Comment #21 on issue 2496 by [email protected]: Add announcements to the upper
right corner of the website
http://code.google.com/p/lilypond/issues/detail?id=2496
Regarding comment #17: this is a case of 2.15.38 not getting out, not a
case for 2.16 not getting out. A release candidate is not a stable
release: what has blocked the stable release so far is that no release
candidate survived the 2-week settling period without a (usually unrelated)
critical regression being discovered. A delay here means that we have more
time to discover regressions.
Basically our release criteria are such that we'll get a stable release
once everybody is tired of looking for bugs and funding regressions.
So it makes little sense to let the stable release excitement determine the
treatment of this bug.
In fact, we likely don't want to have it stomp over the next release
candidate again, and it is not like we will release 2.16 tomorrow.
What is at issue here is that the bug precludes 2.15.38 from being
released. Since the release procedure even for unstable releases is
costly, a fix should be tested well enough before the procedure gets
started again.
To me it looks likely that the lines after
# Website post-processing
in make/website.make are at issue here, but I have no idea how this is
intended to work and whether or not the tweet.xml needs to be present on
the installed website or not.
It is not exactly news by now that changes of the website structure require
checking the respective commands to work as well.
