10:10 mat 

Add a Perl CORE QA check.

This will use Module::CoreList to tell you if a dependency you added can
be removed, or only conditionally added.

Approved by:	bapt
Sponsored by:	Absolight
Differential Revision:	https://reviews.freebsd.org/D7832

 

09:30 mat 

Fix file modes after extracting.

Some port's archives contains files modes that are a bit too restrictive
for some usage.  For example:
BUILD_DEPENDS=		${NONEXISTENT}:foo/bar:configure
When building as a regular user, dependencies are installed/built as
root, so if the archive contains files that have a mode of, say, 600,
they will not be readable by the port requesting the dependency.
This will also fix broken distribution files where directories don't
have the executable bit on.

OSVERSION 1100077 is after base r283997:

  Change directory permissions in pre-order.
  In this order, it won't try to recurse into a directory for which it
  doesn't have permission, before changing that permission.
  This follows an existing behavior in other BSDs, linux, OS X.

PR:		213574
Submitted by:	mat
Exp-run by:	antoine
Sponsored by:	Absolight

 

12:51 mat 

${RM} already has -f.

PR:		213570
Submitted by:	mat
Exp-run by:	antoine
Sponsored by:	Absolight

 

17:22 tcberner 

Introduce new extensible virtual categories for KDE

A lot of KDE Ports share MASTERSITES, LICENSE and so one, as they are
released as a bundle upstream, however, there was not really a clean
way to share this information.

Using these new categories, we can simplify the Makefiles for the diverse
KDE ports.
At the moment we support the virtual category
	* kde-kde4
In the future, this will be extended to
	* kde-frameworks
	* kde-plasma
	* kde-applications

PR:						213406
Differential Revision:	https://reviews.freebsd.org/D7645
Exp-run by :			antoine
Reviewed by:			mat, rakuco
Approved by: 			portmgr (mat), rakuco (mentor)

 

19:42 amdmi3 

Most commonly used build systems support silent builds, when they
hide actual commands executed and only show short summary line (like
"CC foo.c"). CMake and ninja enable this by default, some autotools
using ports do as well. This is unacceptable because we need complete
build logs at any time, so we now switch to verbose build logs
unconditionally. Note that this change deliberately affects ALL
builds and not only package builds on cluster, because we need to
be sure that user experiencing failure can always provide informative
build log regardless of settings and without rerunning the build.

Change summary:

- Always do verbose builds for cmake, ninja and GNU configure (the
  latter includes check if --disable-silent-rules is actually supported
  by the configure script; there are isolated cases when it's not true)
- Remove CMAKE_VERBOSE, NINJA_VERBOSE and
  CONFIGURE_ARGS=--disable-silent-rules from all ports which set them
  for this is no longer needed
- Revert hacks for --disable-silent-rules support priorly committed
  to biology/ncbi-blast+ and net-p2p/mldonkey - no longer needed as well

Submitted by:	amdmi3
Reviewed by:	mat
Exp-run by:	antoine
Approved by:	portmgr (mat, antoine)
Differential Revision:	D7534

 

21:34 bapt 

Rework the cross build framework so that it uses the external toolchain
It allows to cross build everything wich is not expecting target dependencies

 

19:23 tijl 

- Replace Mk/bsd.linux-apps.mk and Mk/bsd.linux-rpm.mk with
  Mk/Uses/linux.mk.
- Replace USE_LINUX=yes with USES+=linux and USE_LINUX=(.*) with
  USES+=linux:\1 in all ports.
- Replace USE_LINUX_APPS with USE_LINUX in all ports.
- Use INSTALL_SCRIPT instead of INSTALL_PROGRAM to install scripts in some
  ports.
- When USE_LINUX_RPM is defined, simplify the way DISTFILES and EXTRACT_ONLY
  are defined.
- Remove BRANDELF_DIRS and BRANDELF_FILES handling.  In the very rare cases
  that it is still necessary ports can run ${BRANDELF} from post-patch.
- Remove AUTOMATIC_PLIST handling.  Only one port used it.
- Fix Linux MASTER_SITES.
- Replace OVERRIDE_LINUX_BASE_PORT and OVERRIDE_LINUX_NONBASE_PORTS with
  default versions framework.
- bsd.port.mk:
  - Move Linux related bits to Uses/linux.mk, except USE_LINUX_PREFIX.
  - Put USE_LINUX_PREFIX handling after USES processing.
  - Define DOCSDIR, DATADIR, etc. after handling USE_LINUX_PREFIX so it can
    give these variables a different default value.
  - When a package needs to run Linux ldconfig check before installation if
    Linux support is enabled.
- emulators/linux_base-*:
  - Use USES=linux and remove duplication.
  - Remove files/lp.  FreeBSD or CUPS lp(1) should work.
  - Remove files/yp.conf.  No longer seems to be used.
  - Remove pkg-deinstall and move pkg-install into pkg-plist.
  - Update pkg-descr and pkg-message.
  - Fix handling of ldconfig cache in pkg-plist.
- devel/fb-adb: Use a Linux shell to run a Linux script but patch the script
  to use FreeBSD mkdir so mkdir -p $path creates $path and not
  /compat/linux/$path.

