Maven and Scarab suck? And all along I thought I was on a bleeding edge! ;)
================================================================= Jeffrey D. Brekke Quad/Graphics [EMAIL PROTECTED] http://www.qg.com > -----Original Message----- > From: Stephen Haberman [mailto:[EMAIL PROTECTED]] > Sent: Sunday, July 21, 2002 1:28 AM > To: 'Turbine Torque Developers List' > Subject: RE: Requirements to submit patches > > > Hi Ben, > > Since no one else has replied yet, I'll give you my two cents > as to what > I've gathered in the last month or so of sending in patches every once > in awhile. This is by no means formal...a response from a dev would be > much better. > > > * Use case > > I don't think I've seen a formal use case for a patch before, though I > imagine if you did a good one, it would only help your case to get the > patch committed. > > > * Patches > > Everything in cvs diff/patch form, with -u, really helps. Attached > instead of included in the email is preferred, I think. > > > * Test case > > If you can run the test suite, "ant test", and get a successful build, > it'll help your case. What'd be really cool, though I don't > know if the > test suite is currently setup for it, is if you could submit > a test case > that is broken in HEAD, but then your patch fixes it. > > (e.g., looking through the archives, if you had had a test > case for your > setByPeerName that the devs could drop in via patch and have to > automatically work in the test suite and pass, it would have been > convincing). > > (Also, I just noticed Runako Godfrey's patch he submitted with a test > case and it makes a very convincing argument...I can't see this not > being committed). > > I had mentioned refactoring the test suite into something that's more > extensible and easy to add test cases for bugs to a week or so ago and > it got good reception but, alas, it is yet another thing I > can't seem to > find time for. > > If you're up for it, I think ideally the test suite should be to the > point where if a user finds a bug in the templates, they > should be able > to non-trivially add a project-schema.xml + java code to show > what their > patch fixes. > > > What about? > > * documentation > > Documentation is good if your patch changes the behavior reflected in > existing docs, though I think it's perfectly acceptable to post your > patch first saying that if it'll get committed, you'll write the docs. > If it won't be committed, there's so sense spending the time changing > the docs. > > Other than your concerns, I think the biggest problem is that the > community is a bit fragmented. There are great developers listed in > project.xml, but they're all off working on Scarab, T3, Maven, OJB, > Jelly, etc., etc., etc. > > On a short side note, Jason van Zyl makes the comment that > cross-pollination of ideas across projects is what results is great > software, and I generally agree with him, as these guys come up with > great stuff. But Torque doesn't seem to currently be the > darling project > to work on, so everybody is off cross-pollinating in other projects. > > Though I dunno, I tend to over react about such things (you can search > the archives and find the previous time I gave a similar, > though longer, > soap box speech :-). Martin lurks and commits things he sees as > interesting, as do most of the other devs. And they're pretty > good about > responding to bugs. > > BUT, the most damning thing is that when non-trivial patches (or even > trivial, as in your setByPeerName patch) are posted and the devs who > should have the responsibility of fostering the community and giving > good feedback, even to bad/non-neccesary patches, don't say a damn > thing. > > Really, people, this pisses me off more than anything. Patches get > ignored a lot and it just kills the community. > > If patches are bad, just tell us. Feedback is good. It'll make > wanna-be-contributors more mature, make better patches, and > in the long > run improve the Torque and it's community. > > Perhaps we could use Scarab to track patches/new feature, > etc.? I'm not > familiar enough with Scarab to know if it's a good idea, but perhaps > it'd help remind devs to go through the patch queue every > week or so and > at least respond? I imagine it'd be a lot more organized > that a folder > full of email each dev has to remember to parse through. > > Anyway, I apologize; this email got a little long, as mine tend to do. > Does anybody even make it to this point in my emails? > > Hm...Maven and Scarab suck! > > (Ha, now we'll see if anybody reads passed the first few > paragraphs and > refutes my spurious remarks. :-) > > - Stephen > > > -- > To unsubscribe, e-mail: > <mailto:[EMAIL PROTECTED]> > For additional commands, e-mail: > <mailto:[EMAIL PROTECTED]> > -- To unsubscribe, e-mail: <mailto:[EMAIL PROTECTED]> For additional commands, e-mail: <mailto:[EMAIL PROTECTED]>
