I took a look at the remaining failures to build 7.8 in Xenofarm. There are only 2 of them left:
"localhost"*: Runs out of memory when exporting documentation. Not much of an issue.
"tomoyo": Fails to build Nettle because apparently the libnettle bundle requires m4 to build.
Now m4 isn't a very burdensome requirement, but is it supposed to be there for exorted builds? If so I'll just ask the owner of tomoyo to install it, but usually m4 tasks are taken care of when exporting a source package.
* Awsome name. I must take a look at the node name generation in the client again.
On Thu, Jul 09, 2009 at 10:05:02AM +0000, Peter Bortas @ Pike developers forum wrote:
"tomoyo": Fails to build Nettle because apparently the libnettle bundle requires m4 to build.
that would be because it is using the bundled nettle 2.0 now, is it?
because i remember that i fixed that before by packaging nettle 1.5 m4 is installed now., thanks for the notice!
greetings, martin.
On Thu, Jul 09, 2009 at 10:05:02AM +0000, Peter Bortas @ Pike developers forum wrote:
"tomoyo": Fails to build Nettle because apparently the libnettle bundle requires m4 to build.
that would be because it is using the bundled nettle 2.0 now, is it?
I shouldn't think so. The snapshot still bundles Nettle 1.15. If it's fixed in 2.0 it would be one reason to upgrade, but for lack of obvius advantages in upgrading I'm holding off on the 2.0 train until it's "stabilized" a bit.
On Thu, Jul 09, 2009 at 10:50:29AM +0000, Peter Bortas @ Pike developers forum wrote:
that would be because it is using the bundled nettle 2.0 now, is it?
I shouldn't think so. The snapshot still bundles Nettle 1.15.
oh, i thought i read that grubba committed changes to make 2.0 work...
in that case something went wrong with the detection of a system nettle... -> turns out that aparently nettle:devel is no longer on my system. fixing...
On Thu, Jul 09, 2009 at 10:05:02AM +0000, Peter Bortas @ Pike developers forum wrote:
"localhost"*: Runs out of memory when exporting documentation. Not much of an issue.
strange. the machine(*) has 500MB and is not running any large applications, actually pikefarm is the only thing that really excercises the machine. and i think it worked before.
- Awsome name. I must take a look at the node name generation in the
client again.
that is really wierd, because the hostname was detected correctly before, and no configuration was changed on the machine. hostname; still returns 'nessie' and gethostname() too.
* yes, 'localhost' is my machine too :-)
ah, right, and 64bit needs more space than 32. i raised the data segment and virtual memory values by 50% let's see if that is enough.
thanks for the hint!
greetings, martin.
On Fri, Jul 10, 2009 at 05:20:24AM +0200, Martin Baehr wrote:
On Thu, Jul 09, 2009 at 09:55:02PM +0000, Peter Bortas @ Pike developers forum wrote:
What does "client.sh --nodename" say?
nessie
wait, that's wrong, it comes out with nessie when i run it as root, but not when i run it as a user. huh?
greetings, martin.
Not strange, but unfortunate. If you are root the result will always be uname -n, together with a warning that you shouldn't run the script as root. If not a number of other methods will be used to to determine the host name. The reason is that one of the methods is "hostname --fqdn" and that will _set_ the hostname to "--fqdn" on some OS.
I suspect this will fix it:
--- client.sh~ 2008-06-29 05:28:08.000000000 +0200 +++ client.sh 2009-07-11 01:01:27.000000000 +0200 @@ -209,7 +209,7 @@ echo "WARNING: You are running client.sh as root. Don't do that!" >&2 else if hostname --fqdn >/dev/null 2>&1; then tmp_node=`hostname --fqdn` - if [ X$tmp_node != Xlocalhost.localdomain ]; then + if [ X$tmp_node != Xlocalhost.localdomain -o X$tmp_node != Xlocalhost ]; then cur_node=$tmp_node fi else
On Fri, Jul 10, 2009 at 11:00:02PM +0000, Peter Bortas @ Pike developers forum wrote:
if [ X$tmp_node != Xlocalhost.localdomain ]; then
if [ X$tmp_node != Xlocalhost.localdomain -o X$tmp_node != Xlocalhost ]; then
should be: if [ X$tmp_node != Xlocalhost.localdomain -a X$tmp_node != Xlocalhost ]; then otherwise it's always true...
that works.
greetings, martin.
both machines are fully green now, thank you for discovering the issues...
greetings, martin.
pike-devel@lists.lysator.liu.se
-
Marcus Comstedt (ACROSS) (Hail Ilpalazzo!) ＠ Pike (-) developers forum -
Martin Baehr -
Martin Bähr
-
Peter Bortas ＠ Pike developers forum