summaryrefslogtreecommitdiff
path: root/usr.sbin
AgeCommit message (Collapse)Author
2009-12-31when separating command from parameters in smtp session, the parser triesGilles Chehade
to use ':' as a separator then fallbacks to ' ' so that it can detect the command names that contain more than one words (MAIL FROM and RCPT TO) or the one word ones (HELO, DATA, ...). this is incorrect and the parser can get confused if the parameter to any command contains a ':', for example "HELO [ipv6:...]" cause the parser to lookup for command "HELO [ipv6". fix this by using ':' as a delimiter for 'mail from' and 'rcpt to', while using ' ' as a delimiter for all other commands. fixes bug 6285/system reported by Lionel Le Folgoc <lionel@lefolgoc.net>
2009-12-31Instead of passing the config via arguments to the childs on bootup issueClaudio Jeker
a config reload as first step in bootup. This allows childs to start with an empty config and a lot of special cleanup code can bite the dust. Testing by myself and sthen@ with a few configs (more testing welcome). Seems like a good idea henning@ & sthen@
2009-12-31stupid, stupid, stupidMarc Espie
2009-12-31need to tweak from to_update to self, as this kind of thing happens allMarc Espie
the time.
2009-12-31oops, actually report issuesMarc Espie
2009-12-31fix for error message logging.Marc Espie
2009-12-31always create a problem entry for libraries we don't findMarc Espie
2009-12-31complain if we see weird @libs (so, not in system).Marc Espie
optimization: for system libs, only parse stuff with .so in the name
2009-12-31add meat to the "can't install" stupid handle messageMarc Espie
2009-12-31XXX until I fix it for realMarc Espie
2009-12-31more/better cache for bad cases, and trap self after merge.Marc Espie
2009-12-31sha256 is the common case, no need to delay loading.Marc Espie
2009-12-30don't bother clearing the line, since we do that automatically when weMarc Espie
write new info. This lets messages persist a bit longer.
2009-12-30more changes and optimizations: create unique dewey objects, so I canMarc Espie
compare them directly and avoid storing the string. on typical use through a large update, 2/3 cache hits. also, remove switch on <,> by writing subclasses. Actually makes things clearer.
2009-12-30... and this should fix todd's case, that of not finding anything andMarc Espie
resorting to default...
2009-12-30also track stuff I couldn't install, and use it for negative cache lookups.Marc Espie
fix some looping if stuff can't install for whatever reason.
2009-12-30fix pkgspec handling, actually do filter by regexp.Marc Espie
shave some memory off packagenames by not storing non-existent p/v. cache all packagename construction (happens two times out of 3).
2009-12-30Correctly parse the host address in the FEC TLV.Michele Marchetto
ok claudio@
2009-12-30store url where we retrieved the package.Marc Espie
2009-12-30ouch, now that we keep packages, take them into account on done.Marc Espie
2009-12-30a few small optimizations:Marc Espie
- be more specific in what we trim. - iterate over list directly, don't build it. - don't create pattern objects for the most common case (one subpattern)
2009-12-30take merging of sets into account when creating a header.Marc Espie
2009-12-30infrastructure for url elementsMarc Espie
2009-12-30actually cache architecture, as intendedMarc Espie
2009-12-30prettyMarc Espie
2009-12-29do "not reinstalling" in a more intelligent way, by moving things intoMarc Espie
kept, and then, just nop'ing a set with nothing to do...
2009-12-29temporary fix for .libs update issue.Marc Espie
2009-12-29a few more tweaks:Marc Espie
- allow for merged updates if we want them - deplist becomes a hash, maybe useful for big sets - simplify loop checking. We don't actually care whether we find any. - be more aggressive in keeping cached stuff, create solver or conflict_cache if need be.
2009-12-29the way Tracker works, installed packages are always marked uptodate as well,Marc Espie
so remove some unneeded checks from solver's perspective.
2009-12-29ouch, got slightly longer, some stuff to fix soonishMarc Espie
2009-12-29for now, show cache statistic if -F stat_cache (still in need of improvements)Marc Espie
2009-12-29simplify solver handling, rely on cache to avoid recomputing stuffMarc Espie
(large performance gain).
2009-12-29rework dependency solving around cached items, so that redoing the sameMarc Espie
lookup will be faster
2009-12-29performance gain: cache conflicts.Marc Espie
2009-12-29prepare to cache conflicts in sets: object to do so, code to merge conflictsMarc Espie
and solver, actual auto-gen of conflict cache and solver.
2009-12-29improved check for loops:Marc Espie
- detect as many loops as we can - don't merge sets twice
2009-12-29move check for already installed UP, which allows me to deal with UpdateSetMarc Espie
and tracker globally to avoid loops.
2009-12-29prepare for smarter sets, with cache infoMarc Espie
2009-12-29replace a few remaining UNIX terms with .Ux macros; this diffIgor Sobrado
completes the changes from Oct 22.
2009-12-28kill old dependency adjusterMarc Espie
2009-12-28use new forward dep adjust.Marc Espie
2009-12-28new method for adjusting forward dependencies, much simpler.Marc Espie
2009-12-28erase Requiring entirely while deleting, to avoid further issues.Marc Espie
if size_only, don't even pretend to delete the list (ways faster, no checksum checks)
2009-12-28make objects chainable.Marc Espie
Provide a new "erase" method that removes everything
2009-12-28an actual difference between -r and -u. If we replace, don't look as hardMarc Espie
for conflict matching.
2009-12-28prepare to handle forward dependencies in a simpler wayMarc Espie
2009-12-28add -F carp as an extra debugging device: ^T handler will show a stackMarc Espie
backtrace
2009-12-28move forward dependency checking to very late, so we can reuse the result.Marc Espie
uncache conflict checking, since it might need to be redone now...
2009-12-28move code around a bit, more functions. Especially do the exec checkMarc Espie
ways later, when we're certain we're not going to change the set any longer, so we don't need to go back.
2009-12-28nitsMarc Espie