PR:		211645
Exp-run by:	antoine
Approved by:	portmgr (antoine)

 

22:08 mat 

Introduce IGNORE_opsys/osrel/arch, similar to the BROKEN_* ones.

Sponsored by:	Absolight

 

15:23 mat 

Add a stage-qa check to check for the existence of SONAME in .so's.

If a port provides .so.X files, they have to have a SONAME for them to
work correctly.

While there, incorporate a reverse soname checks in proxydeps.
A port that needs a .so.X from another port which does not have a
SONAME.

Reviewed by:	bapt
Sponsored by:	Absolight
Differential Revision:	https://reviews.freebsd.org/D7454

 

14:50 mat 

Add GH_SUBDIR, automatically moves a secondary distfile to the right
place inside ${WRKSRC}.

Before:

GH_TUPLE=	Regaddi:Chart.js:f13f99b:chart_js \
		FVANCOP:ChartNew.js:77e7f87:chartnew_js
post-extract:
	@${RMDIR} ${WRKSRC}/database ${WRKSRC}/3rd/Chart.js
	@${MV} ${WRKSRC_database} ${WRKSRC}/database
	@${MV} ${WRKSRC_chart_js} ${WRKSRC}/3rd/Chart.js

After:

GH_TUPLE=	Regaddi:Chart.js:f13f99b:chart_js/3rd/Chart.js \
		FVANCOP:ChartNew.js:77e7f87:chartnew_js/3rd/ChartNew.js

Sponsored by:	Absolight
Differential Revision:	https://reviews.freebsd.org/D7583

 

08:20 tcberner 

Replace Mk/bsd.kde4.mk by Mk/Uses/kde.mk in preparation for KDE Frameworks and
Plasma5 ports

At the moment KDE ports use bsd.kde4.mk to handle their dependencies. When
working on the ports for KDE Frameworks and Plasma5 it seemed to be more
reasonable to create a new kde.mk instead of adding an bsd.kde5.mk.

The kde.mk in this review is a stripped down version of the one we are using in
the KDE Test repositories plasma5 branch [1] to only contain the parts relevant
to the current KDE4 ports in the portstree [2].

Changes to the KDE Ports needed by this:

Replace USE_KDE4 by USE_KDE [3]
Add USES=kde:4 [4]
[1] http://src.mouf.net/area51/view/branches/plasma5/KDE/Mk/Uses/kde.mk
[2] The version in the plasma5 branch also handles frameworks/plasma5 and
    handles MASTER_SITES via a KDE_DIST variable similar to bsd.qt.mk for Qt
    Ports -- I chose to leave this out for now, as the diff is already large
    enough.
[3] I chose USE_KDE instead of USE_KDE4, USE_KDE5, USE_KDEX as the version we
    want is already specified as argument to kde:<arg>
[4] For KDE Frameworks and Plasma5 ports this would be kde:5

PR:             210667
Approved by:    portmgr, mat (mentor), rakuco (mentor)
Reviewed by:    mat, rakuco
Differential Revision:   https://reviews.freebsd.org/D6961

 

13:09 mat 

Fixup the PLIST_SUB_SED creation.

PR:		211995
Reported by:	koobs
Sponsored by:	Absolight

 

21:07 mat 

Put back PLIST_DIRS handling.

