[v7r0] M2Crypto: fix isVOMS#4490
Conversation
|
I'm not convinced this is working, from #4481: |
|
This is the |
|
I did give a quick glance to the test though. Although far from trivial to guess what it is supposed to do, I am under the impression that it is supposed to fail, like if there was no voms extension. But given where the exception is it means we already did decode part of the extension. So maybe the subject is just very weird.. I don't know, I let the experts check |
|
I ll add a more detailed message in the exception tomorrow |
|
ProxyDB test contain a simple debug, that print every step of loops of a tests cases. Also this test was fixed in #4469 : added debug about situation where the test method that must create VOMS proxy didn't do it, for more clear understanding what's going on. Test check that result of If I understand correctly, when |
|
The fact that we raise an exception is, in DIRAC, indeed an issue, while it should be returning |
|
So conclusion here; the error reported by @chrisburr has nothing to do with this fix, so I think it can go as is. The error reported will be looked at by @TaykYoku in another PR |
|
In my test proxy a subject is Also, during the tests, I got Maybe need to add all Distinguished Names? |
|
That's weird but easily testable. |
|
Apparently the underscore is not the issue as such. I generated a certificate with That's how I read it And the output If you send me by email the proxy that fails, I can try to run the same piece of code on it and debug |
|
but again, this conversation has nothing to do here :-) |
|
Merging (needed for other PRs) |
I finally understood why it was not showing up in the unit tests. It's because
opensslkeeps some memory, and since the pyGSI tests were running before theM2Cryptoone, the namevomsExtensionswas known. I changed the order in which we test to make sure M2Crypto starts from scratch.Also, the test
isVOMSis now done by retrieving the extension by OID.BEGINRELEASENOTES
*Core
FIX: X509Chain.isVOMS returns correct value
ENDRELEASENOTES