This changes the buildlink3.mk files to use an include guard for the
recursive include. The use of BUILDLINK_DEPTH, BUILDLINK_DEPENDS,
BUILDLINK_PACKAGES and BUILDLINK_ORDER is handled by a single new
variable BUILDLINK_TREE. Each buildlink3.mk file adds a pair of
enter/exit marker, which can be used to reconstruct the tree and
to determine first level includes. Avoiding := for large variables
(BUILDLINK_ORDER) speeds up parse time as += has linear complexity.
The include guard reduces system time by avoiding reading files over and
over again. For complex packages this reduces both %user and %sys time to
half of the former time.
* manpage spelling fixes, from Claus Assmann <ca+tinycdb (at) esmtp.org>.
* little mods to allow compiling tinycdb by C++ compiler,
from Olly Betts <olly (at) survex.com>.
* use program_invocation_short_name on GLIBC, (modified) patch
from Dmitry V. Levin <ldv (at) altlinux.org>
* manpage fix (cdb_findnext() prototype),
from Dmitry V. Levin <ldv (at) altlinux.org>
* (somewhat silly) GCC-4.x "signedness" warnings fix, modified patch
from Dmitry V. Levin <ldv (at) altlinux.org>
* more signed vs unsigned char* fixes in various places
* Makefile: always build libnss_cdb.so with libcdb_pic.a, no nss-shared
target: to avoid extra dependency from /usr/lib/.
* Makefile: use map files for lib*.so, with explicit list of exported
symbols. This, in particular, avoids exporting of libcdb symbols by
libnss_cdb.so.
* mark all internal routines as internal_function (defined as
__attribute__((visibility("hidden"))) for GCC)
* Makefile: add tests-shared, to use cdb-shared for testing
* Makefile: allow to specify which binary (shared vs static) to install
in install target, by using INSTALLPROG variable
* Makefile: pass -DNSSCDB_DIR=... to compiler when building .lo files,
to allow setting of system config dir (/etc by default) on command line.
For nss_cdb module.
* Makefile: use $(CP) instead of cp, to be able to specify `cp' options
(CP="cp -p")
* Use unlink(tmpname) + open(O_CREAT|O_EXCL) instead of
open(O_CREAT|O_TRUNC) when creating the new CDB file.
And use O_NOFOLLOW if defined.
This also works around some (probably) possible symlink attacks.
* Add -p perms option for cdb utility, to specify mode (permission bits)
for newly created .cdb file (default is 0666&umask(), -p forces the given
mode).
* allow tmpname (cdb -t) to be `-', to mean no temp file gets created.
Also check if tmpname is the same as database name and don't try to
rename() if it is.
* rewrite nss_cdb-Makefile a bit: simplify it, and use more sane
permission scheme for /etc/shadow
* fixed a typo in cdb_seek.c, -- it segfaulted if passed NULL dlenp
pointer. Thanks Daiki for the patch. Closes: #383417
* use MAP_FAILED to test mmap() return value, instead of hardcoded -1.
* libcdb-dev replaces tinycdb<0.76
and add a new helper target and script, "show-buildlink3", that outputs
a listing of the buildlink3.mk files included as well as the depth at
which they are included.
For example, "make show-buildlink3" in fonts/Xft2 displays:
zlib
fontconfig
iconv
zlib
freetype2
expat
freetype2
Xrender
renderproto
RECOMMENDED is removed. It becomes ABI_DEPENDS.
BUILDLINK_RECOMMENDED.foo becomes BUILDLINK_ABI_DEPENDS.foo.
BUILDLINK_DEPENDS.foo becomes BUILDLINK_API_DEPENDS.foo.
BUILDLINK_DEPENDS does not change.
IGNORE_RECOMMENDED (which defaulted to "no") becomes USE_ABI_DEPENDS
which defaults to "yes".
Added to obsolete.mk checking for IGNORE_RECOMMENDED.
I did not manually go through and fix any aesthetic tab/spacing issues.
I have tested the above patch on DragonFly building and packaging
subversion and pkglint and their many dependencies.
I have also tested USE_ABI_DEPENDS=no on my NetBSD workstation (where I
have used IGNORE_RECOMMENDED for a long time). I have been an active user
of IGNORE_RECOMMENDED since it was available.
As suggested, I removed the documentation sentences suggesting bumping for
"security" issues.
As discussed on tech-pkg.
I will commit to revbump, pkglint, pkg_install, createbuildlink separately.
Note that if you use wip, it will fail! I will commit to pkgsrc-wip
later (within day).
creating and reading constant data bases, a data structure introduced
by Dan J. Bernstein in his cdb package. It may be used to speed up
searches in a sequence of (key,value) pairs with very big number
of records. Example usage is indexing a big list of users - where
a search will require linear reading of a large /etc/passwd file,
and for many other tasks. It's usage/API is similar to ones found
in BerkeleyDB, gdbm and traditional *nix dbm/ndbm libraries, and
is compatible in great extent to cdb-0.75 package by Dan Bernstein.
CDB is a constant database, that is, it cannot be updated at a
runtime, only rebuilt. Rebuilding is atomic operation and is very
fast - much faster than of many other similar packages. Once created,
CDB may be queried, and a query takes very little time to complete.
