[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

Re: Sendmail update



Jef Spaleta wrote:
chabotc has confused the crap out of me, with mixed messages, like a
drunk girl a frat party passed out with her head in my lap. I for one
don't think its so great the chabotc has rolled up
packages...considering the previous disdain chabotc has shown for
running rawhide packages, as part of the beta process. I have to
question chabotc's motivations...and I don't see why any of us should
explicitly trust chabotc packages...especially unsigned packages.

I am sorry to have confused you Jef. If you are truely interested, i will try to clarify a few points.


First of all, my email about "don't run rawhide" was triggered by some poor soul who upgraded totaly to rawhide packages.. And subsiquently broke his system (lucky for him, it was "just" non functional icons in konq. and not a total whipe-out). Then i went to state that rawhide is the current non-tested & non-tested for interoperability output of rh's build tree, so there is no guarantee something will work.. if thats what you hope for, then don't use it.

Whats more, as we come closer to release, or just after release, rawhide will be more of the experimental playground again it has always been.. It is not a pool of updates for the current release, it are packages that might, or might not ever make it into redhat and you should expect a lot of broken stuff in there... thats the whole point of rawhide (it is _not_ a continuous redhat beta). If people are tought now that they should run rawhide packages, without fully knowing what that entails and what the risks are, then that is a bad thing for them and we did them a disfavour.

That is not to say that 'fixes' also first show up in rawhide before they are tested and put into either the next beta, erata/updates or the next redhat release.

Now the package update you refer to in your reply was the sendmail update, where i expressly mentioned i got that package from the rh 8.0 updates repository, and _not rawhide_; And that the only difference is the 'erata define' in the spec file. How you missed that and imidiatly jumped to a reasoning that i send out mixed messages about this is beyond me, and seems designed only as flame material. (please correct me if i'm wrong). Your comment seems to be a little confused, at best, in this regards. So for now i'm assuming your main point of the email was telling me to tune it down a bit, because in some way i steped on your toes and your feelings hurt.

So my apologies for causing that emotion with you, that was not my intend. Since i am not employed to assist people on this list, i can get pissed of when i personaly feel people are motivating (in what ever form) people to try things that could destroy their system. People are already confused about what they should and should not expect with beta's, and these people then using rawhide? That thought scares me, and it scares me to think what it could do to beta's and rawhide's availability to the public. Not to mention what it can do to the unexpecting end-user who finds a kernel bug can infact blow up his whole instalation and every other partition on his drive.

To address the other point: I do feel it is a bad idea to run rawhide packages to much on their beta system. The point of the beta is to do public QA on that combination of packages.. Sure you can get a specific fix from a redhat engineer's home dir or rawhide, but at large it's that specific combination of packages that we are testing

You could also choose to test rawhide packages, thats what it is there for, but do not confuse that with testing a redhat beta! With the beta you can have some expectations that it might work, with rawhide you do not. (Have you ever read the mission statement and intent for rawhide?) Confusing a beta testing process with the rawhide testing process, and the miss-understanding the risks attached to either, is a bad thing for the end-user, and for redhat.

You do make one other good point, there is no reason why you should trust my packages... i could've put backdoors in them, or whatever evil tricks. Thats why i explicitly explained how people could build these them selves from a trusted source, and only if that fails and you feel wurried about being rooted/hacked/loaded with worms, then the update would be a good idea.

Now i feel this thread should come to an end.. I try every time (though not always in as friendly tones, see the not employed part) to explain in detail why i'm saying this, and i see a responce with a smart wording designed to be inflamatory.. Not something i want to spend to much time on.

If you feel so inclined, feel free to ask any of us here about release processes, rawhide and all those issues, but please leave the personal, quipy attacks in /dev/null and try to understand why i took the time to write all those (and this) long emails.. if not, i sugest using procmail and :0: * ^From chabotc xs4all nl to /dev/null so you don't have to be bother by me anymore.

-- Chris





[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]