My eyes were playing tricks on me.

PR:		211953
Reported by:	dim
Pointy hat:	mat
Sponsored by:	Absolight

 

12:16 mat 

Move USE_BDB and PLIST_DIRSTRY to the unsupported section, all the ports
have been converted.

Sponsored by:	Absolight

 

09:11 mat 

Add regexps capacity to PLIST_SUB.

This adds the possibility to use regular expressions for the makeplist
stage of the PLIST_SUB life.

From time to time, the values are too generic, and they get in the way of other
stuff.

This adds the possibility to have a `VAR_regex=regex` that will be used instead
of the `VAR=string` to search for possible replacements.

For example, in lang/perl5*, there is PERL_ARCH=mach, which will get
replaced in paths if a file is called, say "machine", will end up being
"%%PERL_ARCH%%ine".  Adding PERL_ARCH_regex="\bmach\b" will ensure only
full words are replaced, so machine will stay machine, but "lib/mach/foo
"will still be replaced by "lib/%%PERL_ARCH%%/foo".

Reviewed by:	bdrewery
Sponsored by:	Absolight
Differential Revision:	https://reviews.freebsd.org/D7335

 

09:11 mat 

Split the PLIST_SUB_SED variable in smaller logical groups.

This will allow expanding it more easily.

Sponsored by:	Absolight

 

08:54 mat 

Unbreak make missing.

Sponsored by:	Absolight
Differential Revision:	https://reviews.freebsd.org/D7504

 

14:29 mat 

Don't overquote the DEPRECATED message, it ended up looking silly.

Something like:

-----------
This port is deprecated; you may wish to reconsider installing it:

EOL\ upstream\ since\ 10\ Jul\ 2016.

It is scheduled to be removed on or after 2016-08-20.
-----------

PR:		211421
Reported by:	Miroslav Lachman
Sponsored by:	Absolight

 

12:48 mat 

Prefix the PKGMESSAGES variable with an _ to show it is private.

Sponsored by:	Absolight

 

12:45 mat 

Add a PKGMESSAGES variable that allows the framework to have more than
one PKGMESSAGE file.

This allows the framework to add messages to ports, without touching
their PKGMESSAGE file.

Sponsored by:	Absolight

 

12:09 mat 

Always include bsd.default-versions.mk in bsd.port.mk.

The variable defined in it are now always available after including
bsd.port.pre.mk.

PR:		210666
Submitted by:	mat
Exp-run by:	antoine
Sponsored by:	Absolight
Differential Revision:	https://reviews.freebsd.org/D6933

 

17:57 bapt 

Do not try to register '/' as a directory when PREFIX is set to '/'

 

19:12 adamw 

Extend the USES=php match so that it recognizes php:web as well, to prevent
an erroneous deprecation warning.

PR:		210822
Approved by:	portmgr (mat)

 

13:26 mat 

Catch up two instances of USE_OPENSSL.

Sponsored by:	Absolight

 

11:31 mat 

Replace bsd.openssl.mk with USES=ssl

Add a qa hint about needing, or not, USES=ssl.

Fix ports doing silly things, like including bsd.openssl.mk directly.

PR:		210322
Submitted by:	mat
Exp-run by:	antoine
Sponsored by:	Absolight
Differential Revision:	https://reviews.freebsd.org/D6866

 

03:21 ohauer 

- add _ypldap to USERS_BLACKLIST and GROUPS_BLACKLIST

PR:		210282
Submitted by:	ohauer
Approved by:	portmgr (mat@)

 

13:19 mat 

Move bsd.php.mk to Uses/php.mk

PR:		210323
Submitted by:	mat
Exp-run by:	antoine
Sponsored by:	Absolight
Differential Revision:	https://reviews.freebsd.org/D6867

 

11:25 mat 

Add a few checks for (MASTER|PATCH)_SITES groups.

While make(1) is ok with variable names having quite a lot of strange
characters in them, the fetch code mostly uses sh(1), where variable
names can't include a dash (or pluses, or many other things).

PR:		210251 210198
Submitted by:	mat
Exp-run by:	antoine
Sponsored by:	The FreeBSD Foundation, Absolight
Differential Revision:	https://reviews.freebsd.org/D6779

 

11:25 mat 

Extract create-manifest.

