Date: Mon, 13 May 1996 20:51:09 -0600 From: Nate Williams <nate@sri.MT.net> To: Josh MacDonald <jmacd@CS.Berkeley.EDU> Cc: Michael Smith <msmith@atrad.adelaide.edu.au>, freebsd-current@freebsd.org Subject: Re: version of makeinfo in -current Message-ID: <199605140251.UAA14151@rocky.sri.MT.net> In-Reply-To: <199605140012.RAA22477@paris.CS.Berkeley.EDU> References: <199605131155.VAA19647@genesis.atrad.adelaide.edu.au> <199605140012.RAA22477@paris.CS.Berkeley.EDU>
next in thread | previous in thread | raw e-mail | index | archive | help
> > > This seems to be a recurring thing for me to mail the lists, I'm > > > using FreeBSD as a development machine, and every piece of GNU software > > > I use I have to install from the current release, its kind of annoying. > > > > Lots of us use FreeBSD as development systems. > > Not many are using c++, from what I can tell. Hmm, I'm using C++ with no problems, but I tend to write --C++ code, where I don't take into account most of C++ new (mis)features. > > Lots of us have large investments in code that works correctly with the > > toolchain as it is. Including the system itself. > > > > If all we did was follow the often senseless faddism that seems to > > permeate the GNU world, all our time would be spent fiddling with tools, > > and no real work would ever be done. > > > > (This is why people just ignore the endless "when will you be upgrading > > to gcc 2.7.x" - there are bugs in 2.7.x that the FSF have said won't > > be fixed until 2.8.0, and the work and stress involved would be substantial.) > > This is silly, the same argument applies to the 2.6 -> 2.7 transition. > 2.8 will have bugs too, but for some of us, there are BIG differences, > especially when you try to use g++. As long as you don't rely on template support, the version we are using works for most everything. Templates are broken in 2.6, but there are as much if not more brokeness in the C optimizer/generator in 2.7 as compared to 2.6, so we're sticking with the 'known' quantity vs. the unknown quantity. > The same argument applies for the latest versions of diff, RCS (these two > are already upgraded), binutils, etc. The reason new versions get released > is because there are new features and/or bugs fixed. And the versions of RCS and diff were upgraded (sort as well) becuase they were bug-fix only releases. Gcc 2.7 was *far* from a bug-fix only release, so there was no reason to upgrade to it as far as stability goes. > If there wasn't, then there would be no reason for new versions, would > there? The reason for gcc 2.7 was for a rash of new features, but with very few bug-fixes. Gcc 2.7 needed at least one more minor upgrade before it could be considered stable, but the FSF decided against it for some reason. > who use these utilities find those bugs, and each time they install a > new FreeBSD system, they say to themselves, damn, now I have to > download 14 packages off of prep.ai.mit.edu and compile the latest > version. Why? Name the 14 packages you need to install from prep? Name more than 5 you need? (I can name 2, gcc and libg++) Nate
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199605140251.UAA14151>