[TYPO3-core] RFC #9553: Bug: Form validation script jsfunc.validateform.js sometimes fails in IE

[Vladimir Podkovanov]

Xavier Perseguers wrote:
> Hi,
> 
>> By the way I think form validation should be rewritten in the future 
>> from scratch. IMHO it should be server side as
>>
>> 1. in case of an error in js validation func emails just sending 
>> without checking required fields
>>
>> 2. client side checking could be easily tricked by editing js or 
>> switching js off
>>
>> 3. there is old bug 3527 re: non-english unicode labels in form; that 
>> is not resolved because old browsers have missed javascript function 
>> decodeURI() (one of reasons)
>>
>> 4. bug in current thread also shows that better do server side 
>> evaluation than rely on valid javascript execution in every browser
> 
> IMHO, client-side check is just a way to quickly check that the form 
> *seems* to be valid, it is convenient in many cases to do this because 
> it saves bandwidth and can perform some quick checks, mainly misses from 
> the user.
> 
> However, server side checks should *always* be done for the reasons you 
> gave above and because sometimes a script is accessed "legally" from 
> another script on another server without relying on the html form itself 
> but by performing a direct POST request.
> 

Yes, I think it is good way and IMHO it should be configurable from 
content element and in FORM cObj - choosing client side processing, 
server side processing or both should be used.
And if we leave client side js checking then unicode labels still should 
be handled.

Vladimir