PR:		210198
Submitted by:	mat
Exp-run by:	antoine
Sponsored by:	The FreeBSD Foundation, Absolight
Differential Revision:	https://reviews.freebsd.org/D6779

 

11:25 mat 

Remove unnecessary evals that do-fetch was using.

Turns out that env(1) knows how to parse a properly quoted string
using -S "string", it makes the double eval used for the fetch
command to be unnecessary.

This is because running:
eval "foo $(escape $bar)"
is silly when this works:
foo $bar

So remove escaping and quoting, and sillyness.

PR:		210198
Submitted by:	mat
Exp-run by:	antoine
Sponsored by:	The FreeBSD Foundation, Absolight
Differential Revision:	https://reviews.freebsd.org/D6779

 

11:24 mat 

Simplify do-fetch, and remove now unused code.

Turns out that the DEFAULT, and other groups sites were handled in
a different way, ending up having the same code twice, but in
slightly different places.

PR:		210198
Submitted by:	mat
Exp-run by:	antoine
Sponsored by:	The FreeBSD Foundation, Absolight
Differential Revision:	https://reviews.freebsd.org/D6779

 

13:00 mat 

Grmbl, poudriere still calls this unused empty target, so putting it back.

Reported by:	antoine
Sponsored by:	Absolight

 

12:42 mat 

Remove DO_NADA targets.

No need to have make(1) compute a target order graph for targets that
are empty.

Sponsored by:	Absolight

 

18:41 bapt 

Fix make check-vulnerable

 

15:58 mat 

Extract the larger bsd.port.mk targets into separate scripts.

Refactor all the fetch code so that there are not 6 slightly different
versions of it but one that does it all.

The targets that have been extracted are:
- check-vulnerable
- do-fetch
- fetch-list
- fetch-url-list-int
- fetch-urlall-list
- checksum.
- makesum.
- check-checksum-algorithms

Run the fetch code directly from make makesum instead of calling make
fetch, this is because some port change the options with OPTIONS_*_FORCE
when make(makesum) to be able to add all distfiles in one go, which was
a nice, non working, idea.

PR:		208916
Submitted by:	mat
Exp-run by:	antoine
With hat:	portmgr
Sponsored by:	Absolight
Differential Revision:	https://reviews.freebsd.org/D5997

 

11:59 mat 

A new stage-qa test has been added, it reports all shared libraries
dependencies that are not part of the port list of dependencies.  It help
finds what is called proxy dependencies.

A is needed by B, and B is needed by C.  If C also needs A, then it needs to
be registered, and this check will tell you to do so.

Right now, it is only reporting the problems, but if you add
PROXYDEPS_FATAL=yes to your environment, it will give an error and will force
you to fix the dependencies.

PR:		195203
Submitted by:	yuri rawbw com (earlier version)
Sponsored by:	Absolight
Differential Revision:	https://reviews.freebsd.org/D6531

 

22:55 bdrewery 

Re-enable code from r415530 with safer MAKE_ENV support.

PR:		D6271
Sponsored by:	EMC / Isilon Storage Division

 

22:24 bdrewery 

Disable the change in r415530 until an exp-run is done.

Passing things in MAKE_ARGS tends to break some obscure ports.

Reported by:	antoine
PR:		D6271

 

22:06 bdrewery 

Use MK_* to prevent debug files if possible.

A longstanding bug with the base /usr/share/mk files is that /etc/src.conf
*does* get pulled in when building ports which use /usr/share/mk.  This despite
the documentation saying it is only used for "the FreeBSD tree".  If users
have a WITH_DEBUG_FILES=yes in their src.conf then the previous
WITHOUT_DEBUG_FILES=yes would hit an error.  Fixing pulling in src.conf is
risky.  MK_* overrides work fine here after recent stable/9 and stable/10
updates.

If you run into the WITH_ vs WITHOUT_ problem on a release build then you can
protect your src.conf value with something like:
  .if !${.CURDIR:M*ports*}
  WITH_DEBUG_FILES=yes
  .endif

PR:		D6271
With hat:	portmgr
Sponsored by:	EMC / Isilon Storage Division

 

20:53 jonathan 

Fix "need root" logic for PORT_DBDIR.

