Developers/Release process

From sigrok
Revision as of 18:20, 17 October 2018 by Abraxa (talk | contribs)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to navigation Jump to search

This is a list of steps we perform when creating a new release:

  • Check if the manpages are up-to-date.
  • Test that make distcheck works without errors and creates a .tar.gz file. Unpack that file somewhere.
    • Test that no files are missing, and no extra/unneeded files are in there (non-public header files, *.o files, and that kind of stuff).
    • Test that building from that unpacked directory works without errors.
    • Test that installing from there works.
    • Test that running/using the newly installed library/program works fine.
    • Test at least the demo driver, one hardware LA, one non-default input- and output driver, and one protocol decoder. The more tests the better, of course.
  • Do all of the above tests on all supported platforms if possible, e.g. Linux, Windows, Mac OS X, FreeBSD.
  • Write release notes (containing user-visible, important changes) in the respective NEWS file(s).
  • Increase the respective package version number.
    • If there were any backwards-incompatible changes in libsigrok and/or libsigrokdecode, increase the respective lib version numbers too.
    • Then, push your current status, including the version number change commit via git push.
  • If everything works OK, tag the new release in git via (for example): git tag -a "libsigrok-0.1.0" -m "libsigrok 0.1.0 release" <hash>. Replace <hash> with the commit hash that should be tagged.
    • Verify that the tag is placed correctly via git tag or gitk.
    • Then, push it via git push --tags.
  • Now create the final tarball via make distcheck as described above, and upload it.
  • In case of a client program (sigrok-cli / PV), create all pre-built binaries, upload them and make them available on the wiki for future reference.
  • Announce the new release in the blog and on the mailing list.