Jo-Philipp Wich [Wed, 17 Jun 2015 17:27:37 +0000 (19:27 +0200)]
xinetd: fix musl compatibility
Pass HAVE_RLIM_T via TARGET_CPPFLAGS since configure uses that define
but never actually declares it.
Without doing that, `config.h` tries to declare `rlim_t` itself which
leads to `config.h:126:16: error: 'long long long' is too long for GCC`
with musl on at least x86_64.
Yousong Zhou [Wed, 17 Jun 2015 13:37:24 +0000 (21:37 +0800)]
libpam: bump to 1.2.0
- Add configure options --disable-nis, --disable-regenerate-docu
- 000-OE-libpam-xtests.patch is not relevant in OpenWrt
- 001-no_nis.patch was dropped because we now --disable-nis
- 002-no_yywrap.patch was dropped be cause it was fixed in 1.2.0
- 003-no_doc was dropped because we ignore doc/ with
--disable-regenreate-docu
- 004-fix_lib64 was replaced by new 0001-build-use-host_cpu...
- pam_rhosts will not be built with musl because ruserok{,_af{
are not available
- pam_lastlog will not be built with musl because logwtmp is missing
[sqm-scripts/luci-app-sqm] Document how to disable shaping on a per direction basis
sqm-scripts for a long time interprets a "Down- or Upload speed" of zero as
an indication that the shaper should be disabled. Note that really shaping
an individual direction down o zero will make the link effectively dead
for tcp (think reverse ACK traffic). Son instead of allowing the user to
configure something broken, 0 was "over-loaded" to denote no shaping
since several years, but that information has not been documented visibly
to the users. This commit aims at fixing that oversight.
Jo-Philipp Wich [Tue, 16 Jun 2015 22:10:05 +0000 (00:10 +0200)]
pptpd: fix musl compatibility
- Ship a net/ppp_defs.h replacement header since musl does not provide one
but pppd/pppd.h provided by pppd is needing it.
- Replace deprecated MSG_TRYHARD flag with the proper MSG_DONTROUTE one
- Unconditionally include string.h in bcrelay to avoid undeclared memset()
and strcpy() warnings.
Jo-Philipp Wich [Tue, 16 Jun 2015 14:32:21 +0000 (16:32 +0200)]
aiccu: fix musl compatibility
- Fix redefinition of struct ethhdr
- Restrict the use of res_ninit() and res_nquery() to glibc
- Fix a warning in dn_skipname.c by declaring ns_name_skip() static
- Refresh patches
heil [Thu, 11 Jun 2015 17:03:14 +0000 (19:03 +0200)]
haproxy: add patches from upstream
- [PATCH 1/2] BUG/MEDIUM: stats: properly initialize the scope before
- [PATCH 2/2] BUG/MEDIUM: http: don't forward client shutdown without
- [PATCH 3/8] BUG/MINOR: check: fix tcpcheck error message
- [PATCH 4/8] CLEANUP: checks: fix double usage of cur / current_step
- [PATCH 5/8] BUG/MEDIUM: checks: do not dereference head of a
- [PATCH 6/8] CLEANUP: checks: simplify the loop processing of
- [PATCH 7/8] BUG/MAJOR: checks: always check for end of list before
- [PATCH 8/8] BUG/MEDIUM: checks: do not dereference a list as a
- [PATCH 09/10] BUG/MEDIUM: peers: apply a random reconnection timeout
- [PATCH 10/10] DOC: Update doc about weight, act and bck fields in the
- [PATCH 11/14] MINOR: ssl: add a destructor to free allocated SSL
- [PATCH 12/14] BUG/MEDIUM: ssl: fix tune.ssl.default-dh-param value
- [PATCH 13/14] BUG/MINOR: cfgparse: fix typo in 'option httplog' error
- [PATCH 14/14] BUG/MEDIUM: cfgparse: segfault when userlist is misused
Daniel Golle [Tue, 9 Jun 2015 01:05:53 +0000 (03:05 +0200)]
lxc: fix build on mpc85xx
Initialize ret to 0 so compiler no longer complains about
monitor.c: In function 'lxc_monitor_open':
monitor.c:212:5: error: 'ret' may be used uninitialized in this function [-Werror=maybe-uninitialized]
Florian Fieber [Fri, 5 Jun 2015 14:07:56 +0000 (16:07 +0200)]
vim: Fix editing crontabs with crontab -e
Per default, vim creates a new file and keeps the original file as backup.
This leads to crontab checking the backup file for changes instead of the
one actually being edited, resulting in any changes being silently dropped.
Since busybox's crontab saves temporary crontab files in "/etc/crontabs.*"
vim's default of "backupskip=/tmp/*" doesn't circumvent this.
This commit fixes this issue by appending "/etc/crontabs.*" to "backupskip".
Daniel Golle [Fri, 5 Jun 2015 15:11:07 +0000 (17:11 +0200)]
xl2tpd: fix xl2tpd options
The previous commit left an unwanted string in the options passed
to xl2tpd-control. Remove that as well.
Also set procd respawn parameter now that I'm at it.
Daniel Golle [Fri, 5 Jun 2015 02:22:23 +0000 (04:22 +0200)]
gnurl: no longer force-enable TLS-SRP
It might not be available in GnuTLS and is not strictly required
by the gnurl manifest, see
http://www.git.taler.net/?p=gnurl.git;a=blob;f=README;h=f7521aa431a4e77a58dc558880ff54648d69e072;hb=HEAD#l22
Daniel Golle [Thu, 4 Jun 2015 23:27:52 +0000 (01:27 +0200)]
gnunet: don't override default to 'n'
Overriding DEFAULT to 'n' prevents packages being build unless
explicitely selected even if ALL is set.
Instead, set 'DEFAULT:=y if PACKAGE_gnunet' for the default
packages and otherweise don't touch DEFAULT.
Daniel Golle [Thu, 4 Jun 2015 12:44:40 +0000 (14:44 +0200)]
libseccomp: work-around prctl.h mess
musl doesn't allow including both, sys/prctl.h and linux/prctl.h.
Thus include linux/prctl.h only on GLIBC systems (that includes
uClibc which defines __GLIBC__)
Daniel Golle [Thu, 4 Jun 2015 11:46:32 +0000 (13:46 +0200)]
lxc: allow building with seccomp
seccomp was unconditionally disabled before.
Also a non-escaped newline in a place where it should be escaped
made the build process skip some CONFIGURE_ARGS.
Fix CONFIGURE_ARGS and add an option to allow building with seccomp.
git.99rst.org / openwrt-packages.git / log