When PORT_DBDIR is set to a user directory, we shouldn't need su to
manipulate options. This also shouldn't depend on INSTALL_BY_USER, as
building != installing.

Approved by:	portmgr (bapt)
Differential Revision:	https://reviews.freebsd.org/D6349

 

18:20 emaste 

Record TIMESTAMP in make makesum

This is intended to support development and prototyping for ports
reproducible build efforts which require some concept of a "last
updated" time. It is being committed now so that timestamp entries
will be populated "for free" as port distfiles updates happen.

Submitted by:	bapt (earlier version)
Reviewed by:	bapt, bdrewery
Approved by:	portmgr (bdrewery)
Differential Revision:	https://reviews.freebsd.org/D6031

 

10:12 mat 

Try to simplify how libperl.so is looked for.

PR:		209123
Submitted by:	mat
Exp-run by:	antoine
With hat:	perl@
Sponsored by:	Absolight
Differential Revision:	https://reviews.freebsd.org/D6107

 

07:28 mat 

Remove quotes here, ${m} is already quoted.

With the quotes, it ends up doing, for example:

echo ""It looks like the ocaml>=3.11:/home/mat/work/freebsd/ports/lang/ocaml
depends line has an absolute port origin, make sure to remove \${PORTSDIR}/ from
it.""

Which ends up failing with an error:

cannot create =3.11:/home/mat/work/freebsd/ports/lang/ocaml: No such file or
directory

Noticed by:	antoine (via qat)
Sponsored by:	Absolight

 

12:06 mat 

Add a warning about using PORTSDIR in a *_DEPENDS line.

Sponsored by:	Absolight
Differential Revision:	https://reviews.freebsd.org/D6075

 

16:17 mat 

Move MySQL support from bsd.databases.mk to Uses/mysql.mk.

Also, USE_MYSQL can't happen after bsd.port.pre.mk because it is a USES.

PR:		208971
Submitted by:	mat
Exp-run by:	antoine
With hat:	portmgr
Sponsored by:	Absolight
Differential Revision:	https://reviews.freebsd.org/D5951

 

16:17 mat 

Move USE_BDB support from bsd.databases.mk to its own file.

Also:
- Deorbit WANT_BDB_VER, one can use USES=bdb:<ver> instead.
- USE_BDB can't happen after bsd.port.pre.mk because it is a USES.

PR:		208971
Submitted by:	mat
Exp-run by:	antoine
With hat:	portmgr
Sponsored by:	Absolight
Differential Revision:	https://reviews.freebsd.org/D5951

 

16:14 mat 

Remove USE_FIREBIRD from bsd.databases.mk, replaced by USES=firebird.

Also:
- replace USE_FIREBIRD by USES=firebird.
- convert ports depending on databases/firebird25-client directly to
USES=firebird.

PR:		208971
Submitted by:	mat
Exp-run by:	antoine
With hat:	portmgr
Sponsored by:	Absolight
Differential Revision:	https://reviews.freebsd.org/D5951

 

16:13 mat 

Remove USE_SQLITE from bsd.databases.mk, replaced by USES=sqlite.

While there replace USE_SQLITE=x by USES=sqlite:x.

PR:		208971
Submitted by:	mat
Exp-run by:	antoine
With hat:	portmgr
Sponsored by:	Absolight
Differential Revision:	https://reviews.freebsd.org/D5951

 

09:46 mat 

Add the messages from check-deprecated to the package +DISPLAY so that
users have a bigger change to see it.

Sponsored by:	Absolight
Differential Revision:	https://reviews.freebsd.org/D6048

 

03:28 ohauer 

- add unbound to USERS_BLACKLIST (sync with src/etc/master.passwd)

Approved by:	portmgr (mat@)

 

18:39 mat 

Extract the users/groups creation into a separate shell script.

This adds a few features:
- it checks that the UID line has the right number of fields [1]
- it tells the user that there are groups that he may want to remove [2]

PR:		208800, 173318 [1], 157546 [2]
Submitted by:	mat
Sponsored by:	Absolight
Differential Revision:	https://reviews.freebsd.org/D5939

 

21:26 mat 

Remove superfluous, non working, code.

If there is one file in PKGPOSTDEINSTALL, it gets overwritten by itself,
and if there's more than one, it'll fail.

Noticed by:	antoine
Sponsored by:	Absolight

 

15:46 swills 

Add user NOTICE for MAINTAINER=ports@

PR:		206666
Submitted by:	cperciva

 

11:00 mat 

USE_RC_SUBR=yes has not done anything for a long time, remove it from
the few remaining ports and make it give an error.

Sponsored by:	Absolight

 

15:27 jbeich 

bsd.port.mk: sync INSTALL_* comments with Porter's Handbook

PR:		197613
Approved by:	portmgr (mat)

 

10:37 jbeich 

Add CONFIGURE_OUTSOURCE support

If CONFIGURE_OUTSOURCE is defined HAS_CONFIGURE invokes configure script
and builds the port out of source tree, under ${WRKDIR}/.build by default.
The feature is inspired by USES=cmake:outsource and USES=qmake:outsource.

PR:		208294
Exp-run by:	antoine
Approved by:	portmgr (mat)
Differential Revision:	https://reviews.freebsd.org/D4157

 

10:36 jbeich 

Consistently prefer ${CONFIGURE_CMD} over ./${CONFIGURE_SCRIPT}

This also fixes SET_LATE_CONFIGURE_ARGS with custom CONFIGURE_CMD.

PR:		208294
Exp-run by:	antoine
Approved by:	portmgr (antoine), perl (mat)
Differential Revision:	https://reviews.freebsd.org/D4158

 

01:49 swills 

CentOS ports: for c6_64, install 32bit libraries alongside

PR:		206943
Submitted by:	xmj

 

13:29 mat 

Remove ${PORTSDIR}/ from dependencies, Mk and categories a, b, and c.

With hat:	portmgr
Sponsored by:	Absolight

 

13:18 mat 

Fixup some whitespace at the beginning of lines problems.

With hat:	portmgr
Sponsored by:	Absolight

 

01:23 bapt 

Remove the now unneeded ${PORTSDIR} from dependency definition in
The infrastructure Makefiles

PR:		206569
Exp run by:	antoine
Differential Revision:	D5047

 

10:59 tijl 

Revert local changes accidentally committed in previous commit.

Approved by:	portmgr (implicit)

 

10:53 tijl 

Install net/libnet headers and libraries back in the location where other
ports expect it.  The files were put in a different location to avoid a
conflict with net/libnet10 but this port has been removed a while ago now.

PR:		208122
Exp-run by:	antoine
Approved by:	portmgr (antoine)

 

12:59 mat 

Once upon a time, when you ran make package and the ${PACKAGES}
directory did not exist, you ended up with the package in the port
directory.  Make that true again.

PR:		207895
Sponsored by:	Absolight

 

16:48 mat 

Fix fetch-list.

PR:		207875
Sponsored by:	Absolight

 

17:38 bapt 

Fix the fetch-specials target when dependencies are defined without ${PORTSDIR}

 

16:49 bdrewery 

Use more proper WITHOUT_ rather than NO_ for disabling symbols.

PR:		207168
With hat:	portmgr

 

15:54 bdrewery 

As r403467 intended, also avoid splitting symbols for kernel modules.

This should fix the build of x11/nvidia-driver-340.

Reported by:	dchagin
PR:		207168
With hat:	portmgr

 

19:43 kwm 

Move Mk/bsd.gnome.mk and Mk/bsd.mate.mk to Mk/Uses/.

To use the GNOME or MATE components activate it by adding gnome
or mate to USES. The usage of USE_GNOME/INSTALL_ICONS and for
example GLIB_SCHEMAS has stayed the same.

Like with USES, the use of USE_GNOME and so after bsd.port.pre.mk
is now forbidden. And adapt ports that where still doing that.

Exp-runs done by:	antoine@

PR:		205432
Reviewed by:	antoine@, mat@
Approved by:	portmgr (antoine@)
Differential Revision:	https://reviews.freebsd.org/D3653

 

18:30 miwi 

Welcome Centos 6.7

- Upgrade all linux-c6- to CentOS 6.7
- Cleanups

PR:                     205846
Submitted by:           xmj
In Collaboration with:  allanjude, netchild, xmj
Exp-run:                antoine
Sponsored by:           Perceivon Hosting Inc.
Differential Revision:  D3428

We'd like to thanks for all the feedback and comments.

