#foswiki 2017-05-16,Tue

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

WhoWhatWhen
GithubBot[distro] gac410 pushed 1 new commit to Release02x01: https://git.io/v9F9z
distro/Release02x01 12bcd28 George Clark: Item13883: Preparing for release.
[00:41]
***GithubBot has left [00:41]
FoswikiBothttps://foswiki.org/Tasks/Item13883 [ Item13883: Documentation changes for master and 2.1 ] [00:41]
................ (idle for 1h16mn)
***ChanServ sets mode: +o Lynnwood [01:57]
GithubBot[distro] gac410 pushed 1 new commit to Release02x01: https://git.io/v9FdO
distro/Release02x01 438d0fc George Clark: Item14234: Fix unit tests for ERP...
[02:10]
***GithubBot has left [02:10]
FoswikiBothttps://foswiki.org/Tasks/Item14234 [ Item14234: Clicking pencil for first row (erp_row=0) opens editor on entire table. ] [02:10]
.... (idle for 16mn)
GithubBot[distro] gac410 pushed 1 new commit to master: https://git.io/v9Fdp
distro/master 4804978 George Clark: Item13883: Merge branch 'Release02x01'
[02:26]
***GithubBot has left [02:26]
FoswikiBothttps://foswiki.org/Tasks/Item13883 [ Item13883: Documentation changes for master and 2.1 ] [02:26]
GithubBot[distro] gac410 tagged FoswikiRelease02x01x04_RC2 at 6d78a43: https://git.io/v9FFe [02:28]
***GithubBot has left [02:28]
vrurggac410: Are you available? [02:28]
gac410sure what's up. I'm working on getting 2.1.4 RC2 released/installed [02:28]
vrurgBriefly. In Logger::Compatibility, line 132 on master – it die's if cannot write to a log file.
To my view it's too much for a log file to crash an application.
Should there be a proposal on changing this behavior?
I think die must be replaced with `print STDERR` or anything that would send the message to web server's log.
[02:29]
gac410hm If there is a situation where the logfiles are not writable, it's possible for stuff to happen that is not attributed. You could quietly run along for significant periods with no event logging.
It's one of those things that are a severe misconfiguration - should never happen. I prefer to tell the user and refuse to allow normal operation.
[02:33]
vrurgA broadcast message would be much more preferable. I plan a bufferrized messaging/broadcasting for v3, but so far combining a front-end warning with sending the messages to web server's log must be sufficient. [02:35]
gac410This is a one-time misconfiguration that I don't think I've ever seen other then when I totally f-up directory permissions. A normally installed foswiki won't encounter it
If things are really wrong, die is appropriate IMHO.
[02:35]
vrurgThink of a bug causing logfiles to become owned by root on some configurations. Usually it happens over nightly log rotation. A user might not be able to reach the application until the admin wakes up in the morning and fixes this. Not pretty nice situation too. [02:36]
gac410NO WAY do I want the system operational with the logging quashed. No log, No changes, No view. [02:37]
vrurgWhat is so bad about temporary logging to web server logs? [02:37]
gac410The logs in foswiki are an operational part of the API. "eachEventSince" iterator [02:38]
vrurgDunno, my view is that an app must run for as long as it doesn't affect it's internal structure/data. Anything else must not prevent a user accessing information he would require or do his duties.
I know. But sometimes... Ok, as we cannot agree on this in a few minutes discussion then it's definitely a subject for a proposal.
Perhaps somebody would come up with a better solution for this.
[02:39]
gac410The logs are an internal structure. Extension can use the eachEventSince API to find out what's going on. [02:40]
vrurgBut in my setup, for example, I have operators taking care of network monitoring. If a remote node goes down – they have to get contact info on wiki and report it. But when logfile becomes unwritablle after the rotation then they simply wait until morning doiing nothing.
And that's a problem much worse than anything else.
Ok, I'm not distracting you anymore. Thanks!
[02:42]
gac410Well once you fix your rotation script to use the correct userid, it should never happen again, right? I hope this isn't a monthly occurrence :D
Actually the foswiki logs rotate themselves. You shouldn't need to touch them with os tools. Granted it's monthly.
I'm trying to come up with which plugins used the eachEventSince API. :( Can't come up with it.
[02:44]
vrurgRight, and there was a bug in foswiki under apache causing the logs to become root-owned after the rotation. Perhaps due to the master apache process being owned by root and thus causing foswiki to use wrong uid.
Not sure if the bug has been fixed though I remember a related task item was filled in.
