[BTC-dev] The Bitcoin Foundation: STATE OF BITCOIN ADDRESS

Tue Mar 3 03:54:50 UTC 2015

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA512

 ..::[ The Bitcoin Foundation: STATE OF BITCOIN ADDRESS ]::..

     [        Date: 2015.02.28                          ]
     [   Co-Chairs: mod6 [R.01] && ben_vulpes [R.02]    ]

0x00] Introduction

   It is with great honor and privilege that The Bitcoin Foundation embraces
   this opportunity to address the public on the state of current progress,
   obstacles and continuing steps in our mission.

   February was another month with intense activity for The Foundation.  Within
   one day of the January Address being published, we started seeing a defect
   that had not been seen before; a failure to Verify Signature on a TX within
   block 168`001.

   The defect was not consistent across environments and build configurations
   but was reproduced on enough environments to cause concern.  This prompted
   us to hold up The Foundation publication of a patched milestone release
   package as stated in the January Address.  Further details are contained
   herein.

0x01] Accomplishments

   [ Verify Signature Failure at block 168`001 Resolved ]:

     As stated in the Introduction, during the first days of February we were
     working through completing regression testing for the release of v0.5.3.1
     with applied patches.  When we discovered during a blockchain full-sync
     regression test, a defect appeared in block 168`001.

     Upon a close examination of the log [R.03], it was a tx [R.04] where it
     encountered a Verify Signature failure.

     This issue was investigated for over a week with a number of different
     people from #bitcoin-assets involved to help to reproduce this defect.
     The Foundation thanks all of those in #bitcoin-assets who worked hard
     to test and reproduce this defect.

     mod6 upgraded the local his OpenSSL version from v0.9.8o to v1.0.1g and
     no longer encountered the defect [R.05].  Many people re-tested with the
     newer version and this seemed to resolve the issue.  Although, it is
     inconclusive at this time as to why this problem occurred and, why it is
     inconsistent across different configurations and environments.
     More investigation is needed here as time permits.  It is also worth
     noting that while other individuals did successfully pass block 168`001
     using OpenSSL v0.9.8o; using OpenSSL v1.0.1g is now recommended.

   [ Static Build Completed ]:

     As a last requirement for the Reference Implementation (R.I.) milestone
     release v0.5.3.1: Allow for a statically linked ELF output binary.
     Currently, the v0.5.3 codebase file `makefile.unix' does include some
     macros for allowing a static build.  However, this does not work without
     two changes.

     asciilifeform [R.06] created a build script `auto.sh' for the
     portatronic project [R.07].  The script is used to statically build a
     cross-compiled ELF binary for an ARM processor.  mod6 modified this script
     and made the necessary changes to build static libraries of:
     OpenSSL v1.0.1g, BDB 4.8.30 and Boost 1.52.0.  These are then (along with
     a number of other necessary libraries) statically built into the ELF
     output binary (bitcoind).  This process has been verified and tested, but
     requires further regression testing.

     The a patch file containing the changes to `makefile.unix' as well as
     a patch file for asciilifeform's original `auto.sh' will be published
     to the btc-dev mailing list [R.08] soon.  These will be also included in
     the R.I. milestone release: v0.5.3.1.

0x02] Complications and Obstacles

   [ Release Package Not Yet Published ]:

     In January's Address The Bitcoin Foundation said that a milestone release
     (v0.5.3.1) was imminent and would be released in February.  This did not
     happen for reasons listed herein.

0x03] Continuing Steps

   [ Patch & Compile Script ]:

     Last month a bourne shell script was written by mod6 to automate the
     applying of patches and building the R.I.  This script was later patched
     by both jurov [R.09] and ben_vulpes [R.02].  Although this script mainly
     seems to work, there are some people who have still reported some issues.
     These mainly relate to local configuration issues or missing tools such
     as `curl` or `gpg`.  Since many people are hopefully going to be using
     the build script to help test patches for the R.I.  mod6 re-wrote this
     script in perl to mitigate these minor problems programmatically.

     Thus far, a handful of people have used this new perl script and it seems
     to be working well.  However, before it is published to the mailing
     list [R.08] or to The Bitcoin Foundation website [R.0A], mod6 asks for
     your help testing the script.  Please inquire at #bitcoin-assets.

     Hopefully the testing on this script will be completed quickly and will
     be made available for general use.

   [ Regression Testing ]:

     To meet our goal of publishing a release of the patched v0.5.3 codebase
     this month; we must regression test the statically built ELF which
     includes all patches in the current v0.5.3.1 configuration.  When testing
     is completed a release tarball will be constructed and a notice will go
     out to the btc-dev [R.08] mailing list.

   [ Portatronic ]:

     asciilifeform [R.06] has continued to put his efforts into a build of the
     R.I. for a field-ready device.  Various patches and configurations have
     been created for testing.  It is exciting to watch this project develop.
     The Foundation thanks asciilifeform for his hard work on this project and
     also danielpbarron [R.0B] for testing and routine updates on the
     portatronic testing status.

   [ OpenBSD ]:

     mod6 worked to produce a working build of the patched R.I. on OpenBSD 5.6.
     With certain patches, the configuration does work and does sync blocks.
     What remains is to create a static ELF build of the R.I. on OpenBSD and
     test that also.  Work is halting on this for now as the main focus will
     be the v0.5.3.1 milestone release.  Stay tuned for more on this.

0x04] Conclusion

   February was a month that started out with complications but finished on a
   high note with challenging problems resolved.

   The Bitcoin Foundation would like bestow our sincerest thanks and gratitude
   to the contributors and community for its support and insight.

[ References ]:
  [R.01]: 027A 8D7C 0FB8 A166 4372 0F40 7217 05A8 B71E ADAF
  [R.02]: 4F79 0794 2CA8 B89B 01E2 5A76 2AFA 1A9F D2D0 31DA
  [R.03]: http://pastebin.com/raw.php?i=vmMmRA92
  [R.04]: 2c2314f353013f920d8fbfde242d7d23ba4cb9b97dc24f481dd0ccfd8f56324c
  [R.05]: http://pastebin.com/raw.php?i=VVz5j1dW
  [R.06]: 1721 5D11 8B72 3950 7FAF ED98 B982 28A0 01AB FFC7
  [R.07]: http://therealbitcoin.org/ml/btc-dev/2015-January/000033.html
  [R.08]: http://therealbitcoin.org/mailman/listinfo/btc-dev
  [R.09]: BBB0 A999 5003 7551 F533 850A 677A BD62 D0AE E7D7
  [R.0A]: http://thebitcoin.foundation/
  [R.0B]: 1559 34BD D16E 8EAF 4493 CB9C B36A E984 9D96 1AC9

EOF
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.13 (FreeBSD)
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=yDud
-----END PGP SIGNATURE-----