 

21:09 bdrewery 

Force DEBUG_FILES to be off when building with /usr/share/mk.  Otherwise
debug symbols are split out and extra files are generated that are not
expected by the plist.  This is after the base system enabled DEBUG_FILES
by default in r291955.

I have pending patches to handle debug symbol splitting for ports, that will
benefit all ports, not just the ones using /usr/share/mk.

With hat:	portmgr
Sponsored by:	EMC / Isilon Storage Division

 

22:18 bdrewery 

Use normal PORTSDIR rather than dp_PORTSDIR so it is passed around properly.

This fixes errors when PORTSDIR is not set and /usr/ports is not used.  The
/usr/share/bsd.port.mk logic will auto set PORTSDIR, but this was not being
passed into children in the dependency scripts.

With hat:	portmgr
Reported by:	bapt

 

09:19 danfe 

Currently, when bsd.port.mk outputs a message addressed to a user, it would
typically pipe it through ${FMT} 75 79 command to ensure that if it gets
overly long, it would be wrapped nicely.  Do the same when outputting IGNORE
message which often gets very long.  While here, add missing section number
in sed(1) references in comments.

Differential Revision:	D4116
Approved by:	portmgr (bapt)

 

15:25 marino 

Enhance "make makepatch" to address two major deficiencies

This update to the "makepatch" target adds the following enhancements:

1) Conserves comments
   If the existing patch has comments, they will be transferred to the
   regenerated patch.

2) Supports multiple patches per file
   If the patch file contains concatenated patches, the makepatch
   target will keep these patches together.  It may change the order
   of the patches the first time, but every time after the multi-patch
   will be assembled in the same order.

Behavioral changes:

A) The "old" patches are not overwritten, but rather archived at:
   ${WRKDIR}/makepatch-tmp/archived-patches

B) Any patch that was not replaced or renamed is deleted by makepatch
   (but it is archived first, see paragraph above)

C) There regeneration messages for the user will show them which patches
   are using "legacy" names formats.

D) Makepatch will do a great job at "cleaning" git patches; it removes
   lines starting with "diff" and "index" in the comments section.

Notes:

E) Should a source file be modified by multiple patches (e.g. two separate
   multi-patches), a composite patch will be generated.  In the above
   example of two multi-patches, one would get the full patch and the
   other no longer patch the source file.

Approved by:		portmgr (mat)
Differential Revision:	D4136

 

17:32 amdmi3 

Don't use cookie for test target

Test target doesn't produce anything and nothing depends on it.
Also it's often useful to run tests multiple times in a row, so
having a cookie which remembers that the tests were already ran
and inhibits subsequent test runs is needless. It also eliminates
the need for retest target.

Approved by:	portmgr (mat)
Differential Revision:	D3875

 

15:54 mat 

When !defined(DEVELOPER), stage-qa is not put in the stage pipeline.
Make it depend on stage, and point people to adding DEVELOPER=yes to
their environment.

Discussed with:	bapt
Sponsored by:	Absolight
Differential Revision:	https://reviews.freebsd.org/D4082

 

14:30 amdmi3 

Remove support for WANT_SDL/HAVE_SDL knobs

WANT_SDL/HAVE_SDL macros allowed a port to check which SDL components
are installed. This goes against the policy of avoiding automatic
dependencies, and there are actually no cases in the portstree where
these knobs are used properly, so axe them out.

Approved by:	portmgr (mat)
Differential Revision:	D4093

 

12:35 mat 

Change the meaning of NO_WRKSUBDIR to force a WRKDIR != WRKSRC.

Right now, NO_WRKSUBDIR means that the extraction does not produce a
subdirectory, and that everything goes straight into WRKDIR.  It is
problematic, because during the build of a port, quite a few files are
created in there, and then, a stage directory, where everything is
installed, and then a pkg directory where the package is created, and
those often conflict, or get in the way, of the building process.

With this, NO_WRKSUBDIR will extract the distfiles directly into WRKSRC
instead of WRKDIR.  In this case, WRKSRC is artificial and is based on
PKGNAME and not DISTNAME, mitigate conflicts with rc files.

