[pacman-dev] [patch] for resolvedeps + some notes
Hi! With this patch resolvedeps will search for other satisfiers, if we don't let it add a pkg from IgnorePkg. Note: this patch assumes my depcheck.diff patch applied. Bye PS: Resolvedeps is still dummy, because pulling a new package to the *list may break an earlier verified dependency of a pkg in the *list (because the new package will overwrite something in the localdb!). The best (but much slower) solution is to call depcheck with the *list parameter (plus syncpkg).
On Sat, Aug 04, 2007 at 03:25:08PM +0200, Nagy Gabor wrote:
PS: Resolvedeps is still dummy, because pulling a new package to the *list may break an earlier verified dependency of a pkg in the *list (because the new package will overwrite something in the localdb!). The best (but much slower) solution is to call depcheck with the *list parameter (plus syncpkg).
I think it's always important to worry about getting clean and correct code first, and then worry about performance later (while trying to keep the first two properties ;)). In any cases, it would be interesting to have a pactest for this, and also possibly a fix, which will probably only causes a negligible slowdown anyway.
> I think it's always important to worry about getting clean and correct code
> first, and then worry about performance later (while trying to keep the
> first
> two properties ;)).
Hi!
Well, the current code is clean and bugfree, however it is dummy. It cannot
corrupt the db, because we do a checkdeps in sync_prepare before sync_commit, so
that will result a dependency error if needed:
1. This can easily happen, because resolvedeps doesn't handle the backward
(requiredby) dependencies...
2. Resolvedeps doesn't guarantee that all dependencies of the populated list is
satisfied neither: see my previous mail.
Bye, ngaba
PS: I also interested in speed tests where you implement the following psuedocode:
resolvedeps(list) {
while(missing = checkdeps(list)) resolve_misslist_by_adding_from_syncrepos(missing);
}
----------------------------------------------------
SZTE Egyetemi Könyvtár - http://www.bibl.u-szeged.hu
This mail sent through IMP: http://horde.org/imp/
On Mon, Aug 13, 2007 at 03:28:43PM +0200, Nagy Gabor wrote: > > I think it's always important to worry about getting clean and correct code > > first, and then worry about performance later (while trying to keep the > > first > > two properties ;)). > Hi! > Well, the current code is clean and bugfree, however it is dummy. It cannot > corrupt the db, because we do a checkdeps in sync_prepare before sync_commit, so > that will result a dependency error if needed: > 1. This can easily happen, because resolvedeps doesn't handle the backward > (requiredby) dependencies... Actually, I also wanted to ask about this. Why is it needed to remove the requiredby handling just for resolvedeps? If it isn't really needed, we could remove that requiredby arg of checkdeps.
Actually, I also wanted to ask about this. Why is it needed to remove the requiredby handling just for resolvedeps? If it isn't really needed, we could remove that requiredby arg of checkdeps. My patch: I wanted to keep the old behaviour. Originally: I don't know, probably the main reason is the speed-up; and that would(?) be strange if resolvedeps added a missing dependency of a locally installed package (it is designed to resolve the dependencies of one package in sync repo). Bye, ngaba
---------------------------------------------------- SZTE Egyetemi Könyvtár - http://www.bibl.u-szeged.hu This mail sent through IMP: http://horde.org/imp/
participants (2)
-
Nagy Gabor
-
Xavier