cpet/openbsd-ports
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
114,518 Commits 1 Branch 0 Tags
Commit Graph

474 Commits

Author SHA1 Message Date
espie
5e63ab5e56 do not check signatures in there 2016-09-14 15:06:53 +00:00
espie
cb486f103d prepare for new sigs 2016-09-14 15:02:41 +00:00
ajacoutot
59b5e044ea Sync make-plist with our pkg archs. 
ok espie@
 2016-09-01 13:12:34 +00:00
espie
c061b0cca1 run dev_mkdb after MAKEDEV. 2016-06-26 22:10:03 +00:00
ajacoutot
19618db5d9 Remove references to SHARED_ONLY; it's unused nowadays. 
Not touching make-plist too much because espie@ already has local changes in that file.

ok espie@
 2016-06-22 10:13:17 +00:00
espie
a11d1f2d8e synch with base. put the call in an eval for now so it doesn't error out, 
will just fuck up display slightly if you don't have current base
 2016-06-16 18:17:44 +00:00
espie
0d5d727916 remove more PLIST_DB references 2016-06-10 12:52:57 +00:00
espie
e38f6715f1 keep going even after warnings during the copy process 
they happen so often...
 2016-06-06 22:14:20 +00:00
espie
4095c1c1c4 fix 2 issues, bogus next (patch by semarie@) and a more dire one: unpopulate 
should use accounted, not known.

Fixes snapshots + cleanup issue, ouchie. Issue found by semarie@ again.
 2016-06-04 17:22:11 +00:00
espie
b7e1532d12 allow not copying CVS subdirs in the ports tree (ignorecvs) 
set checkout_ports to 1 by default if portscvs is defined.
 2016-06-04 12:36:06 +00:00
espie
715a3799e3 let actions=none work as intended 2016-06-04 12:22:02 +00:00
espie
5a5989ff85 a few important changes: 
- run unpopulate_light by default;
- make an explicit list of preserved data in the manpage.
- don't preserve PORTSDIR if we copy_ports. Don't store extra stuff there.
- register fatal errors and exit(1).
- move a lot of things before unpopulate_light so that clean-up is more
accurate.
- recognize wxallowed. Those checks are a bit simplistic yet.
- clean-up old redundant code
 2016-06-04 11:58:36 +00:00
espie
e066008fd5 protect preserved stuff better during the actual copy 2016-05-28 08:42:23 +00:00
espie
7640ab447f fix a nasty update bug 2016-05-26 00:08:02 +00:00
espie
ffbbe7d386 completely kill PLIST_DB here. After all, it's new stuff, 
so people should be able to adapt quickly.
 2016-05-23 13:36:25 +00:00
espie
cc2f4890f5 keep a global cache of distfile checksum while scanning, allows us to avoid 
rescanning the same distfile 50 times...

also display elapsed time at end of build (cosmetic) since that's now trivial
 2016-05-21 12:20:10 +00:00
espie
252e212c37 switch to PLIST_REPOSITORY, keep PLIST_DB glue for a few days 2016-05-20 09:55:25 +00:00
espie
3973d1b6f1 make it possible to directly add lines at end of generated mk.conf. 
make option handling more obvious, no actual change.
 2016-05-19 13:40:27 +00:00
espie
31775ef126 typo in conversion 2016-05-17 16:23:02 +00:00
espie
1d55b5f977 PORT_USER is required for checkout_ports (cryptic error noticed by juanfra) 2016-05-17 16:16:17 +00:00
espie
b4e5bf6a64 tweaks to heuristics 
1/ do subdirs on the command line first
2/ ignore errors from "preferred" dependencies (and don't register anything)
if there are actual errors beyond paths that no longer exist, we'll get them
during the next pass.
 2016-05-17 14:50:36 +00:00
espie
4c52369cd9 fix (for now) 2016-05-17 13:14:45 +00:00
espie
3388fc4cfd change drastically the way LISTING starts up: let dependencies be a permanent 
log.

Now that we can SKIPDIR, doing a first LISTING for most important dependencies
is not a trade-off for the full listing.

It allows dpb to start building most important stuff right away, so that
your machines start earlier.

