1. 03 Jan, 2019 1 commit
  2. 18 Dec, 2018 1 commit
  3. 14 Dec, 2018 3 commits
  4. 13 Dec, 2018 1 commit
  5. 02 Nov, 2018 2 commits
  6. 19 Oct, 2018 3 commits
  7. 26 Sep, 2018 4 commits
    • Vincent Pelletier's avatar
      all: More python3 adaptations. · 8ce08bf9
      Vincent Pelletier authored
      What was not picked up by 2to3.
    • Vincent Pelletier's avatar
      wsgi: Become web-friendly · 719959e0
      Vincent Pelletier authored
      Self-describe site structure in application/hal+json format.
      Add Cross-Origin Resource Sharing support: pre-flight request support,
      same-origin-only origin access control minimal html page. Access control
      decision is stored client-side in a signed & time-limited cookie
      supporting multiple concurrent origins. Origins may be pre-allowed (ex:
      when caucase GUI is served from a trusted server).
    • Vincent Pelletier's avatar
      http: Constrain the certificates caucased https CA may sign. · 7ff81404
      Vincent Pelletier authored
      This makes it safer to trust this CA certificate in general-purpose https
      clients, like web browsers, as it prevents such trusted CA certificate
      from issuing rogue certificates.
      Bump pyOpenSSL to latest version (and, as a consequence of pyOpenSSL
      18.0.0 itself requiring cryptography 2.1.1, bump it as well) as it seems to
      fix a bug related to validating NameConstraints - and anyway fixes
      worrying use-after-free errors.
    • Vincent Pelletier's avatar
      all: Make caucased https certificate independent from CAS. · 17325dc0
      Vincent Pelletier authored
      This is a step in the direction of being browser-friendly: if caucased
      https certificate is issued by CAS CA, then for a browser to trust that
      certificate it would have to trust all certificates emitted by CAS CA
      certificate. This would be very dangerous, as CAS CA does not constrain
      the certificates it may sign, so it exposes users of that caucased to
      rogue certificates.
      Alone, this step is insufficient, as the new internal "http_cas" does not
      constrain certificates yet. This will happen in a separate commit, to
      ease review and regression testing.
      As a consequence of this step, by default client will not check server
      certificate in https. This is consistent with how trust is bootstrapped
      with plain http: maybe client is accessing an unexpected/malicious
      caucased, but in such case issued certificates will be worthless to a
      party which could access the correct caucased. Also, the client
      certificate presented to caucased does not allow that caucased to fake
      being that user, so there is no privilege escalation possible for
  8. 20 Sep, 2018 2 commits
  9. 15 Jul, 2018 2 commits
  10. 14 Jul, 2018 4 commits
  11. 12 Jul, 2018 3 commits
    • Vincent Pelletier's avatar
      Bump copyright. · 94146872
      Vincent Pelletier authored
    • Vincent Pelletier's avatar
      http: Do not start listening on https port before wrapping socket. · 975d1c0b
      Vincent Pelletier authored
      Otherwise, this port will fail https handshake if clients connects too
    • Vincent Pelletier's avatar
      http: Separate bind from netloc. · 50a5d1d3
      Vincent Pelletier authored
      netloc is the public access point to a caucase instance.
      bind is the private access point to a caucase instance, which may be
      different (ex: NAT). Allow overriding netloc address with --bind.
      As a consequence, add support for multiple binds: a netloc may resolve to
      multiple addresses (ex: one IPv4, one global IPv6 and one Unique Local
      As a further consequence, systematically disable automatic IPv4 binding
      when binding to an IPv6 address.
      Also, allow overriding netloc port with --base-port. The same port pair
      will be used on all bound hosts.
      Share SSL context between multiple https sockets.
      To increase binding visibility, print bindings, and print when exiting.
  12. 04 Nov, 2017 2 commits
  13. 03 Nov, 2017 4 commits
  14. 31 Oct, 2017 1 commit
  15. 25 Oct, 2017 1 commit
  16. 24 Oct, 2017 1 commit
  17. 19 Oct, 2017 2 commits
  18. 18 Oct, 2017 1 commit
  19. 14 Sep, 2017 1 commit
  20. 23 Aug, 2017 1 commit
    • Vincent Pelletier's avatar
      all: Major rework. · ecd07d22
      Vincent Pelletier authored
      - Re-evaluate feature set and REST API.
      - switch duration units to days, which are more meaningful than sticking to
        ISO units in this context.
      - Implement the "cau" half of "caucase".
        As a consequence flask password authentication mechanism is not needed
        anymore. As HTML UI is not required internally to caucase, and as
        sqlalchemy is not used to its full extend, get rid of these
        dependencies altogether.
      - Implement REST HTTP/HTTPS stand-alone server as a layer above WSGI
        application, and integrate HTTPS certificate issuance and renewal
        mechanism to simplify deployment: no middleware needed, so from
        gunicorn dependency.
      - Use standard python modules for http client needs.
      - Re-evaluate data retention options:
        - unsigned CSRs are kept forever
        - CRTs are stored in CSR table, and a 24 hour expiration is set
        - CA CRTs: (unchanged, expire when past validity period)
        - CRLs: (unchanged, expire when past validity period)
      - Redispatch housekeeping tasks:
        - CA renewal happens when caucase is used and renewal is needed
        - CRL is flushed when re-generated
        - CSR table (containing CRTs) is cleaned when a new CSR is received
        removing completely the need for these special periodic tasks.
      - Storage parameters are not stored persistently anymore, instead their
        effect (time offsets) is applied before storing (to protect against
        transient retention period reconfiguration from wiping data).
      - Rework storage schema.
      - Implement certificate extension propagation & filtering.
      - Implement "Certificate was auto-signed" extension.
      - More docstrings.
      - Use a CSR as a subject & extensions template instead of only allowing
        to override the subject. Useful when renewing a certificate and when
        authenticated client wants to force (ex) a CommonName in the subject.
      - Reorganise cli executable arguments to have more possible actions.
        Especially, make CA renewal systematic on command start (helps
        validating caucase URL).
      - Increase the amount of sanity checks against user-provided data (ex:
        do not upload a private key which would be in the same file as the CRT
        to renew).
      - Extend package classifiers.
      - Get rid of revocation reason, as it seems unlikely to be filled, and
        even less likely to be read later.
      - (almost) stop using pyOpenSSL. Use cryptography module instead.
        cryptography has many more features than pyOpenSSL (except for certificate
        validation, sadly), so use it. It completely removes the need to poke
        at ASN.1 ourselves, which significantly simplifies utils module, and
        certificate signature. Code is a bit more verbose when signing, but much
        simpler than before.
      - add the possibility to revoke by certificate serial
      - update gitignore
      - include coverage configuration
      - include pylint configuration
      - integrate several secondary command:
        - caucase-probe to quickly check server presence and basic
          functionality, so automated deployments can easily auto-check
        - caucase-monitor to automate key initial request and renewal
        - caucase-rerequest to allow full flexibility over certificate request
          content without ever transfering private keys
      - add a secure backup generation mechanism
      - add a README describing the design