[02:48]
gac410foswiki used the wrong uid? Strange. The only time I've seen the log ownership change is if somone runs any tools or bin scripts under the root user.
That can be nasty. indeed.
[02:49]
vrurgJust out of curiosity I've rechecked my configs – nah, foswiki rotates the logs on its own. [02:52]
gac410yeah. The Loggers do it by renaming the file.
It used to copy the file. Which was AWFUL. Busy systems would never rotate.
[02:53]
vrurgThough I don't remember when did the problem happen to me last time. Could be fixed by now. [02:53]
gac410The logger should never change ownership. It runs under control of CGI, no way to change ownership that I can think of. [02:54]
vrurgAnyway, a bug like this is always possible. If not this particular one then another. So, a proposal would be useful anyway. [02:55]
gac410The easiest way to trip up things is to run a script as root. bin/view tools/tick_foswik tools/mailnotify, etc. [02:55]
vrurgAs far as I remember, what I was checking the rotation code it simply creates a new file with open. Why was it receiving wrong ownership is beyond me. [02:56]
gac410y. I've never seen it happen, except as I said, someone accidentally running a foswiki script/tool as root, [02:56]
vrurgOk, it's not that much time left and I have a new bug to track down. [02:57]
gac410Installing the foswiki tasks in the root crontab vs. the web-server owers' crontab is a killer.
gac410 needs to copy the code to f.o and install it.
[02:57]
vrurg🤞🏻 [03:00]
gac410That just came through as a unknown unicode box. 01f91E [03:01]
vrurgHm, sorry. It was 'crossing fingers' emoji. [03:05]
gac410ah...
foswiki.org is now running Foswiki-2.1.4-RC2
[03:05]
vrurgCongrats and g
g'night!
[03:15]
gac410g'night [03:15]
***gac410 changes topic to: Download: NEW RC: https://foswiki.org/Download/FoswikiRelease02x01x04_RC2 Logs: http://irclogs.foswiki.org/bin/irclogger_log/foswiki - Slack: https://foswiki-slackin.herokuapp.com/ - Bugs: https://foswiki.org/Tasks - Next Release meeting Monday 29 May 2017, 1300Z
gac410 changes topic to: Download: NEW RC: https://foswiki.org/Download/FoswikiRelease02x01x04-RC2 Logs: http://irclogs.foswiki.org/bin/irclogger_log/foswiki - Slack: https://foswiki-slackin.herokuapp.com/ - Bugs: https://foswiki.org/Tasks - Next Release meeting Monday 29 May 2017, 1300Z
[03:27]
.................................. (idle for 2h48mn)
ChanServ sets mode: +o MichaelDaum [06:15]
GithubBot[distro] MichaelDaum pushed 1 new commit to Item14288: https://git.io/v9bJE
distro/Item14288 f17cef8 MichaelDaum: Item14288: working on dialogs as well as......
[06:21]
***GithubBot has left [06:21]
FoswikiBothttps://foswiki.org/Tasks/Item14288 [ Item14288: rewrite to support pluggable edit engines ] [06:21]
GithubBot[distro] MichaelDaum pushed 1 new commit to Item14288: https://git.io/v9bUv
distro/Item14288 2ac8e71 MichaelDaum: Item14288: merge 'origin/master'
[06:28]
***GithubBot has left [06:28]
............................. (idle for 2h21mn)
MichaelDaumDeep recursion on anonymous subroutine at lib/Foswiki/Engine/Legacy.pm line 48. [08:49]
GithubBot[distro] MichaelDaum pushed 1 new commit to Item14288: https://git.io/v9bGa
distro/Item14288 d0e7192 MichaelDaum: Item14288: removing magic 2px from height calculation
[08:56]
***GithubBot has left [08:56]
FoswikiBothttps://foswiki.org/Tasks/Item14288 [ Item14288: rewrite to support pluggable edit engines ] [08:56]
......................................... (idle for 3h21mn)
***ChanServ sets mode: +o Lynnwood [12:17]
.... (idle for 18mn)
ChanServ sets mode: +o gac410 [12:35]
......... (idle for 44mn)
gac410MichaelDaum: What branch is the deep recursion in? [13:19]
MichaelDaummaster
pseudo-install MetaDataPlugin
[13:24]
gac410hm strange. I'll try to recreate it. Not even sure when the Legacy engine is even used.
wow nasty ... almost forced a reboot.
Not sure what all the IO was, but my system was non-responsive for a while.
wtf is different about that extension. This is really strange.
