Codecov Report

Merging #1901 (74d478c) into master (d456913) will decrease coverage by 0.49%.
Report is 3 commits behind head on master.
The diff coverage is 52.95%.

@@             Coverage Diff              @@
##             master    #1901      +/-   ##
============================================
- Coverage     44.59%   44.10%   -0.49%     
+ Complexity     3743     3720      -23     
============================================
  Files           166      165       -1     
  Lines         12728    12728              
============================================
- Hits           5676     5614      -62     
- Misses         7052     7114      +62     

There are still some changes in my vendor tree that I made to help things pass it seems. So working on that next...

Looking at my diffs I added some toArray() methods to the following. I am guessing that just adding them to release-6.1 / master in saml2 would be the way to go...

It seems github doesn't want to allow me to fork both saml2 and saml2-legacy. I could do a PR from a branch in the main saml2 repo or remove my saml2-legacy and fork and PR from there but either of those seem a little messy

Files I added toArray to:
saml2/src/XML/mdui/DomainHint.php
saml2/src/XML/mdui/GeolocationHint.php
saml2/src/XML/mdui/IPHint.php
saml2/src/XML/shibmd/Scope.php

This is strange, because you have all the admin-permissions you could possibly want on those libraries. saml2-legacy you can forget about.. That one only has the v4 version of the library.

For saml2 v6 I've created a pull request here.. Please review and let we know if you expect to need more classes to implement toArray/fromArray.

I now have the fork https://github.com/monkeyiq/saml2 I suspect that the saml2-legacy and saml2 repos were similar enough, or maybe even forked from each other in the past that github was not wanting to let me make a fork of both of them on my account.

The main saml2 lib fork is more useful so I have deleted the saml2-legacy fork I had in order to make this new fork happen.

The new MetadataParser uses SimpleSAML\SAML2\XML\shibmd\Scope::toArray() which I had a basic version of. saml2/master has a more real version of toArray() in it. I am not sure if the procedure is to

• cherry pick things back from master to release-6.1 here
• just copy the toArray() methods in PRs
• something else.

I imagine I should be using saml2 6.1 branch because that is what the composer.json is calling for. It almost works against saml2/master.

It's not really a bugfix to add those helper-methods, so I will clone master into a new release-6.2 branch and tag it

ok so composer update seems to still have it working on the 11 tests in ./tests/src/SimpleSAML/Metadata/MetadataParserTest.php

I was looking for other example XML files with EntitiesDescriptor to add to the new MetadataParserTest.php. With a rgrep on the tree there doesn't seem to be other pickings. There are a few examples in sec 2.6 (page 23) of "Metadata for the OASIS Security Assertion Markup Language (SAML) V2.0".

I will take another look at the MetadataParser.php and it's test next.

I have also been trying to get ds:Signature creation / validation going for the MetadataParser. Sort of like validateSignature did. I have some nasty code that runs but does not verify it's own signatures correctly at the moment :/

It is also possible that the keys I am using from SignerTest might not match up in the way that I had hoped.

I was looking for other example XML files with EntitiesDescriptor to add to the new MetadataParserTest.php

There should be a few examples in the saml2-library. The interoperability-tests there download a few from the Czech and Greek edu-federations.

ok so I now have it working for signing and verifying. Sans a few nasty hacks in there and really nasty code in the current scratch zone.

But at least doing what I wanted to get to. Which is creating signatures on <md:EntitiesDescriptor>, saving to XML again, reloading, and verifying the sig.

Very nice work @monkeyiq !

So that is the test cleaned up a bit for less eye bleeding impact.

Do I understand correctly from the failing test that you need toArray/fromArray implemented for the shibmd:Scope element? If yes, then that's something I can take care of