[...@spacehopper.org: relayd: allow a fallback table for relays]

2010-07-02 Thread Stuart Henderson
I've been running with most of this diff for ages, just got round to adding the pfe.c part to fix relayctl show ho / show su output. Any comments, oks? Index: parse.y === RCS file: /cvs/src/usr.sbin/relayd/parse.y,v retrieving

WANTLIB/LIB_DEPENDS semantics change

2010-07-02 Thread Marc Espie
The following patch changes the semantics of LIB_DEPENDS/WANTLIB, hopefully to make things simpler for ports maintenance. Historically, all libs were registered in LIB_DEPENDS. In order to handle updates, packages had to register all needed libraries. The current mechanism is that a package

Re: WANTLIB/LIB_DEPENDS semantics change

2010-07-02 Thread Marc Espie
Faster patch... this removes a lot of old cruft, and yields a _print-package-args that's about twice as fast, instead of potentially slower. Specifically, always grab lib lists off the ports tree proper, evaluate them once, and use the fact that resolve-lib has known how to handle multiple

Re: WANTLIB/LIB_DEPENDS semantics change

2010-07-02 Thread Marc Espie
On Fri, Jul 02, 2010 at 04:33:43PM +0200, Marc Espie wrote: Faster patch... this removes a lot of old cruft, and yields a _print-package-args that's about twice as fast, instead of potentially slower. Turns out it must be slightly more complicated. Namely, we have to bring LIB_DEPENDS

Call for testing: IPsec diff

2010-07-02 Thread Reyk Floeter
Hi! I need people to test the following IPsec diff on existing setups running -current. This diff will add some cool features for the next release but I first need regression testing with plain old setups (ipsec.conf with static keying or isakmpd); preferrably on IPsecs that are running closely

Trouver des professionnels pour vos travaux

2010-07-02 Thread Philippe
Si vous ne pouvez pas lire cet email, suivre ce lien (http://front.datafnxjune25.com/php/emailing/view_mail.php?CODE=586BBGXG_416219HASH=df0fe6426bc7f9674e80137ab1433bd7) Untitled Document Fini les traveaux compliquis

Re: Call for testing: IPsec diff

2010-07-02 Thread Reyk Floeter
Hi, updating one side should be enough. reyk On Sat, Jul 03, 2010 at 01:15:50AM +0400, Vadim Zhukov wrote: 2010/7/3 Reyk Floeter r...@openbsd.org: Hi! I need people to test the following IPsec diff on existing setups running -current. ??This diff will add some cool features for the

Re: WANTLIB/LIB_DEPENDS semantics change

2010-07-02 Thread Marc Espie
Hopefully final patch... moving this up from make to sh means I must escape a bit more stuff... Index: bsd.port.mk === RCS file: /home/openbsd/cvs/ports/infrastructure/mk/bsd.port.mk,v retrieving revision 1.1007 diff -u -p -r1.1007

Re: WANTLIB/LIB_DEPENDS semantics change

2010-07-02 Thread Stuart Henderson
On 2010/07/02 23:32, Marc Espie wrote: Hopefully final patch... moving this up from make to sh means I must escape a bit more stuff... people who are confused might find this easier to understand with an example: the diff below results in no packing list change, it's equivalent to what is there

Re: Enable speedstep on nehalem (i3/5/7) cpus

2010-07-02 Thread Jonathan Gray
On Sat, Jun 26, 2010 at 11:48:15PM -0600, Gordon Willem Klok wrote: Hello this should make speedstep work on nehalem cpu's (i3, i5, i7s) where est uses acpi to get the available states. Please let me know of any issues This is not changing the places where it interprets PERF_STATUS on