Re: Important bgpd fix

By the way, that bgpd diff is in the snaps.

Re: Important bgpd fix

On Sat, Jan 11, 2014 at 03:07:37PM +, Stuart Henderson wrote: > I'm running this on one router without seeing any problems yet, > however it does not have any graceful-restart peers so it's not exactly > a great test. > > Has anyone else tried this at all yet? Benno is running it on one of ou

Re: [PATCH] nginx.conf(5): Less EXAMPLES, more HISTORY

Hi Donovan, Donovan Watteau wrote on Sat, Jan 11, 2014 at 04:18:59PM +0100: > OpenBSD now provides an nginx.conf(5) manual page and I love this, > thanks! Florian, who did all the work on the formatting, will no doubt enjoy that statement. :) > But it ends with an EXAMPLES section which only d

[PATCH] nginx.conf(5): Less EXAMPLES, more HISTORY

OpenBSD now provides an nginx.conf(5) manual page and I love this, thanks! But it ends with an EXAMPLES section which only deals with some horrible use of whitespace, which is "not recommended". I don't get it. This manpage is a relief, but it ends in a really sad and not really useful way... I

Re: Important bgpd fix

I'm running this on one router without seeing any problems yet, however it does not have any graceful-restart peers so it's not exactly a great test. Has anyone else tried this at all yet? On 2014/01/02 00:03, Claudio Jeker wrote: > There is a somewhat critical bug in bgpd which got hit by local

update libdrm to 2.4.51

Update libdrm to 2.4.51. Needs the previous drm header diff. Most of the changes concern hardware we don't support (Intel hardware not publically available and the latest radeons) but tests welcome. Index: Makefile.inc === RCS file:

update kernel drm headers for libdrm 2.4.51

Update the drm headers based on libdrm 2.4.51 but make sure we still reject DRM_MODE_PAGE_FLIP_ASYNC. Index: drm.h === RCS file: /cvs/src/sys/dev/pci/drm/drm.h,v retrieving revision 1.13 diff -u -p -r1.13 drm.h --- drm.h 17 Nov

small fix for pkg_add(1)

$ man pkg_add [...] -D name[=value] [...] downgradedon't filter out package versions older than what's currently installed. D FW_UPDATE set by fw_update(1) to separate firmwares from normal p

Re: boot: readline off by one

On Sat, Jan 11, 2014 at 12:43:33PM +0100, Tobias Stoeckmann wrote: > p[1] = *p = '\0'; If this is actually intended behaviour, it's off from read_conf, which ends cmd_buf with just one terminating \0 character.

boot: readline off by one

Hi, in boot, we have an off-by-one error in readline. When the user ends input with enter, the string will be ended twice, like: p[1] = *p = '\0'; Therefore we have to make sure that two bytes are still free, not just one. Not sure why it has to be handled like this, but the fix is easy to