Permalink
Commits on Jun 27, 2018
Commits on Jan 9, 2018
  1. Merge branch 'APNIC-net-apnic-deprecated-tas'

    sraustein committed Jan 9, 2018
    Pull request from APNIC deleting obsolete APNIC TALs.
Commits on Oct 20, 2017
  1. try to get the sucker to push

    randyqx committed Oct 20, 2017
  2. try

    randyqx committed Oct 20, 2017
Commits on Oct 17, 2017
  1. s/asc/gpg/

    randyqx committed Oct 17, 2017
  2. god only knows

    randyqx committed Oct 17, 2017
  3. god only knows

    randyqx committed Oct 17, 2017
Commits on May 17, 2017
Commits on May 15, 2017
Commits on Apr 27, 2017
Commits on Apr 26, 2017
  1. emphasize handle is identity

    randyqx committed Apr 26, 2017
Commits on Apr 19, 2017
Commits on Apr 18, 2017
Commits on Apr 10, 2017
  1. changed wgets for keying

    randyqx committed Apr 10, 2017
Commits on Jan 27, 2017
  1. Accumulate primary keys instead of objects to work around Django cach…

    sraustein committed Jan 27, 2017
    …ing.
    
    This is nasty, and I still don't entirely understand it why this was
    happening.  We collect ca_detail objects during bulk ROA processing,
    so that we can defer manifest and CRL updates until the end of the
    batch.  Somehow, Django's caching code was causing the parent CA's
    issued serial number to roll back as part of this caching, which
    caused us to reuse serial numbers.  Which is (very) bad.
    
    Replacing the collection of ca_detail objects with a collection of
    primary key values for those same ca_detail objects seems to have
    worked, presumably because it lets us force creation of a new queryset
    when it's time for us to process the relevant ca_detail objects.
    
    The question is how many other booby traps like this might be lurking.
Commits on Jan 15, 2017
  1. Wrong Django voodoo in previous fix.

    sraustein committed Jan 15, 2017
    The rubber chicken needs to dance around the circle once, widdershins.
    Obviously.
    
    For future reference, the syntax for forcing queryset evaluation is
    "list(blarg.all())", not "[blarg.all()]".  In this case it doesn't
    seem to be necessary.
Commits on Jan 14, 2017
Commits on Jan 13, 2017
Commits on Oct 10, 2016
  1. Update to OpenSSL 1.0.2j.

    sraustein committed Oct 10, 2016
    There would (probably) be no security issue with continuing to use
    OpenSSL 1.0.2h for RPKI, but it's usually best to stay current.
    
    Update the update-snapshot script to use git instead of svn.
Commits on Sep 30, 2016
  1. warn of xenial only

    randyqx committed Sep 30, 2016
    Signed-off-by: Randy Bush <randy@psg.com>
  2. warn we only support Xenial

    randyqx committed Sep 30, 2016
    Signed-off-by: Randy Bush <randy@psg.com>
Commits on Aug 9, 2016
  1. Jessie comments; spelling/typo/trailing-whitespace cleanup.

    sraustein committed Aug 9, 2016
    Randy accidently pushed a bunch of unsigned commits.  I reviewed
    them before signing this one.  Randy's knuckles have been whacked.
  2. typo

    randyqx committed Aug 9, 2016
  3. relative github ref

    randyqx committed Aug 9, 2016
  4. fix refs

    randyqx committed Aug 9, 2016
  5. fix ref

    randyqx committed Aug 9, 2016
  6. fix ref

    randyqx committed Aug 9, 2016
  7. try to fix ref

    randyqx committed Aug 9, 2016
  8. push the sucker

    randyqx committed Aug 9, 2016
  9. add quickstart

    randyqx committed Aug 9, 2016