Re: [boost] Re: [gmane.editors.lyx.devel] possible boost patch [w asdying in boost/regex]

From: David Abrahams (david.abrahams_at_[hidden])
Date: 2002-07-02 09:06:01

From: "Lars Gullik Bjønnes" <larsbj_at_[hidden]>

> | That's entirely appropriate. I don't want someone patching my library
for
> | me unless it's an emergency and I've become completely unreachable.
They'll
> | probably get it wrong. Likewise, I'd much rather wait until the
situation
> | is intolerable than introduce a bug in someone else's code.
>
> So the the silence is appropriate?

No, silence is not appropriate. That's part of why we brought Bjorn aboard.

> btw. this is also what I regard as one of the current problems with
> boost: it is "my" and "their" code.

It's a practical problem. None of us has the time to become as familiar
with the other libraries as we are with the ones we've authored ourselves.

> Sure I am all for the careful introduction of patches and new code, it
> is the current practice that only the library owner should answer that
> I after.

Should a developer give you some answer even if he doesn't intend to do
anything with your patch?

> I do not even need to get my patches put into boost (we have lived
> with small stuff you do not want in boost for a long time), I only
> want some reaction.

Perhaps you and Bjorn could work together to figure out what an appropriate
reaction might be.

> >> I cannot do much more than send patches and test...
> >
> | That's always appreciated. Which libraries have you sent patches for
other
> | than regex?
>
> limit, thread, lwm ... compile warnings etc.
> (what is this? some kind of a test?)

No, I was curious. I wanted to know if responsiveness (or lack thereof) is
correlated to particular library authors.

What is lwm?
compile warnings is not a library -- did you send this information about a
particular lib?

> >> You mean apart from testing and using boost all the time and sending
> >> patches when I discover I can fix a problem myself?
> >> So IMHO I am already trying to make a difference.
> >
> | Try asking yourself how you could make a difference in the
responsiveness
> | to patch postings.
>
> Yes, how?

I suggested you ask *yourself* <.5 wink>

> >> | BTW, other projects use the (IMO lousy but perhaps better than
nothing)
> |          ^^^^^^^^^^^^^^
> >> | SourceForge patch tracker; it's taken for granted that patches will
be
> >> | forgotten unless they're in the tracker.
> >>
> >> Other projects or Boost?
> >
> | How could I possibly make it clearer?
>
> You your statement is then that "Perhaps Boost should use the SF patch
> tracker?"

Yeah. One problem with posting them to the mailing list is that they
typically get line-wrapped into uselessness.

> | Do we really have a big problem with responsiveness to these
> | patches?
>
> No but it is annoying....
> Actually it seems that the problem is almost more present for the
> small obvious patches than larger ones.
>
> f.ex. adding cpu defines to detail/limits.hpp

That was easy. Unfortunately, Jens has become so busy in the past year that
he almost never posts here anymore, so that may explain why nobody
responded earlier. Perhaps you would like to take over maintenance of
limits.hpp (if Jens agrees)?

> | How long have you had to wait, how many times?
>
> Do I count? (no)
> For me to even begin feeling this as a problem it must have happened a
> few times...

I think you should be more forgiving of such minor incoveniences,
especially from an all-volunteer organization, and most especially when
we're already taking steps to address the issue (by appointing a
maintenance wizard).

> btw. Having the gmane list<->newsgroup as a read-only gateway does not
> help.

It's not a read-only gateway, but the posting address does have to be
subscribed to boost (using "nomail" if you don't want to get mail). People
post through gmane all the time.

-Dave