Age | Commit message (Collapse) | Author | Files | Lines |
|
Bump to 0.105
|
|
Now works on Solaris - fixes pkg/16145 from Darrin B. Jewell
|
|
When using -f with -i, run fetch for mismatched packages
|
|
Report downlevel pkg_install version, too (in -i mode).
Requested by gendalia and agc.
|
|
Fix stupid bug in 1.25 relating to updating packages
|
|
'Makefile' is back in BUILD_VERSION, so add it back in here.
Implement '-f' based on patches from norm@sandbox.org.uk - perform a
fetch for all missing/mismatched packages before any building. Can be
used with -c or -i to just fetch missing/mismatched packages.
|
|
Handle updated format of +BUILD_VERSION - sometimes patches are recorded
with paths, sometimes not. Ideally we would have a show-build-version
target in bsd.pkg.mk which could be sued by us, but that doesn't help
checking against existing installations where nothing has changed except
the way +BUILD_VERSION is formatted.
|
|
change the line number in the configure script.
Mostly by Stoned Elipot, with slight changes by me.
Bump to 0.104.
|
|
Changes summary:
- Support digital signatures in binary packages.
- Use full path name to tar.
- Fix ``pkg_add -u'' for packages without dependencies.
- Add ``-U'' and ``-I'' options to pkg_create.
- Close some memory leaks.
- Add ``-n'' argument to pkg_info to show needed packages.
- Fix pkg_info to not FTP unneccessarily
- Improve version number handling:
- recognition of "pl" ("patchlevel") and "rc" ("release candidate")
strings
- recognition of '_' and "pl" as pseudonyms for '.' (1.2pl2 == 1.2.2)
- handle alphabetic characters properly (1.2e == 1.2.5)
- 64-bit integers are used internally for each component of the version
number.
- Various internal cleanups, bugfixes, and API tweaks.
|
|
Avoid using '..' in LOCALBASE as it might end up in the binaries as part of
TAR_FULLPATHNAME.
Record MANINSTALL setting.
Add a workaround for SHLOCK definition or else LOCALBASE setting breaks it.
|
|
needed anymore, manpage related entries added to PLIST.
|
|
Work around chicken-and-egg problem: 'make makesum' seems to need a
'distinfo' file these days.
|
|
Correct inverted logic for DYNAMIC_MASTER_SITES check
|
|
scripts so that they are properly added and removed if xpkgwedge is
installed as a binary package. Also, add some protection against the case
where ${X11BASE} == ${PREFIX} by checking this isn't true before removing
these font-manipulation programs.
|
|
Unconditionally set PKGSRC_LOCKTYPE to `none' to avoid bootstrapping problem.
Do not generate formatted manual and do not install source manual on Solaris or
else we would need to make this package depends on textproc/groff. IMHO this
should be avoided for this package as it can be considered a prerequisite for
building further packages.
|
|
Handle DYNAMIC_MASTER_SITES
|
|
finish a new one.
|
|
into the equivalent INSTALL_*. This is fallout from the change in
revision 1.915 that removed ${MAKE_ENV} from the environment for a
recursive make.
|
|
include/X11/fonts/fontenc.h
include/X11/extensions/xtrapbits.h
include/X11/extensions/xtrapddmi.h
include/X11/extensions/xtrapdi.h
include/X11/extensions/xtrapemacros.h
include/X11/extensions/xtraplib.h
include/X11/extensions/xtraplibp.h
include/X11/extensions/xtrapproto.h
lib/libXrender.so.1.1
lib/libXTrap.a
lib/libXTrap.so
lib/libXTrap.so.6
lib/libXTrap.so.6.4
lib/libXinerama.so
lib/libXinerama.so.1
lib/libXinerama.so.1.0
OKed by jlam
|
|
|
|
|
|
NetBSD-current, into the NetBSD Packages Collection.
The shlock command can create or verify a lock file on behalf of a
shell or other script program. When it attempts to create a lock
file, if one already exists, shlock verifies that it is or is not
valid. If valid, shlock will exit with a non-zero exit code. If
invalid, shlock will re- move the lock file, and create a new one.
shlock uses the rename(2) system call to make the final target lock
file, which is an atomic operation (i.e. "dot locking", so named for
this mech- anism's original use for locking system mailboxes). It
puts the process ID ("PID") from the command line into the requested
lock file.
shlock verifies that an extant lock file is still valid by using
kill(2) with a zero signal to check for the existence of the process
that holds the lock.
