- 01 Dec, 2015 1 commit
-
-
Franco Fichtner authored
-
- 16 Sep, 2015 1 commit
-
-
Franco Fichtner authored
-
- 09 Aug, 2015 3 commits
-
-
Franco Fichtner authored
-
Franco Fichtner authored
o Only probe for key press on bootup. o Rebuild autodetect feature: fixup second use case of get_interface_list() to filter out devices without a good link state, bring back interface up for it to work and mute kernel messages that would clobber stty behaviour.
-
Franco Fichtner authored
-
- 10 Jul, 2015 1 commit
-
-
Ad Schellevis authored
-
- 02 Jul, 2015 2 commits
-
-
Franco Fichtner authored
-
Franco Fichtner authored
Unsetting LAN, later setting it will strip the defaults from the interface section so try to avoid that by always settings defaults.
-
- 01 Jul, 2015 8 commits
-
-
Franco Fichtner authored
-
Franco Fichtner authored
-
Franco Fichtner authored
-
Franco Fichtner authored
Always unload LAN if previously set. Otherwise no code changes.
-
Franco Fichtner authored
-
Franco Fichtner authored
-
Franco Fichtner authored
Boy, this is a fun one! :) Still some work to do, but seems really good for headless installation or just being plain sloppy. Needed do adapt timeout() to return the actual input char, so that we can do the `any key' prompt.
-
Franco Fichtner authored
-
- 10 Jun, 2015 1 commit
-
-
Franco Fichtner authored
-
- 24 Apr, 2015 1 commit
-
-
Franco Fichtner authored
-
- 02 Apr, 2015 1 commit
-
-
Franco Fichtner authored
-
- 24 Mar, 2015 1 commit
-
-
Franco Fichtner authored
Either the RAM requirements have increased, or hardware has increased. Simply stating the arbitrary and obvious while doing things of little value to `fix' the situation it is probably better to just force an old-fashioned out of memory, which is mostly the point where the user realises that it is not going to work anyway.
-
- 14 Mar, 2015 1 commit
-
-
Franco Fichtner authored
We can do without those on the install media. Instead, route necessary code parts through is_install_media(). A lot of stale and forgotten code got ditched in the process--e.g. when did we ever run without an /etc/fstab?
-
- 05 Mar, 2015 1 commit
-
-
Franco Fichtner authored
/tmp is unlikely to change. If, however you decide to change it to somewhere else the likelihood of someone else hardcoding /tmp just to annoy you is so high that it simply makes no sense to keep any softcoding in this regard. More batches coming, but since the usage is so widespread, I'll push it bit by bit.
-
- 24 Feb, 2015 1 commit
-
-
Ad Schellevis authored
- remove the use for pfSctl, replaced with a new configd control script to be able to remove the check_reload_status port - replace use of global $g['booting'] because some scripts use this outside the main loop which causes strange behavior ( for example dhcp client which starts a newwanip on connect ). replaced which check on file /var/run/booting which is set in rc script
-
- 29 Dec, 2014 1 commit
-
-
Franco Fichtner authored
Also reapply a few after recent merge/rebase action in the CP.
-
- 10 Dec, 2014 1 commit
-
-
Franco Fichtner authored
Make "src" the implicit "/usr/local" so that nobody can sneak in weird files in the base file system or potentially overwrite them. This is the last time, I swear... :)
-
- 09 Dec, 2014 1 commit
-
-
Franco Fichtner authored
-
- 15 Nov, 2014 1 commit
-
-
Franco Fichtner authored
A one-port setup is essentially WAN-only. The good thing is the GUI attaches to the WAN port and is automatically reachable. This helps with general testing of all things around builds. While there, append a whitespace that has been missing for years.
-
- 27 Oct, 2014 1 commit
-
-
Ad Schellevis authored
-