summaryrefslogtreecommitdiffstats
path: root/usr.sbin/sysmerge (follow)
Commit message (Collapse)AuthorAgeFilesLines
...
* /var/db/sysmerge -> /usr/share/sysmergeajacoutot2014-07-152-11/+11
| | | | ok deraadt@
* Also ignore examplessum.ajacoutot2014-07-132-4/+5
|
* egsum -> examplessumajacoutot2014-07-121-2/+2
|
* Extend output a bit.ajacoutot2014-07-122-5/+7
|
* Use sha256(1) for compared files, just like we do for examples.ajacoutot2014-07-121-7/+6
|
* Initial support for /etc/examples.ajacoutot2014-07-122-8/+49
|
* /var/games/tetris.scores will move away from the etc set.ajacoutot2014-07-102-5/+3
|
* restore_sum: remove newly created sum file(s).ajacoutot2014-03-271-1/+2
|
* login.conf(5) is not special anymore.ajacoutot2014-03-262-5/+4
|
* Rework the way we fetch and verify sets to simplify the process.ajacoutot2014-03-241-68/+81
| | | | | | | | | | | | | Don't fetch SHA256 twice when we are updating from the same repo. Better output. Simplify extract_set (becomes extract_sets and deal with all sets at once). Make sure we cannot pass -s xetcXX and/or -x etcXX. Bring consistency in condition evaluations. Drop some unused variables. with inputs from rpe@
* Fix trap: do not remove the entire WRKDIR -- we may have our backed up filesajacoutot2014-03-221-2/+2
| | | | in there.
* Also remove TEMPROOT in error_rm_wrkdir.ajacoutot2014-03-221-1/+2
|
* Make sure to add missing users/groups _before_ running theajacoutot2014-03-221-60/+53
| | | | | | | distribution-etc-root-var target; otherwise mtree(8) can fail. issue reported and debugged by deraadt@ ok sthen@
* Use ftp -D.ajacoutot2014-03-211-3/+3
|
* Fix small regression when not using a full path to sets.ajacoutot2014-03-211-4/+4
|
* Cope with recent signify(1) changes and simplify getting/checking the sets.ajacoutot2014-03-201-43/+16
| | | | ok sthen@ rpe@
* Drop the check for badly formed *etc sets when we check signatures; it savesajacoutot2014-03-181-3/+4
| | | | | us time and if the sets are signed we can assume they are legit (yes, I am an assumer).
* Always print the key signify(1) is using.ajacoutot2014-03-061-2/+2
| | | | sure deraadt@
* start using the new officially named keys. This is being done beforederaadt2014-03-011-3/+3
| | | | | | any snapshots come out which are signed with that key... but just give it a day or two. from ajacoutot
* Only create DBDIR if it does not exist.ajacoutot2014-02-081-2/+5
|
* When using SM_PATH, fetch the SHA256 file only once since we will not beajacoutot2014-02-071-7/+14
| | | | | | using multiple repos. ok rpe@
* Sync getopts with usage.ajacoutot2014-02-051-2/+2
|
* Really clean up after a run.ajacoutot2014-02-031-10/+14
| | | | | | Factorize the sig file name. Do not through an error when a file got manually removed but is still present in the sum file.
* Drop comma (sthen@).ajacoutot2014-01-281-2/+2
|
* Drop FETCH_CMD and always use ftp(1).ajacoutot2014-01-282-32/+24
| | | | | | | | New fetch() function that wraps ftp(1) with our needed knobs (courtesy of halex@) input from halex@ rpe@ ok halex@
* Only remove COMPFILE if it exists.ajacoutot2014-01-271-2/+3
|
* Move "EDIT" to the top.ajacoutot2014-01-271-7/+7
|
* Enhance errors output and display exactly what is going on.ajacoutot2014-01-271-29/+36
| | | | | | | req. by deraadt@ Don't redirect error messages unless in a couple of places where it actually makes sense.
* Make it clear SM_PATH will fetch the sets corresponding to the currentlyajacoutot2014-01-271-4/+4
| | | | running release.
