Subject: Re: [boost] logging
From: Jamie Allsop (ja11sop_at_[hidden])
Date: 2011-07-17 18:43:30
On 16/07/11 17:18, Andrey Semashev wrote:
> On 07/15/2011 11:46 PM, Gordon Woodhull wrote:
>> I hope that Andrey speaks up and lets us know how his library is
>> coming along.
> Well, it's not going as fast as I'd prefer. Due to various matters I
> have little time to work on v2 lately but I'm still eager to bring this
> library to Boost. Right now I'm refactoring sink frontends to move
> record formatting to the sink frontend level. Also, the async frontend
> will use a new lock-free queue, which should improve performance. There
> are a few design decisions to make and quite some time to experiment
> with the new code, that's why it's taking so long.
Your efforts are appreciated. I think the reason for much of the
interest is that many of us are actually using your library 'as-is'
today. For those that are, we're going to have to refactor once a new
version is available. In that regard we'll have to deal with breaking
> Also there is much to do left, including the mentioned port to Phoenix
> v3. I didn't look into it yet, so I can't estimate how long it will
> take. Taking the chance, I wouldn't mind if someone wants to provide
> patches in this area.
Not sure if this was mentioned or not, but do you have a public repo
with your latest code so interested early adopters can help contribute
patches against the that?
> There are other things planned, some of which were mentioned during the
> review (trivial logging issues, channel name changing, etc.), some were
> identified by users (problems with initialization from settings,
> boundless record queue in async sink). And of course, docs need updating
> (although this is partially done).
It sounds like it might be worth getting the library to a 'stable
enough' state for an addition to boost and simply make it clear in the
docs that certain interfaces might be deprecated in a later revision?
Like I said many are using it now and the wider exposure might prompt
more volunteer help.
> Sorry, I know it takes too long, the review took place more than a year
> ago. I hope I can do better in the coming days.
Good luck! I had quite a few reservations during the review as I recall
but I am using your library because I do think it is a good library and
as I said all the work you've put into it is very much appreciated. I'm
looking forward to seeing some of the improvements you've made.
> Unsubscribe & other changes:
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk