summaryrefslogtreecommitdiff
path: root/usr.sbin/pkg_add
AgeCommit message (Collapse)Author
2015-08-14remove partial to_register from sets when we merge, since we have to do theMarc Espie
work again. Should solve the +REQUIRED_BY problem hopefully.
2015-08-13scaffolding that will display more info for the elusive REQUIRED_BY errorMarc Espie
2015-08-13let pkg_create recreate all packages correctly, again.Marc Espie
2015-07-28make -B with bad checksums cope. Reported by millert@Marc Espie
2015-07-23change the expansion of %m and %c in installpath during -beta to theJoerg Jung
snapshots/ folder ok phessler@ sthen@
2015-07-21switch dpb from sudo to doas (and env as required); ok sthen@Christian Weisgerber
2015-07-12improve the error message for incorrect PKG_CACHE.Marc Espie
(yes, this should error out always)
2015-07-09gc old code, from Jean-Philippe OuelletMarc Espie
2015-06-30cosmetic comma is not really necessary, better to have a raw list of ↵Marc Espie
packages. Pointed out by theo.
2015-06-25use machine_architecture (arch -s) when assembling PKG_PATHs using %m or %a,Stuart Henderson
fixes these variables with arch like powerpc where this differs from "uname -m". reported by/ok aja@
2015-06-06Add rtwn(4) to fw_update. ok espie@, czarkoff@ suggested a similar diffStefan Sperling
2015-05-25allow pkg_add as nonroot to soft-fail when outside of local base.Marc Espie
2015-05-24pass subst to the installer state, so that -Dunsigned would workMarc Espie
2015-05-18identical common code -> refactorMarc Espie
2015-05-18better error in case we can't create tempfilesMarc Espie
2015-05-18simplify progressmeter calls to visit_and: we retrieve the state from theMarc Espie
progressmeter object
2015-05-18have visit_with_count and visit_with_size use the same parameter conventions.Marc Espie
In particular, pass state before the other method parameters.
2015-05-18since the progressmeter is always tied to a state, store the state always,Marc Espie
and not just in the Term case. allows passing less parameters around.
2015-05-15add missing placeholder so that the "UNSIGNED PACKAGES: " line actuallyJasper Lievisse Adriaanse
contains the list of packages. ok espie@
2015-04-20totally forgot about the !uid notation in newuser when adding the earlyMarc Espie
cache for pkg_create. Reminded by sthen@
2015-04-16firmware, not firmwares;Jason McIntyre
2015-04-16tweak previous;Jason McIntyre
2015-04-16document %m, sort %.Marc Espie
2015-04-16add %m as a shorthand, will expand toMarc Espie
pub/OpenBSD/5.7/packages/amd64 for lazy typers. okay aja@
2015-04-16document % sequences.Marc Espie
2015-04-16reorg code, the arch/osversion code should live in a single place,Marc Espie
short and sweet
2015-04-16commit expanded tags for %c, %v, %aMarc Espie
2015-04-08s/inexistent/nonexistent/Todd C. Miller
2015-04-06don't check this twice.Marc Espie
2015-04-06make pkg_info run slightly slower on distant packages, but do check thatMarc Espie
the info is signed and checked. after remarks from tedu@...
2015-04-06mark installed locations as "trusted" so that pkg_info does not check sigsMarc Espie
on them.
2015-04-06move signer_list up so pkg_info can use itMarc Espie
2015-03-29make sure specialfiles get checked. Actual bug, got lost when moving toMarc Espie
extractible/tied files. Could have been mentioned earlier, the fix is obvious.
2015-03-26Backout -DSHORTENED semantics by default for now, because it causedMatthias Kilian
too much problems with the recent changes to net/libnet (in combination with security/dsniff); the PKGSPEC changes. Some more suspicious pudate problems mentioned by landry@. ok sthen@
2015-03-09make -DSHORTENED the default: don't update unless there is an actualMarc Espie
change in dependencies: shared library versions, and explicit version requirements in dependencies. as discussed with aja@ before the lock.
2015-03-04better error reports if PKG_TMPDIR is not writable, okay millert@Marc Espie
2015-02-25Point people at rcctl(8) rather than rc.d(8) in the text displayed whenStuart Henderson
a package has installed an rcscript - it's a good UI, and people needing further details can follow the "see also". ok espie@ brynet@
2015-02-15-i option, so that even pkg_info is not needed.Marc Espie
make the arguments a bit more free-form, so that both radeondrm and radeondrm-firmware work.
2015-02-09oops, this little bit is unneeded and actually breaks some other stuffMarc Espie
2015-02-09full fix for check-lib-depends... previous patch did only half the workMarc Espie
2015-02-06add iwm to the possible_drivers listStuart Henderson
2015-02-05compat for check-lib-dependsMarc Espie
2015-02-03fix odd error recovery (noticed by kili@): fw_update -d with !installedMarc Espie
driver should just say it's not installed and not try to install it.
2015-02-03zap scp urls from quirks update for fw_updateMarc Espie
2015-02-03document the one non-intuitive test there, as I spent 5 mn looking at itMarc Espie
wondering... :)
2015-02-03Fix -aMarc Espie
2015-01-30Fix the default interactive level to install the "default" dependencies.Marc Espie
Actually testing it caused me to write correct (and simpler) code...
2015-01-27fix messages of FwUpdateMarc Espie
2015-01-19add a specialized version of version compares that should not updateMarc Espie
everything all the time, at the bequest of aja. (basically, we updated everything when runtime dependencies changed, even though we actually only care for wantlib, and hard dependencies, but those trigger bumps all over the tree) Experimental for now, turn on with -DSHORTENED. Likely to become the new default if nothing bad comes out of it.
2015-01-17make sure we have an interactivestub... I forgot the -I part from the oldMarc Espie
script.
generated by cgit v1.2.3 (git 2.25.1) at 2024-11-28 06:33:08 +0000