Allow easier extension of the timezone guessing

[heiglandreas]

This will ease customization of timezone-guessing as it is now gets easier to extend that process with own implementations (as long as they implement the appropriate interface)

This is espechially necessary when wanting to actually guess a timezone via the rules defined in the VTIMEZONE-entry (which is currently not done)

This change does not introduce any new logic. It just rearranges the already existing code in a different way to allow creating own TimezoneGuessers that can extend the functionality like this:

class BuesingenGuesser implements TimezoneGuesser
{
    public function guess(VTimeZone $vtimezone, bool $throwWhenUnsure = false): ?DateTimeZone
    {
        return new DateTimeZone('Europe/Busingen');
    }
}
TimezoneUtil::addTimeZoneGuesser(new BuesingenGuesser());

Now every time a timezone is not resolvable the timezone will be resolved to Europe/Busingen instead of UTC (or whatever else is set via date_default_timezone_set())

[staabm]

Hey Andreas,

Really like this direction. Do we have/need test coverage for this code?

[codecov]

Codecov Report

Merging #548 (ff902c6) into master (a7460c5) will decrease coverage by 0.35%.
The diff coverage is 93.00%.

@@             Coverage Diff              @@
##             master     #548      +/-   ##
============================================
- Coverage     98.69%   98.34%   -0.36%     
- Complexity     1818     1848      +30     
============================================
  Files            66       71       +5     
  Lines          4454     4521      +67     
============================================
+ Hits           4396     4446      +50     
- Misses           58       75      +17     

Impacted Files	Coverage Δ
lib/TimeZoneUtil.php	69.64% <86.66%> (-28.09%)	⬇️
lib/TimezoneGuesser/FindFromTimezoneIdentifier.php	92.85% <92.85%> (ø)
lib/TimezoneGuesser/FindFromOffset.php	100.00% <100.00%> (ø)
lib/TimezoneGuesser/FindFromTimezoneMap.php	100.00% <100.00%> (ø)
lib/TimezoneGuesser/GuessFromLicEntry.php	100.00% <100.00%> (ø)
lib/TimezoneGuesser/GuessFromMsTzId.php	100.00% <100.00%> (ø)

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update a7460c5...ff902c6. Read the comment docs.

[heiglandreas]

Should be good by now..

[staabm]

just leaving a few comments. but lets wait what the other devs think.

[staabm]

I am not sure we can drop this field, as other code might rely on it beeing defined.

there is more public api changed with this PR in a non-BC manner

[heiglandreas]

Right!. I'll readd those other public items to stay backwards compatible.

Main question though is whether I shall do it in a way that allows them to be easily removed at a later stage (meaning to duplicate some code) or to have as little duplicate code as possible.

Personally I'd opt for the former version but am happy to modify it to the later one.

[phil-davis]

@heiglandreas the code style is failing. It is easily fixed by:

composer cs-fixer

and then commit and push the changes it makes.

[phil-davis]

https://github.com/sabre-io/vobject/runs/4101303777?check_suite_focus=true
Class Sabre\VObject\TimeZoneUtil referenced with incorrect case

The code analysis is fussy. Please fix that up.

Note: I released https://github.com/sabre-io/vobject/releases/tag/4.3.6 an hour ago. But when we get this PR sorted out, it will be easy to make a 4.3.7 release.

[phil-davis]

Just a couple of tiny comments.

[phil-davis]

@staabm @evert @DeepDiver1975 please review and comment.
This would be at least a minor version bump (new feature - way of processing timezones). Please think if there is any remaining BC break that would require a major version.

[staabm]

lgtm

[phil-davis]

I squashed and rebased to get up-to-date clean CI.

[phil-davis]

LGTM. There is some discussion in PR #553 about tests to cover the functions addTimezoneGuesser and addTimezoneFinder but that can be left until later.