Custom Query (1145 matches)
Results (145 - 147 of 1145)
Ticket | Resolution | Summary | Owner | Reporter |
---|---|---|---|---|
#344 | wontfix | neo doesn't build on Jaunty | jdreed | |
Description |
In its sleeping code, neo uses the obsolete CLK_TCK constant. /usr/include/time.h and time.h(7posix) claim that CLOCKS_PER_SEC should be used instead, but go on to warn that CLOCKS_PER_SEC should not be assumed to be a compile time constant. CLOCKS_PER_SEC is required to be one million (1000000l) on XSI-compliant systems, and indeed is on both RHEL4 and Ubuntu. However, on RHEL4, CLK_TCK is 100, and is undefined on Ubuntu. Defining CLK_TCK as CLOCKS_PER_SEC makes neo sleep for significantly longer (1000 times longer, in fact) than it should between SNMP queries when getting device stats. Defining CLK_TCK as 100 allows neo to function correctly, but this is almost certainly the wrong answer. I've reported this "upstream" to bug-ktools, but can't even reference a transaction number in the discuss meeting because it's private. Yay. |
|||
#353 | wontfix | tty mode login for clusters (a la Ctrl-P) | geofft | |
Description |
It would eventually be nice to provide a TTY login mode, both because it's probably faster than an X login, and in case your X session or the system's X config is broken. This should be easy by adding a GDM custom command to essentially openvt schroot -c login getty. |
|||
#358 | wontfix | debathena-dns-config breaks networking when registration required | geofft | |
Description |
As far as I can tell, debathena-dns-config sets up bind to start resolving from the global root servers, not whatever DNS servers are provided via DHCP. This makes you sad when you're on a network that depends on redirecting web requests via DNS to a registration page. (For instance, the Middle East's network, although free, requires you to click okay on a web page on every session.) I haven't tested this with a MITnet-unregistered machine, but it seems like it would be a problem there, too, e.g., if you just installed debathena-workstation on a new machine from PXE. There appears to be code in /etc/resolvconf/update.d/bind to configure bind's forwarders to be what the other resolvconf sources say your DNS servers should be, which would be a perfect fix. Is that not getting triggered for some reason? |