) [13:33]
MichaelDaumFoswiki::Sandbox is just fine capturing stdout and stderr of an external script ... [13:33]
vrurgemail I guess because it took quite a time to find the cause. [13:33]
FoswikiBothttps://trunk.foswiki.org/System/PerlDoc?module=Foswiki::Sandbox [13:33]
vrurgMichaelDaum: a workaround is always possible. But that's pretty annoying. In my Expect case the initial code has been written before we tried fcgi.
Anyway, PSGI doesn't have those issues. ;)
[13:35]
MichaelDaumi.c.
right
still fetching stderr of a Foswiki::Sandbox::sysCommand("perl -MFoo::Bar -e "1") to test for a dependency isn't really magic
[13:36]
vrurgMichaelDaum: BTW, regarding PSGI: I mostly finished documenting new extensions in Item14237 branch. [13:38]
FoswikiBothttps://foswiki.org/Tasks/Item14237 [ Item14237: Implement Development.OOConfigSpecsFormat proposal ] [13:38]
MichaelDaumgawd, I wished I could just dive into your branch and enjoy [13:38]
vrurgMichaelDaum: not without docs. And when it comes to enjoying... Well, it still needs SO MUCH work because I had no time neither for tainting, nor for normal unicode - just to mention these two... [13:41]
MichaelDaumyet still. I am tempted to hold back any other core refactoring until after this is settled
such as kick out the current user code
[13:42]
gac410Y I do think that if we are going to actually use vrurg's code for Foswiki 3.0, then we really need to make a true commitment to it.
As it stands, it is still going to be very painful to merge master into vrurg's branch(es). But that has to happen sometime.
vrurg: MichaelDaum: Speaking of core changes. As I delve into things like nonce, password reset, and other random string generation, I'm thinking that these should become a core API which then uses a true cryptographically strong RNG
[13:43]
vrurggac410: Just: there is no more 'branches', only
branch.
[13:45]
gac410Ah... great! [13:45]
vrurg14237 is the one I've put all my focus into. [13:46]
gac410Okay someday I shall try some very small merges into 14237. The way to make it manageable:
branch master 5-10 commits ahead of where the last merge occurred. Merge *that* branch, then toss it and move a few more ahead.
lather/rinse/repeat until it's caught up. Painful, but a one-step merge of master ends up with massive conflicts to resolve.
Cherry pick might be an alternative, but then you have no good record of which commits are in which branch.
[13:47]
vrurggac410: No way that massive merge will work, you're right. Not only the conflicts to resolve, but the initial logic behind changes could be lost due to the differences in the core code.
Anyway, I need to finish the docs. Not gonna make any new topics for now. All docs are in corresponding modules.
[13:50]
gac410git cherry Item14237 master | wc -l ==> 465 So a lot of master changes that are not in your branch. [13:51]
vrurgPerhaps later either me or somebody would compile these into new topics.
465... oops...
[13:51]
***ChanServ sets mode: +o Lynnwood__ [13:52]
gac410393 if you exclude translations [13:52]
vrurgStill enough. This is what I was always afraid of. [13:53]
gac410vrurg, If our PerlDoc topic provides usable documentation, then duplicating it into topics doesn't make a lot of sense. [13:53]
vrurggac410: Some common instructions, guidelines are still needed.
I wouldn't expect finding docs about 'how one starts; code styling and basic principles' in PerlDoc on Foswiki::Object or Foswiki::Class, for example.
[13:56]
FoswikiBothttps://trunk.foswiki.org/System/PerlDoc?module=Foswiki::Object https://trunk.foswiki.org/System/PerlDoc?module=Foswiki::Class [13:57]
vrurgAs to extensions – it's still not really clear to me if extensions manager or Empty extension is the main source of information. Etc., etc., etc.... [13:58]
gac410PerlDoc I *think* can be included. So docs could be incorporated rather than duplicated. Or we need some tweaks to that functionality.
EmptyExtension is more of a starting template than detailed documentation.
i've thought that we should have a PerlDoc equivalent that can display templates from the /templates directory
[13:59]
vrurgThis is for sure something to have. I once were missing docs on templates.
But what I do miss very much right now is an easy way to create links to sections/subsections. Especially to methods and (in my branch) attributes.
