Do all the things like ++ or -- rants, post your own rants, comment on others' rants and build your customized dev avatarSign Up
From the creators of devRant, Pipeless lets you power real-time personalized recommendations and activity feeds using a simple APILearn More
nitnip1533186dUgly code can make any framework look awful. I like laravel, but I really dislike the repository pattern because people only seem to use it to be fancy or because they read a medium article that told them to, and then it inevitably falls apart.
Same with Service classes and DI. I only use the former sparingly. Overall, I find laravel way easier to debug than something like cake.
My biggest problem with symfony is how prevalent yaml is.
gatoMalicioso140185d@nitnip i do not think that laravel encourages repositories. the opposite. ORM make developers spread the queries everywhere. even queries with great logic. Model the same. God Objects.
But what i try to say is that even if i try to write good code, i have to open the documentation like a configuration. "how was the name of dates property on model, how was the casting, how was the other thing that takes all parameters at once and create the model".
ok you start to remember the basic. but...
IntrusionCM12270185dIt's dependent on the project and it's size.
Many patterns only make sense with a reasonable project size and _badumts_ human resources.
DI is a necessity in my opinion, repository pattern, too - in a gist a domain driven approach.
What I dislike mostly are annotations and all the kind of magic methods that some framework are very fond of.
Especially gimmicks like "Autotransformation / serialization" which some devs love and sell as an great invention for automation, e.g. that an incoming request gets transformed with no additional call to an DTO like object.
Or the myriad of magic methods in PHP (__set, __get, ...) and their abusal.
Many devs dislike boilerplate code - till they had the joyful job to debug the automatic code.
Be it transformation errors (e.g. special values, float errors, datetime formats, milli -/ microsend timestamps, ...) or just a missing http header like Accept / Content-Encoding and suddenly and encryptic exception with a traceback of 10 k lines....which tells you what and why. But not the simple reason how - unless you sifted for hours through the traceback.
I think boiler plate code which is self explanatory is better then some magic deep nested vodoo fuckery that is slow, unprecise and in an error case an total enigma.