Data validation issues

So here's the story. Our CTO once thought that data validation libraries he've ever used have the following issues:

#1. Most validators check only the data having the described checking rules while I believe the validator must cut all the data for which the validation rules aren’t defined.

#2. Procedural description of the validation rules. Basically, I want to set a data scheme.

#3. Description of the validation rules in code.

#4. The validation stops at the first field with an error and doesn't highlight all the faulty / necessary fields at once.

#5. Non-standardized error messages.

#6. Numeric error codes, not intuitive.

#7. No possibility to check the hierarchical data structures.

#8. A limited set of rules.

#9. Too wide a range of responsibilities.

#10. Inability to conduct an additional data processing.

How do you handle these issues? Our CTO resulted in creating a multi-language validation library LIVR. If you wish, check out it here: blog.webbylab.com/language-independent-val…