GCC 4.5 Release Candidate Is Coming

[phoronix]

Phoronix: GCC 4.5 Release Candidate Is Coming

GCC 4.5 was not in a good shape as of the middle of March with it still not being ready due to 16 outstanding P1 regressions, but over the past two weeks, developers have feverishly been fixing these bugs and the count is now down to zero. For P2 regressions, 17 of them have been fixed too over this time span, which brings the second-tier bugs down to a count of 81. There is also one new P3 regression bringing its count to three...

http://www.phoronix.com/vr.php?view=ODExNA

[KDesk]

Why do the release a software as stable knowing that it has bugs?

So that distros start using it, and start sending fixes?

[mat69]

Why do the release a software as stable knowing that it has bugs?

So that distros start using it, and start sending fixes?

I guess I have a reason, but I'm not sure: Because they always did it that way!

Imo this is mostly a non-issue, any compiler has known problems and often known before the release.

[pvtcupcakes]

Why do the release a software as stable knowing that it has bugs?

So that distros start using it, and start sending fixes?

There is no such thing as bug-free software.

[rohcQaH]

Why do the release a software as stable knowing that it has bugs?

because otherwise nobody would ever release anything. There is no bug-free software, not when you're talking about software the size and age of gcc.

edit: hi pvtcupcakes

[Remco]

There is no such thing as bug-free software.

But there is such a thing as software without known bugs. Of course it's not practical to always solve all confirmed bugs before a release. But it's theoretically possible to do bugfixes until the bug tracker is empty. I'm curious what would happen if all software projects would pick up a policy: no new features until the confirmed bug list is empty.

[rohcQaH]

But there is such a thing as software without known bugs.

only if the developers are totally ignorant, there are no testers/users, or there's some weird definition of "bug" that allows one to classify all bug reports as invalid.

Show me one large project with an empty bug tracker before release.

I'm curious what would happen if all software projects would pick up a policy: no new features until the confirmed bug list is empty.

No new releases for the next 5 years. As a consequence, distributions start randomly shipping git versions. All hell breaks loose.
Oh, and no Linux-user is allowed to buy new hardware, since new drivers = features.

[whizse]

Maybe Remco is thinking of something like this:
http://www.schneier.com/blog/archive...g_a_compu.html

But it's a very very expensive/time consuming process, then there's also the question of how you define a "bug", errors in the math, and dealing with faulty hardware etc...

[V!NCENT]

Show me one large project with an empty bug tracker before release.

The latest KDE 3.5.x release has no known bugs.

[Kano]

Well KDE 3.5 can not mount usb storage >1TB.