6 days, 76 downloads, 1 issue

[Matthias Andree]

David Relson <relson at osagesoftware.com> writes:

> On Sat, 19 Nov 2005 09:26:52 -0300
> Martín Marqués wrote:
>
>> El Sáb 19 Nov 2005 06:22, Matthias Andree escribió:
>> > David Relson <relson at osagesoftware.com> writes:
>> > >
>> > >    bogofilter-1.0.0 ?
>> >
>> > Documentation perhaps, and waiting for more feedback with Berkeley DB
>> > 4.4.16. It passes "make check" for me, which is why I'm advertising it.
>> >
>> > 76 downloads does not sound ample, yet the FreeBSD port is at 0.96.6 -
>> > and it causes the computers to download from SourceForge.
>> 
>> If you need testing I could step up my bogofilter up to the version in Debian 
>> unstable on my local server (not many users, but mainly my test machine).
>> 
>> martin at bugs:~$ apt-cache policy bogofilter
>> bogofilter:
>>   Instalados: 0.96.2-1
>>   Candidato: 0.96.2-1
>>   Tabla de versión:
>>      0.96.6-1 0
>>          50 http://ftp.de.debian.org unstable/main Packages
>>  *** 0.96.2-1 0
>>         800 http://ftp.de.debian.org testing/main Packages
>>         100 /var/lib/dpkg/status
>>      0.94.4-1 0
>>         500 http://ftp.de.debian.org stable/main Packages
>> 
>> The only issue is that I don have db 4.4.x installed, and it doesn't seem to 
>> be packaged for Debian yet. Is that a problem?
>
> Matthias,
>
> As you know, not much code has changed since the current stable
> release, which is 0.96.2.  This morning SourceForge is showing 0.96.6
> downloads at 98 and 0.96.2 downloads at 785, for a total of 883 (not
> counting text files).  These numbers don't include the intervening
> versions or counts for other download sources.
>
> I hadn't been aware of db4.4's release until I saw your CVS commits
> yesterday.  Evidently it new.  Do we even want to mention it in 1.0.0
> documentation?  Given our lack of experience with it, I'm inclined to
> say the answer is "No."

Given we'll be stuck with 1.0.X for a loooong time, and estimating we'll
mostly see usability improvements going into 1.1.X because the core
works, I'm inclined to vote Yes, that is why I made those changes in the
first place. I checked through the "upgrading applications to the 4.4
release" documents of BerkeleyDB, and there are zero changes we need to
make.

Berkeley DB 4.4 has many usability improvments, however we cannot use
them without dropping support for 4.3 and older at the same time, and
I'm not ready to do that, or fork a new datastore_db44.c, or add more of
the ugly #ifdef or #if DB_AT_LEAST() graffiti.

> Martin,
>
> I don't have db4.4 installed either.  It certainly isn't required for
> 1.0.0.  It would be great if you upgraded to 0.96.6 and gave it a
> test.  I don't anticipate any problems, but one never knows.  Surprises
> _do_ happen.

Indeed. Only recently, we figured that the fetchmail 6.3.0 release
candidates (reported against -rc8, -rc9 the current one, -rc10 coming up
early week with the fix) wouldn't let itself be installed on systems
without Python.

I seriously do expect we'll be seeing a 1.0.1 release 10 days after
1.0.0 proper, and I do expect the same for fetchmail only it will be
named 6.3.1 rather than 1.0.1. :)

Those new releases are going to see many more downloads than anything
that looks as though it might moved in the four weeks before.

Oh, and the compelling reason not to let 1.0.0 out the door is we still
haven't completed our security announcements for 0.96.2. Perhaps we
should bend the rules a bit and declare 0.96.6 stable, mention in the
promotion announcement that it supports Berkeley DB 4.4 (and the
previous Berkeley DB versions back to 3.X whatever), and wait 10 more
days. Perhaps I can then find the time to look into the 0.96.2 security
problem, as the fetchmail front is really getting calm now.

-- 
Matthias Andree