Do not overflow the input buffer. This happens on slow hosts that take
long time to recompute the data on start.
Signed-off-by: Paul Fertser <fercerpav@gmail.com>
git-svn-id: svn://svn.openwrt.org/openwrt/packages@36390 3c298f89-4303-0410-b956-a3cf2f4a3e73
When the outside temperature is less than -10C the station reports that
dewpoint is invalid. This patch allows to ignore only that value instead
of ignoring all of the packet contents from the outside sensor (including
temperature and humidity).
This is also sort of fixed (the reported dew point values might be
invalid) in upstream release 5.20.0, but we can't use his repository
because there is no public repository and also he ignores requests on
his semi-private bugtracker; so we need to use the source we can
trust, and currently jgoerzen's branch is the best source out there.
Signed-off-by: Paul Fertser <fercerpav@gmail.com>
git-svn-id: svn://svn.openwrt.org/openwrt/packages@36389 3c298f89-4303-0410-b956-a3cf2f4a3e73
make the syntax more compatible with kernel menuconfig
Signed-off-by: Felix Fietkau <nbd@openwrt.org>
git-svn-id: svn://svn.openwrt.org/openwrt/packages@36357 3c298f89-4303-0410-b956-a3cf2f4a3e73
zsh would use gdbm if available, failing packaging checks.
Signed-off-by: Francisco Borges <francisco.borges@gmail.com>
Signed-off-by: Florian Fainelli <florian@openwrt.org>
git-svn-id: svn://svn.openwrt.org/openwrt/packages@36106 3c298f89-4303-0410-b956-a3cf2f4a3e73
A new version 1.10 of ccrypt encryption tools package has been published in
late 2012. The old version 1.9 dates back to 2009.
Link to changelog: http://ccrypt.sourceforge.net/ChangeLog
October 18, 2012: Release 1.10. This release fixes a number of minor
portability issues and minor bugs. The ccrypt core functionality has been
factored into a convenience library libccrypt.a. A ccguess program was added
to assist in the recovery of mistyped keys. Improved error checking and fixed
minor compiler warnings. Minor improvements to emacs support, and added a
--disable-emacs configuration option. Thanks to bornlibra23 and Johannes
Ruscheinski for bug reports.
I have tested the version 1.10 on ar71xx/wndr3700.
A patch to update ccrypt to 1.10 is included.
Signed-off-by: hannu.nyman@iki.fi
Signed-off-by: Gabor Juhos <juhosg@openwrt.org>
git-svn-id: svn://svn.openwrt.org/openwrt/packages@35635 3c298f89-4303-0410-b956-a3cf2f4a3e73
Installs the shell binary, and all applicable completion functions.
Completion functions specific for other systems (e.g. AIX/Solaris/Redhat/etc)
are not installed.
Compiled against ncurses. Notice the "--with-term-lib=..." at Build/Configure
forcing the preference for ncurses (against ncursesw). To build with ncursesw,
(i) change the dependencies, (ii) remove the '--with-term-lib=...' (iii) add
CPPFLAGS=-I$(STAGING_DIR)/usr/include/ncursesw to the configure command.
Signed-off-by: Francisco Borges <francisco.borges@gmail.com>
git-svn-id: svn://svn.openwrt.org/openwrt/packages@35474 3c298f89-4303-0410-b956-a3cf2f4a3e73
This patch updates less to 457.
Signed-off-by: Peter Wagner <tripolar@gmx.at>
Signed-off-by: Florian Fainelli <florian@openwrt.org>
git-svn-id: svn://svn.openwrt.org/openwrt/packages@35231 3c298f89-4303-0410-b956-a3cf2f4a3e73
The uClibc++ implementation of stringstream operator>> is broken which
fails boblightd at runtime when parsing the configuration.
Due to the failed parsing, std::list unique() is called with an empty
list which causes the program to trigger a segmentation fault, this
is most likely another uClibc++ implementation error.
git-svn-id: svn://svn.openwrt.org/openwrt/packages@35027 3c298f89-4303-0410-b956-a3cf2f4a3e73
Since a8801d8665bcc94c63a798e291a03c88f2af0215 upstream commit
flash_eraseall is a shell wrapper around flash_erase.
Signed-off-by: Paul Fertser <fercerpav@gmail.com>
git-svn-id: svn://svn.openwrt.org/openwrt/packages@35004 3c298f89-4303-0410-b956-a3cf2f4a3e73
This almost caused a nasty accident on my Geos — without the MSR access
it managed to erase *most* of the flash and write the new image, but
didn't manage to change the last few KiB. Thankfully I'd already taken a
copy of the old contents, and could flash them back.
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Signed-off-by: Gabor Juhos <juhosg@openwrt.org>
git-svn-id: svn://svn.openwrt.org/openwrt/packages@34832 3c298f89-4303-0410-b956-a3cf2f4a3e73
The apcups and snmp plugins were using this to store time in minutes, but
nut uses seconds. Upstream wants to make them all use seconds (unlike my
first patch to make nut use minutes to be consistent with the others), so
all we need to do is fix the maximum. If we were *using* the apcups and
snmp plugins, those would probably want fixing to multiply by 60.
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Signed-off-by: Gabor Juhos <juhosg@openwrt.org>
git-svn-id: svn://svn.openwrt.org/openwrt/packages@34831 3c298f89-4303-0410-b956-a3cf2f4a3e73
This works now...
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Signed-off-by: Gabor Juhos <juhosg@openwrt.org>
git-svn-id: svn://svn.openwrt.org/openwrt/packages@34830 3c298f89-4303-0410-b956-a3cf2f4a3e73
There's more to be done here, but this is a good start and covers most
things that people will need to configure.
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Signed-off-by: Gabor Juhos <juhosg@openwrt.org>
git-svn-id: svn://svn.openwrt.org/openwrt/packages@34828 3c298f89-4303-0410-b956-a3cf2f4a3e73