[14:02]
gac410My thoughts are that developers would be more likely to adjust documentation when it is inline with something else that they are changing. So improvements to PerlDoc is probably preferable to copy/paste docs. [14:03]
***ChanServ sets mode: +o Lynnwood [14:05]
gac410Maybe need to extend our perldoc function with section=, method=, sub= or some other way to index into or reference a modules documentation. [14:05]
vrurgI already rewrote the doc handler to take care of module names in the doc and autolink them. [14:06]
gac410Or even support #fragment-id
cool
[14:06]
vrurggac410: Actually, I totally rewrote it and it's now more of a kind of lexical parser. But after writing a plugin for Pod::Weaver I think it must use PPI to parse the code and extract pods. And some Pod parser.
Because current regexp-based approach is bad in respect to permanent Perl's syntax changes – why do we have to follow them instead of just updating the parsing modules?
[14:09]
gac410agreed [14:10]
vrurgAnd, yes, =method, =privMethod, =staticMethod, =attribute, etc. - they all must be there. Plus either a macro for linking like %L{meth:someMethod}%, %L{attr:someAttr}%, or prefixes similar to interwiki's CPAN: (like METH:someMethod) – are to take away the burden of linking for a doc writer.
I'm thinking of it all, but have no time to implement.
[14:13]
gac410well when it comes time to cut/paste docs, code is preferable :D [14:15]
vrurgOki, I'll be gone for work. Will return later today. [14:17]
gac410okay... have fun [14:18]
...... (idle for 28mn)
zak256Which apache configuration is responsible for Foswiki to find Foswiki.pm ? I changed all parameters to fcgi but something is still wrong. It looks only in /my/path/lib instead of /my/path/wiki/lib ...
I am comparing everything with the ApacheConfigGenerator and it seems right.
I am suspecting this one here: Alias /wiki/bin "${foswikiroot}/bin/foswiki.fcgi"
I wrote: Alias /wiki/bin "${SCRIPTDIR}/foswiki.fcgi"
...because I have a bin directory outside of my foswiki-directory
And apart from that for some reason there is no redirection from https://wikihost/wiki/ to https://wikihost/wiki/bin/view/ anymore. But I disabled shorturls, maybe that is the reason?
[14:46]
ok... I now see there is a FastCGIEnginContrib. I thought this is already installed because I have a lib/FCGI folder. Nevermind, I guess that's it then. [15:01]
gac410er.... FastCGIEngineContrib is bundled with Foswiki 2.x [15:01]
zak256Oh, than it is indeed? Because I don't have a file bin/foswiki.fcgi
...which is included in FastCGIEngineContrib.tgz
[15:02]
gac410You are on Foswiki 2.x? [15:03]
zak256yes [15:03]
gac410You must not have copied or installed something correctly. That file is definitely bundled / included in the foswiki release.
I just checked my clean unzip of the Foswiki-2.1.4.zip
[15:04]
zak256zak256 feels stupid
yes it's there
ok then... it must be somewhere else
[15:05]
gac410Is foswiki working other than the redirect? the location of the bin dir is used for foswiki internally to find all the other directories during bootstrap
bin/setlib.cfg iirc gets run early, and it does magic to locate lib, etc. based off of the location of "itself"
Been a while since I looked at that code.
[15:06]
zak256Is it possible that for some reason lib/ must equal to bin/../lib ? [15:08]
gac410tbh I don't recall. If foswiki is generally working, except for the redirect, then that's not the issue. [15:09]
zak256...here in setlib.cfg, line 31: ( $foswikiLibPath ) = ($foswikiLibPath = Cwd::abs_path( "$bindir../lib" )) =~ /(.*)/;
no, it's not working anymore. I think I found the issue for the redirect, but apart from that, it doesn't work
which is clear if lib is expected as bin/../lib
[15:09]
gac410That may be it then. relocating bin is not recommended. [15:10]
zak256ok... I will have to reconfigure this here then :-/ [15:11]
gac410I thought there were ways to relocate it, but tbh I don't really have time today. [15:11]
zak256Well, at least the possibility is there with cfg{ScriptDir}
and it worked with mod_perl
[15:11]
gac410mod_perl doesn't use the bin dir. [15:12]
zak256and fcgi does? ok, then everything makes sense [15:13]
gac410in LocalLib.cfg, you can configure a hard coded override for the libpath
Example in bin/LocalLib.cfg.txt
[15:14]
zak256I see...
This works. Thanks gac410
[15:14]
***ChanServ sets mode: +o Lynnwood [15:30]
.......... (idle for 46mn)
ChanServ sets mode: +o Lynnwood [16:16]
........ (idle for 36mn)
zak256 has left [16:52]

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