Slashdot Mirror
FreeBSD Ports Collection Breaks 10,000 Ports
sremick writes "After breaking the 9,000 mark in July, the FreeBSD ports collection was well on its way of crossing 10,000 by the end of 2003. Sure enough, we made it! According to freshports, the number of ports in the FreeBSD ports tree currently stands at 10,015. This little graph is also nice, though not completely current. Way to go, FreeBSD!"
That title makes it look like changes to the port system broke all the ports. Maybe "exceeds" rather than "breaks"?
I can't say that I don't give a fuck. I've just run out of fuck to give.
Seems development has been rather steady for a while now. I predict 11,000 ports in... 6 months.
Comment forecast: Bits of genius surrounded by a sea of mediocrity.
Yea, but if you factor in my 0 ports, then your 0.03% is comparably astronomical.
Your suggestion is as useful as suggesting that everyone standardise on a single Unix or a single distribution of Linux. Or a single editor.
> Even more man-hours would be saved if people wrote
> in ISO C using only POSIX functionality, without littering
> their code with Linux-isms, or worse still
> distribution-specific things. Creating a package is
> relatively easy once you can make the code compile.
>
But wait, POSIX is considered by *some* to be *broken* such as the threads model, handling of locking and the non-existence of asprintf(3) which I believe orginated in OpenBSD and is now available from all of the Freeix, but not Solaris.
Okay, easy, so let's choose a standard. Oh, picking at random <flame suit=on>How about everyone choose Red Hat Linux 9.0 with patches as of 20030318</flame>. Oh, 99% of developers don't like *that* too much, gee that's a surprise. Alternatively, let's stay permanently in the 1980s or establish a working group that can wrangle for years before making a decision.
There are some very useful functionality which is genuinely operating system-specific (see Linux capabilities). Other things like PAM would have been delayed/hindered if one was forced to implement it on 5 operating systems at once (Solaris, Linux, *BSD). Finally, purely technical differences of opinion will occur, software development is often a game of trade-offs.
Certainly, one should *try* to develop most of the code using POSIX only. In turn libc and kernel developers should think hard before introducing *another* syscall or library call. But OS-dependences will happen, it's the real-world. Deal with it. Any ports model will also have to take that into account.
Ah, the shameless trolling.
I'll humor you, and let's JUST say, (Although, I certainly disagree with your premise however.) FreeBSD is dying. What would your point be? It's still better than Linux, in more aspects than I have time to list here. Besides, The -core team won't evaporate, our community is too large, and too many skilled coders would love a commit bit.
(I've been running linux since kernel 0.99, and freebsd since kernel 2.2, I've got some clues about both os's, as I've hacked around in both kernels extensively too)
Well, going along with your logic, I guess Linux is dying too, since Microsoft still holds so much market share, and will continue to. Had you happened to notice the climb in w2k3 servers? It'd still be a POS no matter if 99% of Internet/Intranet Servers ran it. According to you, the open source community has quite a grim future.
I suppose you draw your conclusions differently,however I don't implement an operating system because it's the "hip" thing to do, or because of a "clique". Nor because, "everyone else does" - I run FreeBSD at every one of our locations, because I live and die based on how well my servers operate, and for the last ~10 years or so, I've been QUITE content thanks to FreeBSD. (There was a time I was quite content with Linux as well, that ended not long after the 2.0 kernel came on the scene, fyi.)
Fact: Clueless is as clueless does.
-mpf p.s. don't forget, freebsd existed when it's user-base was not even 1/1000th of what it is today.