* specifically NOT suppress signify error outputrpe2014-01-231-2/+2
| | | | | suggested by espie@ OK deraadt@
* More output enhancements, put errors and warnings where they belong.ajacoutot2014-01-231-27/+23
|
* Verify both the signature and the checksum in check_sig().ajacoutot2014-01-231-19/+18
| | | | | | | Fix a regression with '-s' and '-x' when not using full path to the sets. Be a little bit more verbose about what we do. ok rpe@
* Do not bother comparing/merging empty files. Regression from previousajacoutot2014-01-221-2/+9
| | | | spotted by kettenis@
* Don't remove the SHA256 file twice.ajacoutot2014-01-221-2/+2
|
* Cleanup and simplify:ajacoutot2014-01-221-39/+61
| | | | | | | | | | | | | | | - rename restore_bak to restore_sum since this is what it does - better clean up after ourselves so only user relevant files are left in the work directory after a run - better handling of symlinks and make sure they match - tweak output: no need to warn the user when we are finished, it is pretty obvious; output that we are veryfing sigs - simplify signature handling: drop get_sig and get signatures at the same time we get the sets -- this allows to get a specific SHA256.sig file for each set check signature in check_sig() right after getting the corresponding set ok rpe@
* Document -S and the new signify(1) use by sysmerge(8).ajacoutot2014-01-211-4/+12
| | | | input and ok jmc@
* Protect variable names.ajacoutot2014-01-211-5/+5
|
* Add and enable signature checking per default.rpe2014-01-211-7/+30
| | | | | | | | Provide new option -S to skip that step. idea and initial work from jturner@ with feedback from deraadt@ ok aja@
* Add support for the SM_PATH environment variable. This changes theajacoutot2013-09-102-7/+22
| | | | | | | | | | | | | | default behavior of sysmerge when no source is specified on the command line: * if SM_PATH is set, then sysmerge will search this path for etcXX.tgz and/or xetcXX.tgz * if SM_PATH is not set (the default), then the behaviour is the same as before and /usr/src is used idea from espie@ man page refactoring jmc@ (who else :-)) "good idea" krw@ ok rpe@ landry@ espie@
* use .Mt for email addresses; from Jan Stary <hans at stare dot cz>; ok jmc@schwarze2013-07-161-4/+4
|
* Do not register symlinks in the sum files. This makes no sense and makesajacoutot2013-03-221-5/+6
| | | | | | | sysmerge(8) output a warning when the target of the link does not yes exist on the system. ok sthen@
* Always compare aliases(5) at the end of the sysmerge(8) run after allajacoutot2013-03-021-5/+7
| | | | | | | | files have already been merged. This is needed in the case where the syntax of smtpd.conf changed to prevent newaliases(8) from failing. issue reported by todd@ ok sthen@
* Rename some functions and bump copyright.ajacoutot2013-01-251-19/+19
| | | | No functionnal change.
* simplify chroot casesrpe2013-01-041-18/+4
| | | | ok ajacoutot@
* Refactor code for extracting and checksumming of (x)etcXX.tgz setsrpe2013-01-031-37/+46
| | | | | | | | | or preparing from /usr/src into extract_set() and prepare_src(). - extend error handling and quoting of user supplied input - move and rename get_sets() to get_set(), change parameter order and allow arbitrary filenames in url pattern (see r1.62 msg) ok halex@, 'go ahead' ajacoutot@
* correction of the SRCDIR checkrpe2012-12-261-2/+2
|
* Combine almost identical code for fetching and checking of therpe2012-12-261-27/+19
| | | | | | | etcXX.tgz/xetcXX.tgz sets into a more generic function get_sets(). Move the SRCDIR check into the getopts while-loop as well. ok ajacoutot@ halex@
* replace one grep with awk patterns and simplify another using -qrpe2012-12-251-3/+3
| | | | ok ajacoutot@
* simplify things by using shell patternsrpe2012-12-241-13/+10
| | | | ok ajacoutot@ halex@
* simplify setting FETCH_CMD and move it upwards to where the otherrpe2012-12-231-6/+2
| | | | | | global variables are defined ok ajacoutot@
Copyright © 1996 – 2023 Jason A. Donenfeld. All Rights Reverse Engineered.