PR:		204056
Submitted by:	mat
Reviewed by:	bapt
Exp-run:	antoine
Sponsored by:	Absolight
Differential Revision:	https://reviews.freebsd.org/D2735

 

15:30 brd 

Fix a bug introduced in r399992, where creating only a user would fail
due to $PW not being set.  This only occurred when just a user was being
created without a group.  Solve this by setting $PW outside of the group
creation and above both the user and group creation parts.

PR:		203489
Reviewed by:	bapt
Approved by:	portmgr (bapt)
Approved by:	bdrewery (mentor)

 

16:25 brd 

Teach the ports framework to handle creating users/groups when
PKG_ROOTDIR is set.  This will enable help support pkg cross installs.

PR:		203489
Reviewed by:	bapt
Approved by:	portmgr (bapt)
Approved by:	bdrewery (mentor)

 

21:49 bdrewery 

Revert r399921 for now due to fallout with libclc using ninja.

 

14:43 bdrewery 

Enable verbose (compiler output) builds for autotools builds when package
building as these commands can be critical for debugging build failures.

Discussed with:	bapt, kwm
With hat:	portmgr

 

19:23 bdrewery 

- Combine clean-depends-list.sh into depends-list.sh
- Refactor how depends-list.sh is called from bsd.port.mk for each variant.

With hat:	portmgr

 

18:37 bdrewery 

Fix test-depends after r399703.

With hat:	portmgr

 

07:36 bapt 

Readd PORTSDIR for now we will only start removing them after 2016Q1 is branched

This gives more time for tools to get updated, available in packages etc before
bothering users

 

16:49 bapt 

Drop the necessity to add ${PORTSDIR} to dependency line

Modify make describe to automatically prepend ${PORTSDIR} if the path for the
port is not absolute

Checked with poudriere, portmaster, portupgrade

PR:		203685
Exp-run by:	antoine
Differential Revision:	https://reviews.freebsd.org/D3866

 

23:22 bdrewery 

Don't check for minimal Pkg version for package building.

Package building always builds Pkg first.  If someone downgrades their
pkg locally this change will hurt them, but then they are already
running an unsupported configuration.

With hat:	portmgr

 

22:44 bdrewery 

Name the 'uname -r' variable the same as the bsd.port.subdir.mk version,
_OSRELEASE.

This will allow passing in the environment consistently from calling script
to avoid looking it up if.

With hat:	portmgr

 

22:41 bdrewery 

Export more sysctl(8) vars.

LINUX_OSRELEASE, _SMP_CPUS (used as default MAKE_JOBS), CONFIGURE_MAX_CMD_LEN.

With hat:	portmgr

 

22:30 bdrewery 

Remove debugging leftover from r399160.

With hat:	portmgr

 

22:28 bdrewery 

Export HAVE_COMPAT_IA32_KERN to sub-makes and ensure it is not exported if
empty.

With hat:	portmgr

 

22:19 bdrewery 

Move the exported variable assignments to where they are looked-up.

This is so we are less likely to carry around old vars that are no longer
needed and to keep logic more coupled.

With hat:	portmgr

 

22:17 bdrewery 

Spell 'OSREL' properly from r399155.

With hat:	portmgr

 

22:14 bdrewery 

Don't check for minimal pkg version in sub-makes.

With hat:	portmgr

 

22:13 bdrewery 

Pass result of 'uname -r' to sub-makes to avoid unneeded lookup.

With hat:	portmgr

 

22:09 bdrewery 

Use a loop to export vars.

This also fixes adding duplicates to .MAKEFLAGS with fmake, which was a
bug present before recent changes.

With hat:	portmgr

 

21:41 bdrewery 

Re-apply r398953 with protection for 'make *config*'.

This will only include the descriptions file if needed to very slightly
lower overhead with -V lookups and package building dependency calculations.

With hat:	portmgr

 

21:38 bdrewery 

Replace escaping for exported vars with :Q feature.

With hat:	portmgr

 

21:36 bdrewery 

Remove SYSTEMVERSION exporting, not needed since r99913.

This is a relic of port.mkversion.

With hat:	portmgr

 

21:19 bapt 

Avoid a useless call to dirname