Doing regex on phone numbers is not a great idea. This is a breaking
change, but a good one. Phone validation is now much stricter, and
allows choosing the country.
The use case for negating a keyset is very confusing, and can
lead to validators that don't do what they expect.
This commit introduces NonNegatable rules, which will throw
a Component exception if you try to wrap them in `Not`.
This change was necessary to ensure proper message reporting
when extra keys exist on the keyset.
This fixes#1349
The `filter_var` function is more of a sanitizer, but we as
a validation library do not care for that use case.
We should treat its sanitizings as a signal for checking if
the type after sanitization matches the option provided.
This fixes#1387
Since "sokil/php-isocodes" is a dependency of our repository already, it
makes sense to use it as a source of currency codes instead of keeping a
list of currencies ourselves.
By using that library, we can also validate currency codes using
different sets.
Signed-off-by: Henrique Moody <henriquemoody@gmail.com>
Inside the "data/" directory, we have files with lists of subdivisions
that need to be updated. We have to update them manually, or we automate
that task with a script and GitHub actions.
The two options are very time consuming and also not ideal. We don't
want to deal with that problem and, thinking that the user of this
library may want to show the data that we validate, we should create a
whole library to make it more usable.
The "sokil/php-isocodes" is a simple library that, even supports
translations. It's frequently updated and has gone to major performance
updates.
I am not fond of the idea of requiring an external library to install
Validation, as I have seen that gone wrong before [1]. Ideally, that
would be an optional dependency for people who would like to use those
rules, but to make that happen, we need to release a MAJOR version.
[1]: d072b4de6a
Signed-off-by: Henrique Moody <henriquemoody@gmail.com>
The PSR-7 has two interfaces that allow us to validate them as files.
This commit will allow some rules to validate those interfaces.
Co-authored-by: Henrique Moody <henriquemoody@gmail.com>
Signed-off-by: Henrique Moody <henriquemoody@gmail.com>
Most Validation errors are sent to Users/Visitors or Clients and as such
might not need to know it was a Key their inputs are being validated
upon.
Co-authored-by: Henrique Moody <henriquemoody@gmail.com>
Signed-off-by: Henrique Moody <henriquemoody@gmail.com>
After the refactoring on the Factory class [1], to throw exceptions of a
specific rule, it is necessary to add the exception namespace of that
rule. That change makes sense when someone wants to create rules from
the Validator class, but when using rules as classes, it's not as handy.
This commit will auto-resolve exception based on the rule namespace,
just as it used to be.
[1]: 1f217dda66
Co-authored-by: Casey McLaughlin <caseyamcl@gmail.com>
Signed-off-by: Henrique Moody <henriquemoody@gmail.com>
Although BCMath is necessary to execute "Ip" and "Iban" rules, we not
required that in the "composer.json" file. That's because if someone
wants to use the library, but doesn't want to use those rules, they
would not need to install that extension.
However, when executing the tests, they will break. This commit will
verify whether the extension exists to test "Ip" and "Iban" rules.
Signed-off-by: Henrique Moody <henriquemoody@gmail.com>
We already supported PHP 8.0 as our constrains in the "composer.json"
file was ">=7.3", but we were not testing it before.
Because of that, I found a bug on "EndsWith" which is fixed now.
Signed-off-by: Henrique Moody <henriquemoody@gmail.com>
Some amazing features had to be ignored because it conflicts with out
coring standards. I hope to soon fix them so we can use PHPStan to its
fullest potential.
Signed-off-by: Henrique Moody <henriquemoody@gmail.com>
All the rules that extend the "AbstractRelated" class overwrite the name
of the inner-rule even when it already has a name. This leads to some
unexpected behavior, as messages won't show the name of the inner rule
but instead the name of the parent rule.
Co-authored-by: Henrique Moody <henriquemoody@gmail.com>
Even though using "filter_var()" to validate domains may is error-prone,
Validation should fully support "filter_var()" as long as the
"FilterVar" exists.