This package is only for Solaris and Linux platforms at present.
|
|
|
|
New feature:
Somewhat rewritten to support run from top-level pkgsrc directory. When run
from there pkgcvsupdate launchs a separate cvs update for each caterory
directory. This is believed to be faster than a single cvs run. Thanks to
Simon Burge for this!
Bugfix:
First update a category Makefile (by way of a non-recursive update in the
category directory) before finding out which packages are included in it.
|
|
- lintpkgsrc - fixed -p to handle relative symlinks and symlink loops
|
|
|
|
Adapt to 'distinfo' change.
Adapt to 'pkg' directory removal.
Automatically rename 'pkg-message' files.
Fix a problem introduced in 1.4 (which perl -w warned about).
|
|
Do not abort if pkg_delete returns non zero exit status in '-Br'
|
|
Patch sent by Amitai Schlair <schmonz@schmonz.com> in private mail.
|
|
packages, since certain arguments passed to the tar command by pkg_create
assume the existence of GNU tar (-T for one), and Darwin's standard tar
doesn't understand these.
|
|
When comparing BUILD_VERSION, we _really_ do not care what PKGSRCDIR was
|
|
lintpkgsrc - if we find one valid match in a DEPENDS entry, don't
bitch about any others expanding to a non-existant package:
Example: postgresql-{6.5.3*,7.0*,lib-*}
With this lintpkgsrc -d now outputs six lines on current pkgsrc.
Note to self: When you use 'cvs c^[P' the second time in an
evening for the same package... its enough.
|
|
lintpkgsrc.pl:
Finally we parse {x,y,z} constructs in DEPENDS, plus handle a
few more make conditionals. Now most of the lintpkgsrc -d output
is related to versions of software no longer in pkgsrc...
Does the fun ever start?
|
|
Significant rototill of lintpkgsrc.pl, converting to a rather twisted
hierarchy of PkgVer Pkgs and PkgList objects with assorted arbitrary
methods, all based on existing usage rather than ideal design.
Actual fixes found along the way:
- Handle := in makefiles
- Better handle parenthesis and "" in make conditionals
- parse bsd.pkg.mk to extract _REQD values (hence the previous line)
- Add missing 'key' in hash iteration
- Do not repeat includes (evolution longer depends on jpeg 45 times)
Still have to handle such delightful constructs as
DEPENDS += {qt2-designer>=2.2.4,qt2-designer-kde>=2.3.1nb1} in a sane way
|
|
|
|
Implement '-C conf' for specifying alternate config file
|
|
x11-links/buildlink.mk file.
|
|
Reduces warnings from 'lintpkgsrc -d' from ~700 lines to ~100.
[It won't get much better than this...]
|
|
properly on Darwin.
|
|
|
|
references of the pkglint package.
_PKGSRCDIR is an internal definition in bsd.pkg.mk, and a few packages
which would like to refer to other packages in the build tree. It should
not be set by users, but neither should it stop a user from building a
package if it is defined, so make it obvious that this is the case.
|
|
but "lintpkgsrc -d" output reduced from >7000 lines to ~700.
First part of diff fixes most ruby* complaints, second part the general
"Unknown package: 'UNDEFINED' version missing" message.
|
|
OK:ed by jlam.
|
|
only take 9 arguments'
Fix by wiz@netbsd.org (thanks!
Thanks for Thomas Klausner for the idea!
|
|
Bump to 3.16.
|
|
|
|
- lintpkgsrc: Set BSD_PKG_MK, and ensure .CURDIR is correct when running
with -D. Also remove an articulated lorry load of single quotes.
Many thanks to wiz who added PKGREVISION before I lamely managed to get
around to it :)
|
|
Bump to 3.14 (next version 3.141? ;)
|
|
as first argument to the editor.
|
|
Nearly a complete rewrite in order to deal with the following situation:
a package's Makefile include a Makefile fragment (like a buildlink.mk file)
which does not exist, so `$MAKE show-depends-dirs' cannot be run. To resolve
this pkgcvsupdate now try to figure out which package provide the missing file
and update it first (and its dependencies), it will update a second time
the crippled package in case the fragment - hopefully now in place - brings
along new dependencies.
Thanks to zuntum for this interesting case I overlooked!
|