Chris Koeritz [Thu, 7 Jan 2021 00:02:33 +0000 (19:02 -0500)]
recording grub mod for system 76 laptops
addresses a freeze-up condition when returning from sleep. doesn't totally eliminate it but makes the sleep process a lot more reliable. usually, if sleep is going to work, it will keep working for months now. before any old sleep might come up with a dead session where machine was running but had no activity at all and had to be totally powered off to return to normal state.
Chris Koeritz [Thu, 25 Jun 2020 14:42:11 +0000 (10:42 -0400)]
updated to work on centos
centos was bitching about:
git rev-parse @
and
git pull --tags --all
on centos 7.
revised to use:
git rev-parse HEAD
and
git fetch --tags --all
instead.
Chris Koeritz [Wed, 17 Jun 2020 15:26:45 +0000 (11:26 -0400)]
updated to use java itself to get JAVA_HOME
this is the best approach possible if the desired version of java is already in your path.
our previous implementation is a guessing game and is dysfunctional. "is", since it's still in place if the call to java fails.
Chris Koeritz [Wed, 10 Jun 2020 19:33:50 +0000 (15:33 -0400)]
cleaner implementation of keep_awake
will still put process in background, but now you can control it from the shell.
unfortunately, these will probably still zombie out if the connection gets dropped for some reason (either because we're not printing frequently enough or for causes out of our control).
Chris Koeritz [Mon, 1 Jun 2020 21:02:20 +0000 (17:02 -0400)]
settled on whichable func over direct call to which
the whichable function does the which but also hides errors from not finding the command in question. this is much preferred over seeing errors about "gvim not found" etc.
Chris Koeritz [Mon, 1 Jun 2020 14:19:23 +0000 (10:19 -0400)]
simplified sudo function
tossed out a lot of my changes after discovering the -i option to sudo, which starts a legitimate fresh shell of the type i needed. now, the only weirdity i have to deal with is that the PATH variable still seems to be passed in from sudo even though i've asked for that fresh shell. setting it to null prior to doing the sudo fixes that. works a bit better now.
Chris Koeritz [Wed, 27 May 2020 16:19:06 +0000 (12:19 -0400)]
moved to trusted x.org forwarding
need trusted mode on most of my connections; just don't connect to some random place that you don't trust over ssh, and you won't get a bunch of freaky windows popped on your screen.
Chris Koeritz [Wed, 27 May 2020 13:43:05 +0000 (09:43 -0400)]
fully functional death star
or, not so much. fixed sudo and resolved the weird mac issue for extract_current_wd or whatever function.
added the mac bash init into the bashrc templates used by the connect feisty script.
this is now working right again on all platforms, i think, although it is probably not yet clobbering every user specific environment variable. soon though.
Chris Koeritz [Wed, 27 May 2020 03:05:46 +0000 (23:05 -0400)]
vital fixes to sudo function
not sure the initial thrust is done yet, of having root not get a pile of wrong feisty meow variables when i sudo to it with -E flag, but this does at least get the shell options to not still expand for the one case we need. we will fix that to be more specific to enable mac to participate, since mac does actually need aliases expanded.
this also crucially fixes the IMPORTED_XAUTH variable which may never have been working right before? saw a multi-line answer on my host, so taking just the first line of output. not sure that's right yet, but we can't have multiple lines when we add the things in xauth the way we do it now.
projects / feisty_meow.git / log