1
0
mirror of https://git.FreeBSD.org/ports.git synced 2024-11-23 00:43:28 +00:00
Commit Graph

1017 Commits

Author SHA1 Message Date
Doug Barton
d2c1853804 INDEX-6 -> INDEX-8
Submitted by:	gjb
2011-09-28 01:44:31 +00:00
Erwin Lansing
2b35911821 Add support from FreeBSD 10.x 2011-09-27 06:29:35 +00:00
Chris Rees
5d4b8b15b0 Add option (-y) to check for similarly named ports.
PR:		ports/159841
Submitted by:	crees (me)
Approved by:	garga (maintainer)
2011-08-29 14:54:48 +00:00
Stanislav Sedov
bfc1333181 - Use optparse instead of old deprecated method.
- Kill ruby 1.6 related stuff.
2011-08-23 09:23:54 +00:00
Stanislav Sedov
d39e30b3dc - Fix options processing (optparse puts options into hash indexed by strings,
not symbols).
2011-08-23 09:17:46 +00:00
Stanislav Sedov
90d33c8b8c - Make it work with ruby 1.9. 2011-08-23 08:42:31 +00:00
Steve Wills
c1c18af0fe - Back out category check for now 2011-08-16 00:09:12 +00:00
Steve Wills
4a09663ed3 - Remove obfuscation of submitter email at request of dougb
- Add check that port is in right directory as I've been bitten by this
2011-08-15 21:21:41 +00:00
Emanuel Haupt
e7cc2ccf33 Fix readme building. 2011-08-12 06:56:44 +00:00
Renato Botelho
390fe1afe5 Change shebang to #!/usr/bin/env perl, as required by bapt@ 2011-07-27 17:18:26 +00:00
Erwin Lansing
2b43202957 Remove duplicate words in comment.
Submitted by:	ohauer
2011-05-02 09:38:04 +00:00
Florent Thoumie
03074365a5 Remove portbuild scripts from pcvs, as they now live in svn/projects. 2011-04-24 16:37:51 +00:00
Mark Linimon
3fcbbb8d24 Add some definitions to try to limit the number of times we let builds
fail before we stop.  This should help avoid spamming people's mboxes.
2011-04-22 18:32:13 +00:00
Mark Linimon
9fd443c7ba Allow the top-level build to be destroyed. This is useless for
pointyhat, but is intended for pointyhat-[west|east] to be able to
remove no longer needed specialized -exp runs.  (They remain in the
errorlog archive; they can merely be removed from zfs.)

Replace this test with an anti-foot-shooting measure so that you cannot
remove 'latest' unless you have already removed 'previous'.
2011-03-24 06:24:04 +00:00
Pav Lucistnik
95f63674f3 - Stop caring about src-6 snaps and expire checkout snaps rather more
aggresively
2011-03-21 21:43:58 +00:00
Mark Linimon
068948ee07 Add an error entry for gmake. 2011-03-08 22:03:18 +00:00
Mark Linimon
547dbc0a08 Add a test for gmake 3.82 regressions. (3.82 is not in the tree yet).
While here, move segfault down: sometimes this is a false positive for
a config failure.
2011-03-08 21:53:53 +00:00
Mark Linimon
46a7d368e1 Add an alternate spelling for a link. 2011-03-04 06:11:51 +00:00
Mark Linimon
4a3b759d6b Add an overly-large regexp to isolate "packages that fail to build with
clang".  This will probably need to be trimmed down some so as to not
pessimize processing time.

At this time this change will be restricted to pointyhat-west.
2011-02-27 07:08:01 +00:00
Mark Linimon
f7fd413b58 One more pass through this file to bring it up to the latest processonelog:
- deprecate assert, chown, perl5, python, stdio, sysvipc, varargs

 - add the existing nested_declaration, termios, utmp_x

While here, fix formatting.

Pointy hat to:	linimon, for letting this get so far out of sync
2011-02-27 06:55:38 +00:00
Mark Linimon
e4a1d6f3b8 For efficiency's sake, deprecate the following error types, which we no longer see:
chown, perl5, python, stdio, struct changes, sysvipc, USE_XLIB, varargs

Add the following new error types that we are now seeing:

  clang-bug, cpusetsize, sem_wait

Tweak the definitions for linker_error and threads to be more accurate.
2011-02-27 06:33:45 +00:00
Mark Linimon
7b36c962d5 Sheesh. The errors "f77" and "uname" were never even documented. Add
the into the section for obsolete errors.
2011-02-27 06:16:01 +00:00
Mark Linimon
0b94659e59 Move some of the error types that were removed from the script some time
ago to the "obsolete" section.
2011-02-27 06:01:20 +00:00
Mark Linimon
5d2451c2a6 Add the following new error classifications, to sync with upcoming work
to processonelog (not yet committed):

  clang
  clang-bug
  cpusetsize
  sem_wait

The former two will show up if an -exp run is done with clang as the default
ports compiler, instead of gcc.  The latter two seem to be ports regressions
in 9-CURRENT.
2011-02-24 22:28:22 +00:00
Mark Linimon
2fc5780a86 Remove 6.x links since it is EOL. (For now, the error logs remain).
While here, change the misnomer Full to Previous, but leave the links
alone until we upgrade pointyhat to the latest codebase.
2011-02-22 03:07:11 +00:00
Erwin Lansing
90b43d628c Make table headers more selfexplanatory
Submitted by:	gerald
Feature safe:	yes
2011-02-05 16:42:00 +00:00
Erwin Lansing
7b3fc2c221 7.1 runs EOL today, start building against 7.3
Feature safe:	yes
2011-01-31 09:29:22 +00:00
Mark Linimon
e93f4465dc Change the "unclean build" handling case to not retry. Previously, it
was handled the same as "waiting for chroot" and retried.  In certain
circumstances (e.g. a prerequisite package shown as being avaialable
when it really wasn't), this would cause a package build to loop
forever.