Caveat: if that list of dependencies changes (ports renames), then you will
see errors because ports no longer exist.

I figure I have time to make errors in the first listing "not relevant"
before that becomes an actual issue.
 2016-05-16 13:47:18 +00:00
espie
2495ed5e36 use skipped dirs for initial build 2016-05-16 10:37:44 +00:00
espie
c5dddced40 better days display. somehow I thought perl would work like C and do 
integer division.
 2016-05-16 09:50:59 +00:00
jca
2b6787982d Use cmp -s and grep -q, instead of redirecting to /dev/null 
ok espie@
 2016-05-15 22:43:18 +00:00
rpe
fa8d7a6eb0 fix typo 
OK espie
 2016-05-15 20:31:05 +00:00
espie
6154b12d66 variables for simpler dpb builds 2016-05-15 14:14:45 +00:00
espie
56d223a5ab cosmetic shit for aja instead of writing useful code. 
okay aja@
 2016-05-15 14:02:01 +00:00
espie
b7438ad0d0 have check_mount be removable as well (bad idea) and add 
actions=none
to reset to nothing
 2016-05-15 13:25:25 +00:00
espie
6db7739891 tweak interface, add extra to copy extra files, 
mkconf_tail to add a tail to generated mk.conf.
zap copying doas.conf, which is not really that useful especially
since chroot is nosuid often.
 2016-05-15 11:06:57 +00:00
espie
e0fc66e45a get rid of the last chdir 2016-05-15 10:24:38 +00:00
espie
cda8f8d456 remove more chdirs from the main process 2016-05-15 10:00:16 +00:00
espie
38a933b11a zap yet another chdir 2016-05-14 22:49:23 +00:00
espie
b280a55f4f smarter copy of sys includes, no chdir 
don't copy obj/ or CVS/
 2016-05-14 15:19:48 +00:00
espie
487a90448c start removing uses of chdir() in main process 2016-05-14 15:00:07 +00:00
espie
549ccc20b7 kill unneeded special case from copy_sys 2016-05-14 13:55:42 +00:00
espie
e60df16782 copy doas.conf if it exists (convenience request from landry) 
use devno to filter out mount points (more sturdy when symlinks happen)

FIX the regexp used for mount points which did kill unpopulate too much
 2016-05-14 11:16:29 +00:00
espie
79bdd956dd not yet 2016-05-14 10:05:55 +00:00
espie
afa0dc0dd5 if one of your dirs is a symlink that doesn't point anywhere, then 
stat will fail, which is okay for LOGDIR which only exists locally.
 2016-05-14 09:19:09 +00:00
espie
4f9e4555ae accidentally broke preserved mountpoints, fix that 2016-05-14 09:04:40 +00:00
espie
fb8f1bf77c always do mkpath, THEN check for nfs mounts, so that non-existing 
dirs (at first) get classified correctly.

thx aja@ for testing on a virgin dir
 2016-05-13 22:40:28 +00:00
espie
06e0c3279c tweak user message, requested by dcoppa@ 2016-05-13 14:30:06 +00:00
espie
36978edd58 check for absolute symlinks in the chroot. 
report anything that we don't know about (from the base system)
 2016-05-13 14:17:24 +00:00
espie
a892c3669a things that live strictly in the chroot can mostly be run there. 
this solves the issue of resolving symlinks within the chroot
and matching them against nfs dirs, at the expense of a few more processes
 2016-05-13 13:37:22 +00:00
espie
29bd19a65a try to copy mtimes and use that to check 2016-05-13 02:38:03 +00:00
espie
1752ecd04d surprised this one didn't break more things 
thanks ken/stuart
 2016-05-12 13:14:13 +00:00
espie
685cbeaac4 add the actual option "chown_all" 2016-05-11 12:51:07 +00:00
espie
602498e17e add some smarts to the ports directory checker: whine if WRKOBJDIR or 
LOCKDIR are under nfs (often not a good idea), don't chown recursively
by default, and don't even when under nfs
 2016-05-11 12:50:02 +00:00
espie
b159ab1537 prevent unpopulate from wiping out ports subdirectories 2016-05-10 16:13:31 +00:00