FOSTwitterBundle icon indicating copy to clipboard operation
FOSTwitterBundle copied to clipboard

consider switching twitter lib

Open lsmith77 opened this issue 14 years ago • 12 comments
trafficstars

https://github.com/jwage/Twitter

cc @kertz, @jwage

lsmith77 avatar Nov 03 '11 12:11 lsmith77

That lib seems to be using pecl extension for oauth (most of the times it won't be installed). This is not good.

kertz/twitteroauth uses it's own implementation, but that one is not written in the spirit of symfony.

There is also https://github.com/kohana/oauth which is as far as I know Oauth 2.0 client implementation but I did not try it. And Zend has it: https://github.com/zendframework/zf2/tree/master/library/Zend/OAuth (not sure if this is for 2.0 protocol already)

olegstepura avatar Mar 11 '12 18:03 olegstepura

Zend also has some Twitter Service API - why Zend is so much ignored by Symfony community?

olegstepura avatar Mar 11 '12 18:03 olegstepura

The auth side of that library for sure needs some work. I wrote that one weekend when PHP 5.3 was coming out to play with namespaces and the latest version of PHP. I've never personally used it in any project. I think the core API and organization is solid though. It just needs a little polishing love.

  • Jon

On Thu, Nov 3, 2011 at 7:14 AM, Lukas Kahwe Smith < [email protected]>wrote:

https://github.com/jwage/Twitter

cc @kertz, @jwage

Reply to this email directly or view it on GitHub: https://github.com/FriendsOfSymfony/FOSTwitterBundle/issues/13

Connect with me on http://twitter.com/jwage http://twitter.com/jwage and http://about.me/jwage to keep in touch.

Join OpenSky today: http://osky.co/l6a75g

jwage avatar Mar 11 '12 18:03 jwage

@jwage Did you take a look at Zend's implementation before you started your own?

olegstepura avatar Mar 11 '12 18:03 olegstepura

@olegstepura I can't speak for everyone else but in my opinion the quality of the Zend Framework is generally lower than the Symfony standards. I also personally dislike Zend and the way they organize themselves within the "open source" community. And I am still bitter over PEAR. That is why I ignore Zend :)

jwage avatar Mar 11 '12 18:03 jwage

@jwage Well, the quality of some components is good enough. And compared to all other implementations (mostly taking kertz/twitteroauth) OAuth may appear to be the best of all. Not saying it is. But it may be so.

olegstepura avatar Mar 11 '12 18:03 olegstepura

@olegstepura The issue is that the ZF2 code is pretty instable currently as they are refactoring components one by one (and for instance, part of the Zend\Service code was completely broken after the refactoring of the Http Client and fixed only at least a month later). And the ZF1 code is not an option as it forces to put the ZF code in the include path as only part of the codebase was relying on autoloading.

stof avatar Mar 11 '12 19:03 stof

And at the time(2 years ago). I wanted to work on a Twitter library built from the ground up to play with PHP 5.3. It wasn't for use with anything so for my objective it wasn't about using another library or something that works. This was purely a coding practice project for me.

jwage avatar Mar 11 '12 19:03 jwage

Ok, cool. Now it is now. Is Zend\Service suitable for using in FOSTwitterBundle now? I agree that Zend is monolitic library in how it's distributed. That is why KnP labs have decoupled it into composnents. And that decouple makes even more sense after introducing composer. But knp's initiative is not updated frequently.

olegstepura avatar Mar 11 '12 19:03 olegstepura

@stof Can you answer please? Is Zend\Service\Twitter suitable for using in FOSTwitterBundle now?

olegstepura avatar Mar 13 '12 15:03 olegstepura

dunno. I haven't looked at ZF2 recently and I don't use the bundle anyway

stof avatar Mar 13 '12 18:03 stof

@olegstepura i think in general the criteria for any dependency is that it must be reasonably standalone but focused on its task. in the current context this means that if we are to go with Zend's Twitter lib, then it makes sense that it doesn't cover the Http communication, however it should be reasonably easy to implement a different Http lib if we or one of our users decides it makes sense.

the main stumbling block is of course that someone actually implements this. right now we are not totally happy with the current situation as we have to rely on a fork of the upstream lib and in general there seems to be room for improvement.

however we also do not have someone really working on this bundle, we have a hard enough time to keep it maintained as is.

lsmith77 avatar Mar 13 '12 18:03 lsmith77