There is no Bugzilla entry for what I reported. I do not normally report most problems to Bugzilla until the release gets more stable. Usually by the time that the entry is acted on the test release is obsolete and the entry just becomes clutter. If this problem remains in the system then I will report it. About 80% of our problem list vanished from FC2 to FC3 and what will happen with FC3 I do not know. I suspect the bugs will be replaced. Many of these problems are like reporting checksum errors on iso files. It doesn't help the file with the error because it is obsolete. It just helps insure that the next release doesn't have an error.
On Tue, Nov 30, 2004 at 07:04:15AM -0500, gslink wrote:
There is no Bugzilla entry for what I reported. I do not normally report most problems to Bugzilla until the release gets more stable.
....
About 80% of our problem list vanished from FC2 to FC3
Quite likely because somebody else did bother and filed a report.
Yes, a search for already present bug reports could be easier and smarter (on occasions I had troubles finding things which I knew were there) but testing without reporting findings does not sound very productive.
BTW - the problem with a vanishing "/usr" should be fixed in freshly available initscripts-8.00-1. See https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=139656 if you want to know more.
Michal
You forget something. ALL bug reports MUST be properly documented. You can't just report that it doesn't work. The cost of a usefull bug report is anywhere from $500 to $15000 for a bug such as this because it is necessary to spend the time to find out what is really wrong. Some of these bugs cost so much we can't afford to report them. That was the case here. It is simply not obvious what code is actually causing this problem. We have several of these problems but all are going to need lots of digging even to get a case that fails every time and a bug that can't be reproduced is not the proper subject of a bug report.
On Tue, Nov 30, 2004 at 06:10:20PM -0500, gslink wrote:
You forget something. ALL bug reports MUST be properly documented. You can't just report that it doesn't work. The cost of a usefull bug report is anywhere from $500 to $15000 for a bug such as this because it is necessary to spend the time to find out what is really wrong. Some of these bugs cost so much we can't afford to report them. That was the case here. It is simply not obvious what code is actually causing this problem. We have several of these problems but all are going to need lots of digging even to get a case that fails every time and a bug that can't be reproduced is not the proper subject of a bug report.
That's not necessarily the case for an open-source project. Many other people may be experiencing similar problems, and by reporting it, you give a starting point for bringing those people together to all contribute their experiences -- maybe dividing that $15,000 of work over 1500 users, which doesn't seem so bad.
-
gslink -
Matthew Miller -
Michal Jaegermann