| Version 28 (modified by , 17 years ago) (diff) |
|---|
Issues
- Design issue: In my design, the Certificate struct has an ptr to a GnuTLS certificate structure to hold its internal data. This has the consequence that the SSL connection system and X.509 handling must both be provided by GnuTLS; if, say, SILC provides the X.509 handling, and GnuTLS the SSL connections, everything grinds to a gruesome segfaulty halt.
Should I address this? The only workaround I can think of is to keep a pointer to the PEM-formatted certificate in the Certificate struct instead, but that will require any consumers of the Certificate to have PEM encoding/decoding and constantly decode it from PEM for every use. (June 19)
- How to ensure that VerificationRequests? get properly trashed in unusual conditions (such as canceling an SSL connection)? (June 26)
- Jabber.org presents a cert with a CN of "*.jabber.org". How to check this? (26 June)
- Update: This is *almost* valid. "foo.jabber.org" is valid with this certificate, but not "jabber.org" itself! (8 July)
Resolved Issues
- It looks like PKCS12 (the certificate import/export format) is supported by both libNSS and GnuTLS.
- Jabber "connect with fd" problem solved by porting a change from soc.2007.xmpp (29 May, June 26)
- Why doesn't a "not accepted" on a certificate cause the connection to fail? I suspect this is caused by auto-reconnect, so it might be necessary to disable the account involved! (26 June)
- This could be related to Jabber not handling the BAD_CERTIFICATE error code?
- I think this is resolved...at least, I can't reproduce it any more.. (11 July)
TODO
- General paranoia
- Look at how the SILC prpl does its key management, especially the organization of the API used to check certs and interact with the user to verify them.
- Add some way of passing useful error messages back up out of the SSL interface (23 May)
- Fix purple_ssl_init in sslconn.c; it doesn't do anything (23 May)
- Talking to nosnilmot suggests that this ought to just be removed outright (24 May)
- Figure out libNSS everything. (25 May)
- Why am I getting single-byte serial numbers from servers? (25 May)
- Stall the timeouts on the TCP connection while waiting for user input on SSL? (29 May)
- Worry about ensuring that plugins don't kill in-use ciphers/certschemes when unloaded? (29 May)
- Adium uses neither GnuTLS nor NSS. Look into extending the OpenSSL plugin (26 June)
- Guarantee that purple_certificate_verify will not fire its callback until the next return to mainloop. sadrul says to do this with g_timeout_add (June 28)
- Change the things that purple_certificate_verify does such that it is easy for one verification request to make another. This will involve making a "copy" operator for certificates. (7 July)
- Disapprove 42fb4d80947518f87ddcaf43bf2675d8496698de (7 July)
- Reference-count Certificate instances instead of making full copies? (8 July)
- Implemented in f9e47ecff65cbcb1e87b8fedd271ec5b02123fcf, but may need testing (13 July)
- Refactoring. The certificate interface feels too complicated. (11 July)
- HIG compliance (17 July)
- How should a PidginCertificateManager? signal that it *cannot* be displayed? (19 July)
- Degrade gracefully when the CertMgr subsystem lacks important bits. (3 Aug)
Tasks done
- Figure out how to get key fingerprints out of GnuTLS (25 May, 25 May))
- GnuTLS and NSS should probably be configured to use g_malloc and g_free for paranoia's sake (1 June, 26 June)
- e408398aae606a1aafb857413db4fa9736e374ae should be disapproved. It is hacky and stupid (26 June, 7 July)
- Why won't gtkcertmgr.h get properly processed by "make docs"? (13 July, 13 July)
- sadrul is awesome. Typos are not.
- Work out how to use Glib functions for time checking on certificates. (29 May, 13 July)
- POSIX for the win.
- activation/expiration should be made one function in the Scheme, as that's how NSS does it. (13 July, 3 Aug)
