JIm
On Thu, 2005-01-06 at 11:35 -0700, Dan Gregory wrote:
Having read all the esearch posts, it seems to me that what portage really needs is a register/callback mechanism that would allow a package (such as esearch) to register a callback with another package (such as rsync) that gets executed when rsync is rerun/reinstalled/recompiled/etc. In this example, you could just run the command to update the esearch database whenever rsync is rerun. And this could be generically iimplemented so that anyone that has a non-rebuild dependency on another package could use. This solution keeps the nay-sayers happy by not tying esearch directly to rsync and allows other variants of esearch (I don't know of any or I'd give examples) to also be used, or if you don't care about speed searching you don't install it and it works just like it does now. It also solves the what database should we use question because you could have a different esearch for each architecture. Dan Alec wrote: > Luke-Jr wrote: > >> On Sunday 02 January 2005 5:52 pm, Ciaran McCreesh wrote: >> >> >>> On Sun, 2 Jan 2005 17:46:59 +0000 Luke-Jr <[EMAIL PROTECTED]> wrote: >>> | On Sunday 02 January 2005 5:33 pm, Ciaran McCreesh wrote: >>> | > On Sun, 2 Jan 2005 17:24:34 +0000 Luke-Jr <[EMAIL PROTECTED]> >>> | > >>> | > wrote: >>> | > | On Sunday 02 January 2005 4:26 pm, Fabian Zeindl wrote: >>> | > | > Will esearch be included in emerge some time? >>> | > | >>> | > | A slightly better question: Will an esearch database ever be >>> | > | included w/ portage syncs like metadata is? I've got my private >>> | > | rsync distributing it and it works quite well. >>> | > >>> | > Why? If you want to distribute an esearch db, do it separately from >>> | > sync. It's a whole different kettle of fish. Stick it on the web >>> | > somewhere... >>> | >>> | How so? The database is directly based on the portage tree. Keeping >>> | them together makes sense. >>> >>> But it has nothing to do with portage. It's an entirely different >>> project which happens to use the portage tree. Are you suggesting that >>> we distribute data for every single package which happens to use the >>> portage tree in an emerge sync? >>> >> >> >> If it doesn't take much room overall... why not? >> I'm not aware of any besides esearch, though. And you can't say it has >> nothing to do with portage when it is solely a portage tool. That's >> like saying gentoolkit has nothing to do with portage... >> >> > It's not an Official portage tool. Gentoo doesn't maintain it, so > why should they pay for the bandwidth to distribute the metadata with > it? Especially when you can just run a cronjob to update the metadata > for yourself? > -- [EMAIL PROTECTED] mailing list
