On Thu, 19 Feb 2009 13:06:01 +0300 Peter Volkov wrote: > В Чтв, 12/02/2009 в 18:03 +0000, Ciaran McCreesh пишет: > > GLEP 55 *is* the good, workable solution. There still haven't been > > legitimate any technical objections to it. > > This issue was discussed already... [1] ...and there weren't any legitimate technical objections. > If you and think that EAPI is meta-information then it should not be > inside file name and then it's possible to parse ebuild and get EAPI > from some defined-format line. Performance penalties can be mitigated > by some new caching (you know better than me that it's good idea to > re-implement caching in any case). The only thing that can parse ebuilds is bash, and it can only do that once it already knows the EAPI. Another cache won't solve anything since there's no way to generate that cache to begin with -- and a second level of cache would slow things down, not speed them up. > We are already discussing this feature more than year. Many people > voiced concerns about .eapi extension so why don't you try fix same > issue differently? Yes it's harder and probably longer but I don't > believe it's impossible. Because all the alternatives are worse, and none of the objections to the extension have been technical in nature. They've all been "we don't want you to apply anti-mould paint to the rotting bikeshed because it's only available in brown". -- Ciaran McCreesh