Ensure language redirect conditions do not match for too low quality accepted languages

This commit is contained in:
Alejandro Celaya 2024-03-07 10:03:11 +01:00
parent 0bc7412430
commit be8cf56240
5 changed files with 37 additions and 13 deletions

View file

@ -6,7 +6,7 @@ The format is based on [Keep a Changelog](https://keepachangelog.com), and this
## [Unreleased] ## [Unreleased]
### Added ### Added
* [#2041](https://github.com/shlinkio/shlink/issues/2041) Document `color` and `bgColor` params for the QR code route in the OAS docs. * *Nothing*
### Changed ### Changed
* *Nothing* * *Nothing*
@ -18,7 +18,8 @@ The format is based on [Keep a Changelog](https://keepachangelog.com), and this
* *Nothing* * *Nothing*
### Fixed ### Fixed
* *Nothing* * [#2041](https://github.com/shlinkio/shlink/issues/2041) Document missing `color` and `bgColor` params for the QR code route in the OAS docs.
* [#2043](https://github.com/shlinkio/shlink/issues/2043) Fix language redirect conditions matching too low quality accepted languages.
## [4.0.0] - 2024-03-03 ## [4.0.0] - 2024-03-03

View file

@ -8,6 +8,7 @@ use BackedEnum;
use Cake\Chronos\Chronos; use Cake\Chronos\Chronos;
use DateTimeInterface; use DateTimeInterface;
use Doctrine\ORM\Mapping\Builder\FieldBuilder; use Doctrine\ORM\Mapping\Builder\FieldBuilder;
use GuzzleHttp\Psr7\Query;
use Jaybizzle\CrawlerDetect\CrawlerDetect; use Jaybizzle\CrawlerDetect\CrawlerDetect;
use Laminas\Filter\Word\CamelCaseToSeparator; use Laminas\Filter\Word\CamelCaseToSeparator;
use Laminas\Filter\Word\CamelCaseToUnderscore; use Laminas\Filter\Word\CamelCaseToUnderscore;
@ -16,7 +17,6 @@ use PUGX\Shortid\Factory as ShortIdFactory;
use Shlinkio\Shlink\Common\Util\DateRange; use Shlinkio\Shlink\Common\Util\DateRange;
use Shlinkio\Shlink\Core\ShortUrl\Model\ShortUrlMode; use Shlinkio\Shlink\Core\ShortUrl\Model\ShortUrlMode;
use function array_filter;
use function array_keys; use function array_keys;
use function array_map; use function array_map;
use function array_pad; use function array_pad;
@ -27,6 +27,7 @@ use function implode;
use function is_array; use function is_array;
use function print_r; use function print_r;
use function Shlinkio\Shlink\Common\buildDateRange; use function Shlinkio\Shlink\Common\buildDateRange;
use function Shlinkio\Shlink\Core\ArrayUtils\map;
use function sprintf; use function sprintf;
use function str_repeat; use function str_repeat;
use function str_replace; use function str_replace;
@ -85,16 +86,30 @@ function normalizeLocale(string $locale): string
} }
/** /**
* Parse an accept-language-like pattern into a list of locales, optionally filtering out those which do not match a
* minimum quality
*
* @param non-empty-string $acceptLanguage * @param non-empty-string $acceptLanguage
* @return string[]; * @param float<0, 1> $minQuality
* @return iterable<string>;
*/ */
function acceptLanguageToLocales(string $acceptLanguage): array function acceptLanguageToLocales(string $acceptLanguage, float $minQuality = 0): iterable
{ {
$acceptLanguagesList = array_map(function (string $lang): string { /** @var array{string, float|null}[] $acceptLanguagesList */
[$lang] = explode(';', $lang); // Discard everything after the semicolon (en-US;q=0.7) $acceptLanguagesList = map(explode(',', $acceptLanguage), static function (string $lang): array {
return normalizeLocale($lang); // Split locale/language and quality (en-US;q=0.7) -> [en-US, q=0.7]
}, explode(',', $acceptLanguage)); [$lang, $qualityString] = array_pad(explode(';', $lang), length: 2, value: '');
return array_filter($acceptLanguagesList, static fn (string $lang) => $lang !== '*'); $normalizedLang = normalizeLocale($lang);
$quality = Query::parse(trim($qualityString))['q'] ?? 1;
return [$normalizedLang, (float) $quality];
});
foreach ($acceptLanguagesList as [$lang, $quality]) {
if ($lang !== '*' && $quality >= $minQuality) {
yield $lang;
}
}
} }
/** /**
@ -108,7 +123,7 @@ function acceptLanguageToLocales(string $acceptLanguage): array
*/ */
function splitLocale(string $locale): array function splitLocale(string $locale): array
{ {
return array_pad(explode('-', $locale), 2, null); return array_pad(explode('-', $locale), length: 2, value: null);
} }
function getOptionalIntFromInputFilter(InputFilter $inputFilter, string $fieldName): ?int function getOptionalIntFromInputFilter(InputFilter $inputFilter, string $fieldName): ?int

View file

@ -77,7 +77,7 @@ class RedirectCondition extends AbstractEntity implements JsonSerializable
return false; return false;
} }
$acceptedLanguages = acceptLanguageToLocales($acceptLanguage); $acceptedLanguages = acceptLanguageToLocales($acceptLanguage, minQuality: 0.9);
[$matchLanguage, $matchCountryCode] = splitLocale(normalizeLocale($this->matchValue)); [$matchLanguage, $matchCountryCode] = splitLocale(normalizeLocale($this->matchValue));
return some( return some(

View file

@ -75,9 +75,15 @@ class RedirectTest extends ApiTestCase
]; ];
yield 'rule: complex matching accept language' => [ yield 'rule: complex matching accept language' => [
[ [
RequestOptions::HEADERS => ['Accept-Language' => 'fr-FR, es;q=08, en;q=0.5, *;q=0.2'], RequestOptions::HEADERS => ['Accept-Language' => 'fr-FR, es;q=0.9, en;q=0.9, *;q=0.2'],
], ],
'https://example.com/only-english', 'https://example.com/only-english',
]; ];
yield 'rule: too low quality accept language' => [
[
RequestOptions::HEADERS => ['Accept-Language' => 'fr-FR, es;q=0.8, en;q=0.5, *;q=0.2'],
],
'https://blog.alejandrocelaya.com/2017/12/09/acmailer-7-0-the-most-important-release-in-a-long-time/',
];
} }
} }

View file

@ -35,6 +35,8 @@ class RedirectConditionTest extends TestCase
#[TestWith(['es, en,fr', 'en', true])] // multiple languages match #[TestWith(['es, en,fr', 'en', true])] // multiple languages match
#[TestWith(['es, en-US,fr', 'EN', true])] // multiple locales match #[TestWith(['es, en-US,fr', 'EN', true])] // multiple locales match
#[TestWith(['es_ES', 'es-ES', true])] // single locale match #[TestWith(['es_ES', 'es-ES', true])] // single locale match
#[TestWith(['en-US,es-ES;q=0.6', 'es-ES', false])] // too low quality
#[TestWith(['en-US,es-ES;q=0.9', 'es-ES', true])] // quality high enough
#[TestWith(['en-UK', 'en-uk', true])] // different casing match #[TestWith(['en-UK', 'en-uk', true])] // different casing match
#[TestWith(['en-UK', 'en', true])] // only lang #[TestWith(['en-UK', 'en', true])] // only lang
#[TestWith(['es-AR', 'en', false])] // different only lang #[TestWith(['es-AR', 'en', false])] // different only lang