Re: [pcre-dev] Future maintenance arrangements for PCRE

Inizio della pagina
Delete this message
Autore: Daniel Richard G.
Data:  
To: pcre-dev
Oggetto: Re: [pcre-dev] Future maintenance arrangements for PCRE
[Sorry Bob, had to lowercase your subject :-]

Comments:

On Wed, 2007 Feb 21 11:55:11 -0500, Bob Rossi wrote:
>
> I've told Philip this already, but to make it public, I would prefer
> svn because I know it well, and I know it works. What do other people
> prefer?


I doubt that any other SCM will have real traction as a better choice.

> > Should access to the SCS be limited to those with accounts on sesame?
> > Any kind of anonymous access would have to be appropriately protected -
> > I have no knowledge of that side of things. Alternatively, we could put
> > out snapshots as the Exim project does.
>
> You could do it like other open source projects do it.


I.e. anonymous read-only access, and commit access through SSH accounts
given through developers... this seems to work well for most people.

> First, even though it's a pain, I think it would be beneficial to consider
> asking people to sign a copyright assignment form like the GNU project does.
> I like that PCRE is open source and under the BSD license, and would like
> to keep it that way. Does any of you provide on giving money to hire the
> lawyers? :)


I don't know about that. GNU/FSF demands the assignments so that they can
aggressively pursue license violations, IIRC. But with PCRE under the BSD
license, this is somewhat of a moot issue, yes?

It's also a trust issue: an OSS project with copyright-assigned
contributions could unilaterally be changed to a proprietary license by the
assignee(s). Most people are confident enough that RMS et al. won't have
such a change of heart, and so will gladly sign the forms, but I doubt any
of us here have such an ironclad reputation. Bottom line: fewer
contributions.

> Now, as far as using the repository. I think the PCRE repository should
> have read only access to the world.


Yep.

> As far as commit access to the repository, I like gdb and gcc's
> approach. There is a MAINTAINERS file, that describes the maintenance
> procedure. PCRE would probably have only a few categories.
> "Global Maintainer", would have the authority to accept or reject
> patches for the entire project. This, for example, would initially
> be Philip, and any one he deems appropriate.
>
> "Responsible Maintainers", These are developers who have expertise
> and interest in a particular area of PCRE. This, for example, would be
> me for the autotools build system and AnotherUnnamedDeveloper for the
> CMake build system.
>
> and then finally, but not necessarily,
>
> "Write After Approval Maintainers", These are developers who submit
> patches, get peer reviewed, modify patches to everyone's pleasure,
> fix the documentation, .... And then finally when they get permission
> are allowed to commit. We could allow the global maintainers to do
> the committing only, but the revision history would be lost. Also,
> having "Write After Approval Maintainers" is a good way to lure
> suckers, (Oops, I meant developers) into working on your project.


IMHO, all this is overkill. GCC and GDB are enormous projects, with
well-defined branches of interest to different people (different arch
backends, etc.), and lots of activity. PCRE is, all told, a small project.
I think three or four people with commit access can easily keep abreast of
what the others are doing.

> > Bugzilla:
>
> > There aren't many bug reports for PCRE, so it may be that something as
> > heavyweight as Bugzilla is somewhat of an overkill. However, since it is
> > already running on sesame, it might be useful to use it to keep track of
> > what reports there are.
>
> I've never used Bugzilla. I have used Mantis (http://www.mantisbt.org/)
> and it is reasonable.


I don't have much to offer on Bugzilla vs. Trac vs. somethingorother, other
than to implore avoidance of "third leg syndrome." E.g. if Trac has a
discussion forum, but the pcre-dev list is where all the action is, then
the former should be disabled, or otherwise made inaccessible---lest
hapless users come across it and gather that PCRE is an abandoned project.
(You see this problem all the time with SourceForge projects...)


--Daniel


-- 
NAME   = Daniel Richard G.       ##  Remember, skunks       _\|/_  meef?
EMAIL1 = skunk@???        ##  don't smell bad---    (/o|o\) /
EMAIL2 = skunk@???      ##  it's the people who   < (^),>
WWW    = http://www.******.org/  ##  annoy them that do!    /   \
--
(****** = site not yet online)