On 10/13/06, Domenico Andreoli <cavok@debian.org> wrote:
On Tue, Oct 10, 2006 at 01:25:37PM -0400, Jay Berkenbilt wrote: > > If my analysis/understanding is correct, icu, xerces27, boost, and > parrot will need to be hinted into etch together but this is not yet > possible because boost and parrot have other issues right now that > ... boost build looks screwed. pts also looks screwed. ehi.. also http://people.debian.org/~igloo looks screwed... hmm... it looks like boost is ready to enter testing but powerpc, mips, mipsel and of course m68k are still missing.
Boost is screwed by design, and will never be ready for testing. Well, if you want very slow and flakey software that gets SIGSEGV and SIGBUS, maybe boost is appropriate. Contemplate the effects of things like auto_ptr on software maintenance and reliability. Imagine an app with threads. IMHO, I've seen enough of this shit to know. Boost is used by people who can't code, but pretend that they can. It's very much like using lots of 20-letter words in English text: it looks impressive to those who don't know, but the result is a pile of shit. I deal with Boost-infected code every day at work. The bugs are hard to exactly locate, but the general location is easily determined by looking for the Boost shit. It's nearly a given that the bugs will be in the Boost shit.