view release on metacpan or  search on metacpan

Changes  view on Meta::CPAN

  * Documentation

0.56 2019-01-06
  * Made buildtool and few others not executable (MANWAR)
  * Send the "device-signature" header again
  * (Developer) diagnostics are now available in the code proper
  * Better error handling in case of credential failures

0.55 2018-12-03
  * Adapt to site (well, API-data) changes
    BAUFINANZIERUNG accounts will not be returned for the time being

0.54 2018-09-02
  * Add missing prerequisite for JSON, Future
    This addresses RT #127022 , reported by Slaven Rezic

0.53 2018-09-02
  + Add missing prerequisite for Filter::signatures
    This addresses RT #127022 , reported by Slaven Rezic

0.52 2018-09-02
  [Distribution]
  + This is most likely a breaking change. I've tried to keep the old interface
    alive and so far, the test suite passes with what I think are mostly
    data changes and removal of old tests that reached too deep.
  + Merge the API branch, as the old website got switched off.
  + There is no compatibility shim yet to restore the functionality of the old
    module. The tests will currently fail, sorry.
  + All files are now stored and read as UTF-8 indiscriminately.
  + Use the new API, which could be less fragile than the scraping approach
    used before.

0.51 2018-07-17
  [Distribution]
  + Fix prerequisites
    Test::MockObject was missing, MIME::Lite is only a test prerequisite

0.50 2018-07-16
  [Certificate fixes]
  + Fix SSL certificate check to reflect the (systems) merger of Postbank
    and Deutsche Bank
  [Code fixes]
  + Fix handling of session after downloading a CSV file

0.49 2018-04-21
  [Test fixes]
  + Better maintenance detection, mostly for the demo site
  + Better test resiliency, mostly for the demo site

0.48 2018-02-22
  [Certificate fixes]
  + Add some text aliases used/produced by some BSD version of SSL

0.47 2017-07-16
  [Certificate fixes]
  + Add one more numeric identifier for certificates
    This fixes RT #122506, reported by Slaven Rezic

0.46 2017-07-16
  [Code fixes]
  + Make module more resilient against old versions of IO::Socket::SSL
    which don't have OCSP support
    This was reported by the CPAN testers with at least IO::Socket::SSL
    version 1.967
  [Certificate fixes]
  + Update certificate

0.45 2017-04-17
  [Test fixes]
  . Fix test suite that broke due to . (dot) not being in
    @INC anymore, thanks to Ryan Voots

0.44 2016-11-22
  [Test fixes]
  . Remove the EURO sign from the canned account statement
    This should address RT #117204, again; reported by
    Slaven Rezic, thanks!

0.43 2016-10-15
  [Code fixes]
  ! We only enable fingerprint verification for OpenSSL versions
    below 1.0.2. For 1.0.2 and higher, we rely on OpenSSL to do
    The Right Thing.
    Note that Debian Jessie will stay on 1.0.1
  + Make past_days work better (provided by Harald Becker)
  + Reorder module loading so that Finance::Bank::Postbank_de doesn't
    get loaded twice.

0.42 2016-09-03
  [Certificate fixes]
  ! Mozilla::CA 2016-01-04 does not support the G3 certificate from Symantec, but postbank.de
    relies on that certificate. We now check the fingerprint of the postbank.de
    certificate instead of verifying the certificate chain.
    An alternative fix is to fall back to the 2013 version of Mozilla::CA.

0.41 2016-09-02
  [Code fixes]
  + Allow to specify the days to request the statement for
  + Better resiliency against where the EURO sign appears

0.40 2016-06-19
  [Code fixes]
  + Allow more numerical entities in certificate. Yay ASN

0.39 2016-06-12
  . Add license (Artistic license)
  [Code fixes]
  . Adapt to new number/currency format in the statement output

0.38 2015-11-15
  [Code fixes]
  + Adapt to the website changes (new form field names)

0.37 2015-07-31
  + New certificate text
  [Code fixes]
  + Use ->decoded_content when downloading an account statement, in case
    it gets gzipped
  . Allow OIDs and names in certificate
    Some SSL libraries know about the names for the OIDs, but older
    versions of OpenSSL don't.

    The real fix would be to teach OpenSSL the new(er) OIDs