Development

IRCLogs20101216 (diff)

You must first sign up to be able to contribute.

Changes between Version 1 and Version 2 of IRCLogs20101216

Show
Ignore:
Author:
lsmith (IP: 78.53.15.239)
Timestamp:
12/21/10 09:48:46 (7 years ago)
Comment:

--

Legend:

Unmodified
Added
Removed
Modified
  • IRCLogs20101216

    v1 v2  
    33http://doodle.com/queka8t4t7e4v887 
    44 
     5== Progress on the security firewall == 
     6Johannes said "basically, the serialization is now almost finished, there are still a couple of issues to iron out but most of it has been merged, remember-me is still blocked by [http://trac.symfony-project.org/ticket/9359|one of these issues], ACL could be reviewed though, it's the only thing that is truly independent". The documentation is somewhat out of date, so https://github.com/schmittjoh/DoctrineUserBundle/tree/symfony2Security is probably a good place to get an idea of how things work now. Henrik once again brought up how confusing the method name "loadUserByUsername" is. There was universal agreement that its an unfortunate choice and that someone else like handle/identifier/principal might be better than "username". 
     7 
    58== twig tags/filters/default template parameters == 
    6  
    7 == Progress on the security firewall == 
     9Fabien has stated on the list that he wants to get rid of access to _view for twig templates. Opensky, Excersise.com and Liip all stated that currently they heavily rely on this. Jordi explained https://github.com/fabpot/symfony/pull/29: "allows any service that is a templating helper to easily support twig tags". The idea is to reduce the amount of code that has to be written to support both PHP and twig templates. There was also brief discussion about the idea to inject some variables into templates to further reduce the need for specialized tags for twig and to be more flexible (rather than ifrole which for example cannot be negated). 
    810 
    911== RFC: Merging field groups into a form == 
     12Jeremy explained that currently every "field is a 1:1 correspondence with an object property and as you nest fieldgroups in a form, those fieldgroup keys must correspond to objects within your main object". Bernhard's proposed change enables to "create fieldgroups that don't require sub-objects in your main, bound object" but "the complexity is how this will relate to binding form data to the object". For example for a date range: "mainObject.dateRange.startDate" vs. "mainObject.startDate" vs. ..? This question is especially relevant also for constraint and validation error assignment. 
     13 
     14== Extension helpers == 
     15Lukas explained that the goal if these methods is to be available optionally to reduce the amount of boilerplate code in more complex Bundles (f.e. [https://github.com/knplabs/DoctrineUserBundle/blob/master/DependencyInjection/DoctrineUserExtension.php DoctrineUserBundle] and to enable these two be more consistent in the configuration structure. Of course just like DUB extension authors could just put these methods into their code. 
    1016 
    1117== flash api and Message class == 
    12  
    13 == Extension helpers == 
    14  
     18Lukas mentioned the progress that was made since the issue was last discussed. There is a phttps://github.com/fabpot/symfony/pull/198 pull request] still open to also add the ability to remove flash messages. The bigger open question is that contrary to the previous plans there is still [http://trac.symfony-project.org/ticket/9346 no Message class]. One of the key questions is of course i18n. Jordi said that the Message class should do the translation automatically if __toString() is called. So the question was when to translate the string. This will require injecting the translation layer at some point. It was also noted that if any translation is done at runtime then due to the fact that flash messages are serialized this would require that the injected service needs to be able to wakeup. 
    1519 
    1620= IRC logs =