Last modified: 2014-05-18 18:28:03 UTC
Easter eggs are a common piece of software (see: [[Easter egg (media)]]). To my knowledge, MediaWiki doesn't contain many Easter eggs currently, aside from parser irregularities and crude code comments. I suppose naming the "Bad image list" as Tim did is sort of an Easter egg (see: bug 14281#c10), but that's a bit too obscure for my tastes. A neat trick or hidden feature of the software would be nice. Apparently additional languages are a no-go due to ISO rigidness (see: r55092).
Mainly just want myself on CC for this in case anyone tries to commit some completely batshit idea for this bug.
That would be good, but we have to do more important things before.
This is silly. Suggest INVALID. If someone with commit access wants to do this, and it's harmless, then fine, but we don't need a bug for it.
r78200
Is r79712 enough? How many do we need?
(In reply to comment #5) > Is r79712 enough? How many do we need? 42.
r79712 needs something interesting
https://gerrit.wikimedia.org/r/#/c/49162/ addresses brewing issue
Reopening since the easter egg from r79712 has now been removed. It was fun, but had a few problems: * It segfaulted on one dev's server due to its unusual /e modifier recursion. * The MW interface was obfuscated, so could not be maintained. * The data encoding used a lot of fonts, and so caused performance problems in editors and browsers. * The data encoding was vulnerable to corruption by text editors. This ultimately caused it to stop working. * It used /e to demonstrate the security problems of using /e. Since /e is now deprecated for security reasons, I suppose we can congratulate ourselves for raising awareness, but it also means that going forward, /e usage is unnecessary and a liability. (And all that despite it being one of the most heavily-reviewed pieces of code in MediaWiki! Thanks to everyone who bothered to read it, and thanks for appreciating my jokes.) In my spare time, I am working on a new easter egg which will address these issues. The concept is a VM which interprets obfuscated bytecode stored in an image file. The interface between the VM and the rest of MediaWiki will be in maintainable plaintext, but the output generated for a given input will be non-obvious.
It was cute. Although I still think that hashar should have provided his own one. Removed in 4c69569db71d149feff6c4b10ea7a493425d67fd / https://gerrit.wikimedia.org/r/54319
