cpet/openbsd-ports
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
90,724 Commits 1 Branch 0 Tags
Commit Graph

3848 Commits

Author SHA1 Message Date
nigel
2c8b12f938 Ok @espie 2014-03-23 12:24:14 +00:00
espie
5a5389697a zap $_ usage 2014-03-21 11:01:43 +00:00
jasper
4a880144f5 sync with recent hp300/mvme{6,8}8k/m68k removal 
ok aja@
 2014-03-19 13:28:04 +00:00
espie
aeb62cfe9c my $_ 2014-03-18 20:39:45 +00:00
espie
a384b0d3a8 my tracer used lexical $_ 2014-03-17 10:49:19 +00:00
espie
f82962b283 prepare for perl 5.18 
- don't bother declaring $_ for while(<>) loops. assume current default
will still work.
- rename variable for readability elsewhere.
 2014-03-17 10:48:40 +00:00
espie
1af9cd4e04 don't bother re-adding already shunted depends. 2014-03-17 10:47:45 +00:00
espie
cfb2f4c5fd log when we actually run engine, since this is hard to match with perf log. 2014-03-17 10:47:12 +00:00
espie
87ed35ed60 obvious typo. spotted by jasper@ 2014-03-16 17:22:05 +00:00
espie
9c85c8f67b do the depends computation before the fork and after the lock, so we're 
aware of everything going on on the same host. In particular, group together
other ports waiting for lock on the same host, and run one single pkg_add.
since they are now in order, this helps quite a bit.
 2014-03-16 08:52:53 +00:00
espie
bf709a9acc tweak waiting-for-lock mechanism: explicitly keep track of order, 
and have each job relinquishing lock wake up the next one.

To cater for external dpbs, also try to wake up next job if we don't own the
lock.
 2014-03-15 09:51:27 +00:00
sthen
2ea2cd04c4 ugh, forgot to commit this 2014-03-13 22:10:53 +00:00
landry
467d26b5d9 Remove trailing spaces (long line iz looooooong!) that snuck in when i 
failed at copy&paste.
 2014-03-10 21:35:04 +00:00
landry
243a20d699 Default to gcc 4.8 for ports using gcc4 module. 
Stay at 4.6 for hppa & alpha for now, since lang/gcc/4.8 doesnt build there
yet.
Discussed with espie@ pascal@ sthen@ tobiasu@ ajacoutot@
 2014-03-10 20:26:00 +00:00
espie
556d8b312b extract a large part of check-conflicts. aim is to reuse that in 
check-common-dirs, as those two tools do very similar things
 2014-03-10 09:46:08 +00:00
espie
89a8561c00 explicitly prevent tty allocation on distant host 2014-03-10 09:34:54 +00:00
espie
d819111cd9 figure out if we are running as root thru ssh, thus removing the need 
for an extra sudo step
 2014-03-10 09:34:07 +00:00
espie
7205c851fa note to myself, this makes deduplicating paths in the queue complicated. 2014-03-10 09:06:12 +00:00
ajacoutot
02cd405e2b Add an intermediate variable because some tools like python to not 
properly parse variables with trailing spaces and add a bogus "" argument.

input/ok espie@
 2014-03-09 21:40:57 +00:00
espie
4af4841543 allows for a global cache to depends. to be documented once this is fully 
figured out...
 2014-03-09 20:33:37 +00:00
espie
c34dccf81f well, root@echo is a valid hostname 2014-03-09 20:31:04 +00:00
espie
8086dd43f2 be slightly less eager when running junk: take into account dependencies 
that ports already scheduled want.
 2014-03-09 20:30:10 +00:00
espie
ef48774968 tidy: extra space. and make sure we have an offset in the log. 2014-03-09 20:25:19 +00:00
espie
d7208cc508 micro optimization, visible on loads of cores: release swallowed cores at 
end of fake, so they have time to sort themselves out (waiting-for-lock)
before package is finished.
 2014-03-09 20:24:18 +00:00
espie
611e587927 path normalization: introduce can_really_start_build 
that handles held locks, and starts build.

