Re: [RFC] libevent 2.0 transition
On 06/19/2011 04:30 PM, Leo 'costela' Antunes wrote:
Please subscribe to this (being filed) bugreport.
I uploaded a new version of libevent to experimental which I'd like to migrate
to unstable as soon as possible. Since I'm the new guy helping out with libevent
and since this would be my first "bigish" transition, I'd like to run it by the
release team to make sure I'm not stepping on anyone's toes.
The new version is a jump from the 1.* series to the 2.* series, which certainly
breaks the ABI, but shouldn't break the API for most cases. However, since
libevent exposes a relatively big number of internal structures in the headers,
there are ways to abuse the API in non-compatible ways.
I've rebuild all reverse-build-deps against the new version and only 7 from the
32 fail to build (more details below).
Now I intended to write to -devel with all affected maintainers CC'd, requesting
further testing and input and after all have ack'd the problems and have
solutions in place, I'd go ahead with the upload to unstable.
Does this sound like a reasonable plan? Did I miss something important?
About the FTBFSs:
beanstalkd: syntax error; maybe using some changed define?
ladvd: builds correctly, fails on 1 of 6 tests ("HTTP request failed")
forked-daapd: includes event-config.h, which has been moved to
event2/event-config.h (I could provide a compat-symlink)
python-event: bona fide build failure (I couldn't grok the underlying reason
just by looking at the log)
honeyd: doesn't seem related (configure: error: Couldn't figure out how to
Leo "costela" Antunes
[insert a witty retort here]
Mehdi Dogguy مهدي الدڤي