Note to future generations: I did not write this code.

Feature safe:	yes
2011-01-26 10:41:53 +00:00
Mark Linimon
4ddc8569cb Fix typo in fixed-up tar command to suppress a whoooole lot of junk
output on package builds.

Pointy hat:	self
Feature safe:	yes
2011-01-24 01:58:21 +00:00
Mark Linimon
2d9bb86df0 Add an entry for the 'dirent' error.
Noticed by:	Anton Shterenlikht
Feature safe:	yes
2011-01-24 00:44:53 +00:00
Mark Linimon
c53373e936 Add anti-foot-shooting measure.
Feature safe:	yes
2011-01-23 03:06:44 +00:00
Mark Linimon
883391196b Export BACKUP_FTP_SITE for the generalized make.conf files in
bindist-*.tar (so that they can be shared between pointyhat and
pointyhat-west without modification.)

Feature safe:	yes
2011-01-23 02:39:54 +00:00
Mark Linimon
d2628e1ceb Add CLIENT_BACKUP_FTP_SITE as part of a generalization push between
bindist-*.tar files on pointyhat and pointyhat-west.

Feature safe:	yes
2011-01-23 02:34:58 +00:00
Mark Linimon
be8848c7d8 Force the copy of portbuild.* files as files, not symlinks. Adjust
the comment for bindist-* to match.

Feature safe:	yes
2011-01-23 02:26:14 +00:00
Erwin Lansing
4f169b63af Revert to RELENG_7_1 after consultation with re@
Feature safe:	yes (this time for real)
Approved by:	portmgr,re
2011-01-18 15:57:00 +00:00
Erwin Lansing
8e2ba39faf Bump 7-stable to RELENG_7_4 due to a major library version increase
in libcrypto.

Feature safe:	yes (really, really not, but we don't have any other option)
2011-01-14 22:36:32 +00:00
Erwin Lansing
b40874d29c rmdir(1) does not take -f. 2011-01-05 18:34:17 +00:00
Erwin Lansing
fa2e0c2cef Drop support for 6.x and bump 8.x to 8.1-RELEASE 2010-12-21 08:25:08 +00:00
Mark Linimon
6dad2ee054 Remove debugging code that crept in. 2010-12-17 01:34:38 +00:00
Mark Linimon
6f337f74cd Refactor the client-side. It turns out that the evaluation of param.h
was being done before the src tree had been mounted.

This was not a problem in the old codebase, since it simply got the
wrong src tree to start with.

This actually simplifies the code.
2010-12-16 07:38:18 +00:00
Mark Linimon
b0a015135a If the logfile was not correctly copied from the client back to the server,
do not attempt to process it.
2010-12-16 07:34:58 +00:00
Mark Linimon
2935ba5fee Make sure that PKGSUFFIX and LOCALBASE are really defined. (Things work
_very_ badly if they aren't.)

Also, track the refactoring of buildenv so that the proper src tree
bits are picked up.
2010-12-16 07:32:22 +00:00
Erwin Lansing
34b70fee7a Drop support for FreeBSD 6.x. 2010-12-14 14:16:37 +00:00
Mark Linimon
5bc0ea6f26 common.conf has to be copied over as well as client.conf. 2010-12-09 00:50:02 +00:00
Mark Linimon
28905d9035 Refactor some variables that need to be in sync between the server-side
and client-side definitions into common.conf.

While here, slide the 7-stable release to 7.3.  This was driving me
crazy while trying to debug the INDEX and duds problems.
2010-12-01 02:38:04 +00:00
Mark Linimon
5cfbfdffd1 It turns out that some variables have to be shared between the server-side
definitions and the client-side definitions.

While here, add ARCHS_SUPPORTING_COMPAT_IA32, which is needed to get
linux emulation ports to be properly detected on head nodes that are
build from first principles (e.g. !pointyhat).
2010-12-01 02:35:20 +00:00
Mark Linimon
c836d9acad Canonicalize the use of SRC_BASE to fix INDEX and duds builds on head
nodes that don't have /usr/src.  On pointyhat, this "just happens to
work"; on a cleanly installed system like pointyhat-west, it didn't.

Add overrides for Linux compatibility (again, it "just happened to work"
on pointyhat).

Move the code that sets PORTSDIR and SRC_BASE to the server side only.
The code that overrides OSVERSION, OSREL, and BRANCH _must_ remain in
the common code.
2010-12-01 02:32:49 +00:00
Mark Linimon
ac775267a0 Canonicalize the use of SRC_BASE to fix INDEX and duds builds on head
nodes that don't have /usr/src.  On pointyhat, this "just happens to
work"; on a cleanly installed system like pointyhat-west, it didn't.
2010-12-01 02:30:14 +00:00
Mark Linimon
d072a14cd7 Make an rm command quieter; add some debug comments. 2010-12-01 02:28:19 +00:00
Mark Linimon
08037129f0 Although I was going to try to avoid it, it turns out that it is
impossble to have merely "server.conf" and "client.conf" control
everything on a head node.  Some pieces must be refactored into
"common.conf".
2010-12-01 02:27:35 +00:00