Specialize it for  normal builds to remove all build_path_list elements
at once, and choose the smallest one.

in all, this means that the logs will show less variety in what they
build, thus giving better statistics earlier.
 2014-03-09 20:15:10 +00:00
espie
b75849a1af pass logfile to the sizer, so it can output more specific information, 
like its size computations.
 2014-03-09 20:11:33 +00:00
espie
b3b46b315d jobs will register the dependencies they would want, preventing junk 
from removing them (so that we do less delete/re-add cycles, the more
cores you have the more relevant this becomes).

again, found on Yandex machines, thx to 12 cores or more.
 2014-03-09 20:09:53 +00:00
espie
0931d44776 - refactor the unswallow code for readability, and to be able to use it 
at the end of fake
- sort available cores so that if all hosts fire at the same time, the
load will be fair (found out on yandex machines)
- new iamroot property: for chroot distant hosts, it's simpler to connect
as root, then you can directly chroot -u user  and not need the extra sudo
stage.
- more specific "walk_same_host_jobs" to use to signal jobs, for instance.
 2014-03-09 20:08:32 +00:00
espie
6efb77041e color tweaks: 
- use blue for "waiting-for-lock"
- make "frozen" actually readable (black on red, change percent
from purple to white)
- fix host names.
 2014-03-09 20:04:57 +00:00
espie
f5f0d95820 - replace `` with $$() in most places. 
stylistically, I tend to prefer ``, but since we use a lot of fragments,
the parenthetical nature of $() is better.
- give a more complete name to portstree-/inst-  depends lists, so that several
"make package" can run in the same cache.
- create temp dependency files atomically by using mktemp/chmod/mv
(global cache)
 2014-03-09 20:03:27 +00:00
espie
c8571e2c0f restore BUILD_ONCE optimization for full ports builds 2014-03-09 19:47:35 +00:00
espie
06a9ae3967 documentation tweak 2014-02-14 14:30:05 +00:00
espie
b0b03e9d58 cosmetic: print-package-signature -> print-update-signature 
(to be referenced by documentation, to stop confusing crypto signatures
with other signatures).

(old name kept over release, to be scraped completely afterwards, user-visible
change only)

okay sthen@
 2014-02-11 10:34:34 +00:00
zhuk
934244224a Warn about empty lines in PLIST/PFRAG files. Idea supported by jasper@. 2014-02-09 19:26:25 +00:00
zhuk
59eeb06308 Rever previous, it actually makes portcheck do not trigger for problematic 
/var paths at all. A better fix will come later, let's go for low hanging
fruits now.
 2014-02-09 18:11:13 +00:00
zhuk
47f5103e9f Do not warn about hardcoded /var/log/*. 2014-02-09 17:24:48 +00:00
zhuk
c95c81bf60 Fix checks for duplicated assignments. Noted while watching at 
portcheck -A output for audio/soundtouch.
 2014-02-09 16:09:55 +00:00
zhuk
0b6bfe437d Use more precise name for variable, to make logic more clear. 2014-02-09 15:44:20 +00:00
espie
1e496a8a62 typo, avoid creating oodles of empty objects 2014-02-09 15:24:20 +00:00
zhuk
9189b7341e Fix default ignore list for -A, and add "tests" there. 2014-02-09 15:19:32 +00:00
zhuk
28a1c15db0 Unbreak portimport after previous. Clean up getopts call while there. 2014-02-09 15:10:49 +00:00
espie
9efc37aec8 actual progress display... oh, and for_list does the next too :( 2014-02-08 13:50:58 +00:00
espie
05495d6602 add (pkgpaths) to pkgnames 2014-02-08 13:37:27 +00:00
espie
90e985ef25 document extra semantics 2014-02-08 13:22:24 +00:00
espie
fe2b03069f for plists, skip directories (history) and report ba plists. 
user improvement:
with an extra -d, also scan the directories for "current" pkgnames,
and display a ! in pkgnames that are current while reporting conflicts.
 2014-02-08 13:17:19 +00:00
espie
72d7c489b2 style 2014-02-08 10:48:42 +00:00
espie
79886990ba missing annotation 2014-02-08 10:47:59 +00:00
sthen
a52494ad03 remove TEST/TESTS poisons, they break existing ports, found by naddy 
move VMEM_WARNING from the "misspellings" block to a separate section
with a comment (used to exist but retired)
 2014-02-06 12:49:17 +00:00
gsoares
ec2c0bde01 .poison TEST, TESTS and NO_TESTS 
ok sthen@
 2014-02-06 05:12:06 +00:00
zhuk
7e26be973e Fix inconsistence when printing names of files in pkg/ directories. Found 
by regression suite.
 2014-02-04 21:28:30 +00:00