On Thursday 07 July 2005 22:42, Robin H. Johnson wrote: > On Thu, Jul 07, 2005 at 10:52:06AM -0700, Donnie Berkholz wrote: > > > strlen(EBUILD_FORMAT) * 19546 = 249K > > > strlen(EAPI) * 19546 = 77K > > > strlen(EV) * 19546 = 39K > > > > > > Where 19546 is the number if ebuilds in the tree as. > > > > So, the size of the tree is 500216K. This makes the percent increase > > in size for the options above 0.0005%, 0.0002% and 0.0001% > > respectively. In any case, nearly irrelevant to its present size. > > Probably should count some more formatting too. > strlen(EBUILD_FORMAT="") = 16, 16 * 19546 = 306K > strlen(EAPI="") = 7, 7 * 19546 = 134K > strlen(EV="") = 5, 5 * 19546 = 96K > > Don't just count the 4K blocks. Count the actual data instead (du -b). > The tree contains 95912503 bytes of data, including metadata. > Excluding metadata it is 85301916 bytes. > > It's still less than 0.4% increase in the worst case. > > I like the 'EAPI' variant, as it's clearly seperate from 'EV', without > being too long. > > To combat some of the increase, EAPI only needs to go into new ebuilds > that use the functionality. Old ebuilds should default to the current > syntax.
To allow for this to work with current portage versions, perhaps it would be an option to introduce a new extension for .ebuild scripts that use it's functionality. That would allow all non-EAPI aware portage versions to automatically ignore ebuilds that use this. Paul ps. I would also suggest requiring that EAPI can be retrieved by a simple line by line parsing without using bash. (This allows for changing the parsing system) -- Paul de Vrieze Gentoo Developer Mail: [EMAIL PROTECTED] Homepage: http://www.devrieze.net
pgpp5IGEAQkhd.pgp
Description: PGP signature