Age | Commit message (Collapse) | Author | |
---|---|---|---|
2014-08-10 | zap trailing whitespace | Jasper Lievisse Adriaanse | |
2014-08-10 | comment what's going on | Marc Espie | |
2014-08-10 | Ouch... fuck cvs | Marc Espie | |
2014-08-10 | if pkg_create is run as non-root, restory correct group/owner to root/bin. | Marc Espie | |
also, remove write permissions without explicit modes. Allows fake installs to keep directories/files writable while producing correct package. | |||
2014-07-30 | prepare for post 5.6 packages, recognize special case where timestamp exist. | Marc Espie | |
(specifically, this is a nop for 5.6, but it will allow changes to packages without needing anything in pkg_add) | |||
2014-07-27 | make sure makewhatis shows error messages without unsightly CODE() refs | Marc Espie | |
2014-07-27 | - log libraries in a proper way | Marc Espie | |
- mark wanted libraries, then run the db check. Gives you a way to figure out when locate db are behind wrt pkgs | |||
2014-07-27 | ensure log stuff happens, even without context | Marc Espie | |
2014-07-23 | document format changes, including the new @ts annotation that pkg_create | Marc Espie | |
will generate after 5.6. | |||
2014-07-23 | move misleading comment | Marc Espie | |
2014-07-19 | actually use the cve match_location correctly. | Marc Espie | |
problem noticed by aja@ | |||
2014-07-18 | fix sorted output | Marc Espie | |
2014-07-14 | unhardcode db paths. | Marc Espie | |
put pruning of paths back early: these happen to be in the source db for the most part, so if we auto-detect source first, we lose. | |||
2014-07-13 | bye bye src url. Never amounted to much, we went another road a while ago | Marc Espie | |
2014-07-13 | bundling base + packages was slow, as it needed to find lots of files. | Marc Espie | |
different approach: mark the base system in one go, then hand pick the files from the pkglocatedb. quite a few unaccounted files yet, this will get better | |||
2014-07-13 | tedu setlist support, superseded by the new locate dbs. | Marc Espie | |
2014-07-13 | stuff no longer needed, now that base/x release install its own locate dbs | Marc Espie | |
2014-07-12 | don't check_security on newer *after* making sure the set doesn't have | Marc Espie | |
errors. Note that the errors are already signaling enough (conflicts, bad package etc)... And not having any errors mean we actually have proper plists. problem noticed by zhuk@ | |||
2014-07-12 | zap trailing whitespace; | Jason McIntyre | |
2014-07-12 | oh well, remove -Q entirely... I'm probably the only one who ever used that. | Marc Espie | |
2014-07-12 | security check in quirks | Marc Espie | |
2014-07-12 | undocument -Q | Marc Espie | |
2014-07-12 | finish killing old experiment, less confusing code | Marc Espie | |
2014-07-12 | @endfake died a while ago | Marc Espie | |
2014-07-12 | so, pkg_check can now troll thru the whole file system. Either use | Marc Espie | |
a full pkglocatedb, or the system locate dbs... | |||
2014-07-11 | use the merged 4.4BSD.dist mtree if BSD.local is gone | Marc Espie | |
2014-07-11 | display the extraction tally only if verbose | Marc Espie | |
2014-07-11 | with out-of-order extraction, match tempnames to item names | Marc Espie | |
2014-07-10 | make quirks slightly more special, grab the whole packing-list always, | Marc Espie | |
and display the timestamp from the signature if we were signed. | |||
2014-07-10 | prevent users from adding timestamp themselves | Marc Espie | |
2014-07-10 | add a separate timestamp annotation to have this as metadata | Marc Espie | |
2014-07-08 | don't really bother cleaning up the callback, it will go away soon enough | Marc Espie | |
2014-07-08 | in case we say("Some parameter #1 blabla there") but we | Marc Espie | |
forget the parameter, display a decent error message instead of the usual warning... | |||
2014-07-08 | Prettify the list of "required by" package names by sorting it. | Vadim Zhukov | |
okay espie@ | |||
2014-07-08 | cleaner callback handling | Marc Espie | |
- store it inside location, so that belayed archives open can get it - wipe it out when the sizer goes away... to be done slightly better probably ? | |||
2014-07-07 | sometimes, the callback happens when the archive isn't opened... | Marc Espie | |
fix it quickly, to be fixed better later. | |||
2014-07-07 | don't throw strange warnings during pkg_delete... | Marc Espie | |
2014-07-07 | display saved size for now, to be disabled once I'm satisfied this is very | Marc Espie | |
good :) | |||
2014-07-07 | record how many bytes we saved for statistics purposes. | Marc Espie | |
exit the loop thru the same end code every time for consistency. | |||
2014-07-07 | restore progressmeter for big files. | Marc Espie | |
somehow, one extra level of indirection to archive, make it all go away by adding a new method set_callback | |||
2014-07-01 | activate "out-of-order" archives based on history | Marc Espie | |
things to watch out for: - correct handling of multiple files with the same sha - cwd handling. This is already destated, so we can call set_destdir on each entry's cwd, instead of relying on @cwd being in the "right" location as that's no longer the case... | |||
2014-07-01 | turn out-of-order extraction back on. | Marc Espie | |
- separate extraction into extract/tie methods, with slightly overlapping code - early exit when we only have tied entries - mark location as early_exit in that case, since we're gonna close the archive early | |||
2014-07-01 | tweak the ProgressMeter code to always record sizes, even in the stub case | Marc Espie | |
2014-06-25 | revert for now. Some aspects of tied code are currently broken in a fun | Marc Espie | |
way. surprised nobody noticed... | |||
2014-06-17 | oops, committed debug version, to the tied hash was not used at all, | Marc Espie | |
and of course there were 3 obvious errors... fixed and retested. much better now | |||
2014-06-16 | tweak extraction for out-of-order archives: do a first pass | Marc Espie | |
where we register every actual file in the plist by its path relative to cwd. Note every meta info as extracted (for partial packages to make sense), then do a newer extract loop, where each archive member must match by path. That way, we can close archives early as soon as we've extracted every new file. To be used shortly by pkg_create. Note that new packages *won't* be extractible by the old tools. | |||
2014-06-16 | tweak the size_and support to be able to do stuff out of order | Marc Espie | |
2014-06-16 | fix bug introduced by work-around: sets with nothing to do must be marked | Marc Espie | |
uptodate as well. Restores ntogo to working order | |||
2014-06-15 | Correct precedence for returned value. | Andrew Fresh | |
style suggested and ok espie@ | |||
2014-06-12 | display full url if we can for UNSIGNED PACKAGES. | Marc Espie | |
unconfuse mikeb@ :) |