[13:24]
MichaelDaumas soon as I add Foswiki::Plugins::JQueryPlugin,>=4.10,perl,Required. in line 2, does it install just fine [13:42]
FoswikiBothttps://trunk.foswiki.org/System/PerlDoc?module=Foswiki::Plugins::JQueryPlugin [13:42]
MichaelDaumit craps out requiring JsonRpcContrib ... and dives into an eternal loop from there [13:42]
gac410It seems to be a bug in pseudo-install. ... maybe. It's falling through and running code at
main::(./pseudo-install.pl:1412):       my $installed = run();
hm no its normal to fall through that code.
[13:43]
...... (idle for 25mn)
MichaelDaum: I have no idea. I've been single stepping through the debugger trying to find out what's going on. I'm clueless. [14:11]
definitely something with the MANIFEST / DEPENDENCIES. If I hack in a $nodeps=1; in the installFromMANIFEST, then it installs just fine.
I think in processing the dependencies it does a "Require Foswiki:: .... " something, which causes foswiki to run.
[14:16]
gac410 wonders if we have anything that doesn't use the switchboard and would need the old legacy engine. [14:26]
Well the "tools/dependencies" script does successfully use the Engine::Legacy without any looping. [14:31]
Adding +    $Foswiki::cfg{Engine} = 'Foswiki::Engine::CGI'; to the top of the pseudoInstall.pl::installFromMANIFEST() routine also fixes the issue. [14:41]
FoswikiBothttps://trunk.foswiki.org/System/PerlDoc?module=Foswiki::Engine::CGI [14:41]
gac410Or Engine::CLI [14:42]
............. (idle for 1h1mn)
***ChanServ sets mode: +o Lynnwood__ [15:43]
...... (idle for 27mn)
gac410MichaelDaum: I opened Item14399 along with my "fix" though this is just masking the issue I suspect. [16:10]
FoswikiBothttps://foswiki.org/Tasks/Item14399 [ Item14399: Deep recursion in Foswiki::Engine::Legacy line 44 ] [16:10]
MichaelDaumokay thanks. gottta run. will take a look tomorrow. see you. [16:14]
.... (idle for 16mn)
***ChanServ sets mode: +o Lynnwood
ChanServ sets mode: +o Lynnwood__
[16:30]
ChanServ sets mode: +o Lynnwood [16:46]
ChanServ sets mode: +o Lynnwood [17:00]
.... (idle for 15mn)
ChanServ sets mode: +o Lynnwood [17:15]
..... (idle for 21mn)
ChanServ sets mode: +o Lynnwood [17:36]
............................................ (idle for 3h38mn)
ChanServ sets mode: +o Lynnwood [21:14]
.................... (idle for 1h38mn)
vrurggac410: may I ask for a linguistic/terminology help? [22:52]
gac410sure [22:52]
vrurgCurrent Configure is using term 'pluggables'. But I took for a different entity (a pluggable method). Could you suggest a replacement for specs?
I'm stuck on this for the last 20 minutes. ;)
[22:53]
gac410hm Could you explain a bit more what you want to represent.
"Connector"
Insertable Extensible
[22:55]
vrurgAn entity which would autogenerate some a bunch of specs depending on the environment. Like currently there're LANGUAGES, PLUGINS pluggables.
Perhaps 'insertable' would be the best. Extensible may still be confused with 'extensions'.
[22:57]
gac410Whats wrong with just using pluggable? [22:58]
vrurgBecause it may produce another kind of confusion with 'pluggable' keyword I'm using for the new extensions. [22:58]
gac410Expandable spec? [22:59]
vrurgThese are different worlds and never cross with each other. But in a developer's mind they might mix up unpredictably. [22:59]
gac410Definable [23:00]
vrurgWhat if I focus on the fact that it will be a code called? Would change something? [23:01]
gac410the code is called to enhance/expand the "Spec" ... "Specify"
How about "Extend"
[23:02]
vrurgI think 'expandable' is the best. Because it is definitely expands an option into specs.
Ok, thanks! I buy it! ;)
[23:04]
gac410gac410 was plugging words into http://www.thesaurus.com/
aggrandize
[23:06]
vrurgI always forget about these kind of services... Still, I was stuck totally anyway. One of these days when brains are against of any kind of work, you know... ;) [23:08]
gac410;)
(And I wasn't serious about aggrandize )
[23:09]
vrurgI had to find it in dictionary in first place. A word I never seen before.
Poor developers with English worse then mine if we use this vocabulary! :D
Playing with thesaurus while IDE restarts. 'implant' would be a nice one unless so surgically tied... ;)
[23:10]

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