Age | Commit message (Collapse) | Author | |
---|---|---|---|
2010-03-23 | fix callback. This one isn't called that often actually. | Marc Espie | |
noticed by ckuethe | |||
2010-03-22 | clean up the ProgressMeter code, create a specific class if a Term | Marc Espie | |
is around. start using stuff from termcap, specifically, cleareol, and move to hp. use the full line when we can, cram 100% if we can. (this should flicker much less on slow terminals). clean-up the install-and-progress bits to be in progressmeter, this removes some code. have pkg_create follow the same pattern as the others, with a state. This allows us to unify error messages a bit. | |||
2010-03-20 | if a package already contains an url marker (say, local stuff recreated), | Marc Espie | |
don't override it. so that it can actually be installed. | |||
2010-03-07 | erase to the end of "previous line". Avoid dangling % after displaying 100%. | Marc Espie | |
from sthen@. | |||
2010-03-06 | match_locations doesn't work on invalid specs, so don't go there | Marc Espie | |
2010-03-06 | fix very stupid typo | Marc Espie | |
2010-03-06 | don't go up to column 80, until we ask Term::Cap properly for what's going | Marc Espie | |
on. | |||
2010-03-06 | consistent error. exit(1) if something bad happened. | Marc Espie | |
(fixes some errors in ports) | |||
2010-03-01 | slightly terser messages, better use of terminal line | Marc Espie | |
2010-02-28 | progress-meter -> progress meter, which is at least consistent with our | Jason McIntyre | |
other pages; | |||
2010-02-28 | ProgressMeter support. | Marc Espie | |
2010-02-28 | more useful paths (for dpb) | Marc Espie | |
2010-02-24 | revert rev. 1.54 because it doesn't change the rendering with groff | Ingo Schwarze | |
but breaks the build with mandoc(1); ok jmc@ espie@ | |||
2010-02-04 | no -S option; from Daniel Dickman | Jason McIntyre | |
2010-01-27 | actually order things: | Marc Espie | |
beta < (rc|pre) < release < pl (fixes speex updates) | |||
2010-01-26 | downgrade, not downgrades | Marc Espie | |
2010-01-26 | fix merging behavior when merging doesn't occur: | Marc Espie | |
- display message that says "this update didn't merge" - be more specific in return values, because we want to distinguish between we merged some stuff -> compute full depends / we didn't -> proceed thanks to Florent Tribouilloy for being my guinea pig ;) | |||
2010-01-24 | extra indirection: no match gives a reason why, so we can rebuild | Marc Espie | |
report_problems to actually say what's wrong... | |||
2010-01-24 | use libspecs, report problem will be terser for now. | Marc Espie | |
2010-01-24 | use libspecs through Dependencies. | Marc Espie | |
2010-01-24 | add code to unify spec and names as LibraryObject, plus a repository and | Marc Espie | |
match/lookup functions. | |||
2010-01-24 | prevent broken packagenames from creeping in. | Marc Espie | |
2010-01-24 | add code to check packagename for problems | Marc Espie | |
2010-01-19 | simplify caching: return $cached->{$_} //= new_value; | Marc Espie | |
remove stats cache LibSpec as well (1000 c.53.0 !) | |||
2010-01-19 | simplify using LibSpec | Marc Espie | |
2010-01-19 | move generic code to handle libspec in a new file, to clean up stuff | Marc Espie | |
2010-01-19 | fix handling of always-update: it should only come into play when the | Marc Espie | |
normal signature is equal. Otherwise, downgrade prohibition is still in effect. | |||
2010-01-18 | don't bother checking for exec and displaying them if we're not verbose | Marc Espie | |
and we have -D update on. | |||
2010-01-17 | even better messages | Marc Espie | |
2010-01-17 | use Carp directly, it it very inexpensive actually, since it's been rewritten | Marc Espie | |
to be just a very small stub that loads Carp::Heavy on demand. Also, stop wrapping messages in "Expected", doesn't give us anything. Finally, make RequiredBy calls be carp, so that we actually know where they fail... | |||
2010-01-14 | fix new vstat, and make it useful, for instance for -n. | Marc Espie | |
2010-01-12 | allow 'update-only' installs | Marc Espie | |
2010-01-12 | bump this to updater, so that progress is shown again. | Marc Espie | |
2010-01-12 | more auto-cache goodness | Marc Espie | |
2010-01-11 | sometimes, I manage to finish some stuff | Marc Espie | |
2010-01-11 | quirks is special here too, oops. | Marc Espie | |
2010-01-11 | allow installation to proceed even without -r if it's an actual update | Marc Espie | |
(all markers match, conflict and pkgpath, and the pkgname is newer if it's the same stem/flavor combo) | |||
2010-01-11 | works much better if I actually add the option... | Marc Espie | |
2010-01-11 | move the oldlibs code into its separate file, prior to fixing it. | Marc Espie | |
allow replacement if it only touches partial or libs | |||
2010-01-11 | protect gzip against weird file names | Marc Espie | |
2010-01-11 | various cleanup, including a fix from Daniel Dickman; | Jason McIntyre | |
ok espie | |||
2010-01-11 | It's '-D dontmerge' now. | Bernd Ahlers | |
2010-01-10 | fix some mdoc tidbits. Thanks to jmc@ for advice. | Marc Espie | |
2010-01-10 | unify options a bit: use -Dname=value for defines uniformously (e.g., deprecate | Marc Espie | |
-F keyword heavily). | |||
2010-01-10 | use full signatures to avoid downgrades. | Marc Espie | |
remove double check in update that properly belongs in pkg_add replace has_new_sig by has_different_sig, with better diagnostic messages. | |||
2010-01-10 | follow logic: if same packagename, but different signature elements, | Marc Espie | |
something very bad happened. | |||
2010-01-10 | trick: I can actually recognize specs easily, so why not allow them without | Marc Espie | |
-e ? | |||
2010-01-10 | if we make pkg_create more stringent, then we cannot create bogus | Marc Espie | |
packing-lists. Add check for libspec, and REGRESSION_TESTING | |||
2010-01-10 | bump copyright | Marc Espie | |
2010-01-10 | pull the signature comparison code into its own file. | Marc Espie | |
add comparison functions to Packagenames that return undef when it matters. |