Co-authored-by: Henrique Moody <henriquemoody@gmail.com>
There is some confusion about integer literals (as we type them into
source code) and integer values (the actual value they represent).
When casting the integer 08 (without quotes), PHP triggers an error as
integers starting with 0 should have base 8. However, when casting the
string '08' as an integer PHP returns the integer 8.
This commit will change the behavior of the "IntVal" rule, allowing it
to accept any integer type and any representation of an integer as a
string.
Reviewed-by: Emmerson Siqueira <emmersonsiqueira@gmail.com>
Reviewed-by: Wesley Victhor Mendes Santiago <w.v.mendes.s@gmail.com>
Co-authored-by: Henrique Moody <henriquemoody@gmail.com>
I can't explain why, but depending on the environment, some tests of the
"Call" rule was failing.
This commit will add a test for the "validate()" method, and also use a
more clean way to trigger and handle a PHP error.
Co-authored-by: Casey McLaughlin <caseyamcl@gmail.com>
Signed-off-by: Henrique Moody <henriquemoody@gmail.com>
According to Wikipedia, also known as the most reliable source of
information on the whole Internet, Serbian postal codes consist of five
digits [1].
Even though it mentions that a six-digit postal code format has been in
place since 1 January 2005 [2], I have not found any Serbian postal code
which consists of six digits.
Maybe someone would prove us wrong, but we will keep the validation as
six-digits from now one.
[1]: https://en.wikipedia.org/wiki/Postal_codes_in_Serbia
[2]: https://web.archive.org/web/20090917083620/http://www.posta.rs/postanskamreza/postanskamreza.asp
Co-authored-by: Henrique Moody <henriquemoody@gmail.com>
This commit will update the list of postal codes using the command
below:
curl -L http://download.geonames.org/export/dump/countryInfo.txt |
sed 's,\t,\;,g' |
sort --unique |
cut --delimiter ';' --field 1,15 |
sed --regexp-extended "/^#/d; /^[A-Z]{2}\;$/d; s,([A-Z]{2})\;(.+),'\1' => '/\2/'\,,g"
The changes that broke existing tests were reverted.
Signed-off-by: Henrique Moody <henriquemoody@gmail.com>
According to Wikipedia[1]:
> Postal codes in Armenia consist of four digits. Until April 1, 2006,
> they consisted of six digits.
[1]: https://en.wikipedia.org/wiki/Postal_codes_in_Armenia
Co-authored-by: Henrique Moody <henriquemoody@gmail.com>
According to Wikipedia[1]:
> A postal code typically consists of an island code, a hyphen
> separator, and a section code.
>
> There are only three island codes: KY1 for Grand Cayman, KY2 for
> Cayman Brac, and KY3 for Little Cayman.
[1]: https://en.wikipedia.org/wiki/Postal_codes_in_the_Cayman_Islands
Co-authored-by: Henrique Moody <henriquemoody@gmail.com>
According to Wikipedia [1]:
> Until February 2013, Israel postal codes were 5-digit. Officially the
> 5-digit codes remained valid until 31 January 2013, though they
> continue to be widely used.
This commit changes the "PostalCode" rule to allow either 5 or 7 digits.
[1]: https://en.wikipedia.org/wiki/List_of_postal_codes
Co-authored-by: Henrique Moody <henriquemoody@gmail.com>
The intention of the "assert()" method is to show all the errors that a
given input may have. The implementation of the "assert()" method in the
"Each" rule, on the other hand, only reports the first error of each
element of the input.
This commit makes "Each" show all the validation failures of each
element of the input. Also, the implementation of
"AbstractRule::check()" is simply a proxy for the "assert()" method,
and since the "Each" rule extends that class, this commit creates a
custom implementation of the "check()" method.
Co-authored-by: Henrique Moody <henriquemoody@gmail.com>
The "Call" validator, the "assert()" and "check()" methods, define a
custom error handler so it can handle the validation in case of a PHP
error. However, it does not restor the previous error handler when the
validation fails.
Co-authored-by: Henrique Moody <henriquemoody@gmail.com>