[Johann: this email went to me, and I don't think it's been to the
list yet, so I'm resending it there too]
Hi guys...I had an idea about this.....
On Mon, 16 Aug 2004 17:00:31 +0000, Johann Gunnar Oskarsson
<johann@xxxxxxxxxxxxxx> wrote:
> Steve Youngs writes:
>
> [snip]
>
> > o Constant communication between the bug submitter and Us. With as
> > much as possible being automated so it isn't forgotten about.
>
> This shouldn't be *too* automated though, the best way, I think is to
> have a mailing list subset for each bug*, the submitter can choose to
> subscribe to. This should be as easy as replaying to the ack mail.
> Then the submitter can get all traffic regarding that bug, be it
> developer discussion, closed becouse of duplication (in which case,
> the submitter should be subscribed to the previous bug), and finally
> notified when it's closed becouse of being fixed ;)
>
> * I have no idea how to implement this though ;/
>
> Replying to the ack mail also rids us of trying to contact invalid
> addresses and so on and so forth.
>
> But even if the ack mail is not responded to (simply becouse the
> submitter is already flooded with mail and is afraid of more or
> something) it should *not* be forgotten about.
>
I think this could easily be a web-service. We'd have a dedicated
server that takes care of all of the login information for bugzilla.
And then we'd have a standalone front-end and a built-in (in SXemacs)
front-end to the service. We basically facade or abstract all of that
nasty bugzilla information from the user when he/she submits a bug
report. This also allows us to pipe that bug submition not only to
bugzilla via the server backend, but also allows us to pipe it out to
everyone, respond to the user, and even if we make it smart enough
parse the bug report and send it to certain members of the team for
resolution.
Now the question remains.....how does lisp become web-service savvy.
I highly doubt anyone has written a web-service library for it. So my
next question would be can lisp access shared object libraries? I'm
sure it can i'm just wondering how much of a boogle that would be on
performance.
> > o A real live person to get into contact with the bug submitter
> > _within 48 hours_ of the bug report coming in.
>
> Emm, it's not like I'm against this, but could we have a scenario?
>
> > o In absolutely no circumstances will any bug report get lost or
> > forgotten or neglected. I want the bug tracker to send out
> > automated reminders to whoever takes responsibility for the bug.
> > The higher the priority of the bug, the more frequent the
> > reminders.
>
> Or in the case there is no one responsible, it should be sent to the
> one responsible for assigning the bugs ;)
>
> > Feature requests and support requests should work the same way as bug
> > reports. Although I believe that support requests should generate
> > contact from a real live person _within 24 hours_.
>
> Hmm, k, I think I'm pretty cool with that.
>
> > Patch tracking will probably work in a similar way too. I'd
> > definitely want `M-x sxemacs-submit-patch' to automate the process for
> > the user.
>
> Or M-x sexually-submit-patch ?
>
> > For us with patches, I'm not going to put too many restrictions on
> > it. Basically once you have commit privileges you'll be able to
> > commit when or what you like (there may be times when this isn't the
> > case). If you fuck something up you immediately remove it or fix it.
> > Simple.
>
> For the development branch, yes, but what about stuff we have claimed
> as stable? Then there should be manual scutiny of the patches, no?
>
> > Obviously it goes without saying that you'd inform us of your plans
> > here before you start committing oodles of new whiz-bang features.
>
> As will be stated in the PPM, right?
>
> > CVS can easily be made to post diffs to a "sxemacs-patches" mailing
> > list.
>
> But are we so hell bent on CVS?
>
> > Well that's pretty much how I see SXEmacs Issue Tracking. I've only
> > heard of three issue trackers... BugZilla, Gnats, and whatever SF and
> > Savannah uses. I don't know any of them all that well.
>
> And I know none of them at all..
>
> > Give me options folks! Is there anything out there that will give us
> > what I want?
>
> The only thing I can think of is M-x extrapolate
>
> Have fun,
>
> Johann
>
>
--
Samir: No one in this country can ever pronounce my name right. It's
not that hard: Na-ee-ana-jaad. Nayanajaad.
Michael Bolton: Yeah, well at least your name isn't Michael Bolton.
|