From: <dan...@ya...> - 2002-02-15 22:39:25
|
--- "Steve D. Perkins" <mai...@st...> wrote: > > The value of the loose cannon, from my perspective, is that it > facilitates > > the "release soon, release often" approach.... > > I apologize for the miscommunication. By "loose cannon", my gripe > was > directed at the naming convention rather than the particular package > content. I was simply saying that if there's going to be this drive > towards > a consistent naming convention with MinGW subpackages... do we want to > alter > the legacy subpackages to use this naming convention? > A standardised naming convention would be good. I don't think we should rename the existing subpackages since that would cause confusion. That should wait until mingw-1.2 release. But as a template, how would you see the current components being named? I foresee that when 3.1 is released, users will find windows-specific or w95 specific bugs that are not being tested in the gcc dejagnu testsuite or in my test bed. Thus, there will very probably need to be a series of betas or pre-releases before we want to put 3.1 into a new mingw base package. We will probably also want to retain the 2.95 branch of gcc since it is very stable and, in the case of iostream-dependent code, has much faster compile time. How do handle alternative gccs in 'base' distro? Danny http://greetings.yahoo.com.au - Yahoo! Greetings - Send your Valentines love online. |