Infer character encoding for ifAlias and sysLocation (#13248)

* Infer character encoding for ifAlias and sysLocation
Tries to convert character encoding for non-UTF-8 encoded strings.
This will only work for snmp strings that are type-hinted, not ones forced to ASCII with -Oa or similar
Only works for your default charset or Windows-1251 or LATIN1.
You can set your character encoding in .env with CHARSET

Please save us all and just use UTF-8

* style fixes

* less Yoda

* ensure return type

* fall back to passed string

* don't convert strings with any unprintable characters (such as line return)

* Update LibreNMS/Util/StringHelpers.php

Co-authored-by: Jellyfrog <Jellyfrog@users.noreply.github.com>

* Fix CP850

* fix space

Co-authored-by: Jellyfrog <Jellyfrog@users.noreply.github.com>
This commit is contained in:
Tony Murray
2021-09-21 08:47:44 -05:00
committed by GitHub
parent 11dfbc02b7
commit 8a883140cb
6 changed files with 92 additions and 1 deletions

View File

@@ -93,4 +93,36 @@ class StringHelpers
{
return ucwords(implode(' ', preg_split('/(?=[A-Z])/', $string)));
}
/**
* Sometimes devices store strings as non-unicode strings and return them directly.
* NetSnmp parses those as UTF-8, try to convert the string if it contains non-printable ascii characters.
*
* @param string|null $string
* @return string
*/
public static function inferEncoding(?string $string): ?string
{
if (empty($string) || preg_match('//u', $string) || ! function_exists('iconv')) {
return $string;
}
$charset = config('app.charset');
if (($converted = @iconv($charset, 'UTF-8', $string)) !== false) {
return (string) $converted;
}
if ($charset !== 'Windows-1252' && ($converted = @iconv('Windows-1252', 'UTF-8', $string)) !== false) {
return (string) $converted;
}
if ($charset !== 'CP850' && ($converted = @iconv('CP850', 'UTF-8', $string)) !== false) {
return (string) $converted;
}
\Log::debug('Failed to convert string: ' . $string);
return $string;
}
}