>I made it so users can decide to abort the suspension from their own
>hooks/rtcwake script: if that script returns a non-zero exit code (propagates
>through the checkrtcwake() function), suspension will be aborted. Only
>in that case it goes directly to StateDead.

In my testing it also goes directly to StateDead when I push the power 
button once in red led.

>>This is a first attempt to implement regular rtcwake ability in
>>screenlock, so it can wake from deep sleep (crust) every X seconds to do
>>some business like check notifications, and then sleep again.
>
>Testing this patch it works fairly well, but it seems that after some 
>time it stops going back to suspend and just turns then screen on 
>instead...

So, it seems that in checkrtcwake if the time we woke up is "too late" (which it
sometimes is, I've seen a timediff of 4 at least in my testing) then the return
of checkrtcwake is unspecified, so probably at least sometimes 0 (seems like
a bug to have a case with no return value specified?) -- in syncstate when
checkrtcwake returns 0 it goes straight to StateDead and thus wakes up my phone
and turns on the screen.

>This is a first attempt to implement regular rtcwake ability in
>screenlock, so it can wake from deep sleep (crust) every X seconds to do
>some business like check notifications, and then sleep again.

Testing this patch it works fairly well, but it seems that after some time 
it stops going back to suspend and just turns then screen on instead...

Any tricks others are using for this?  I can put it into "caret mode" and then
drop the cursor by tapping, but finger dragging is still scrolling at that
point.  It seems after dropping a cursor I can select by tapping Shift, Right,
Shift, Right over and over -- maybe svkbd needs a way to lock shift to on?

>It sounds as if you missing the second switch of "Line Out" in
>alsamixer.

Thanks!  I managed to fix it by changing "Line Out Source Playback Route" from
Stereo (the default on the image) to Mono Differential using alsamixer.  I have
no idea what this means, but it takes the speaker from barely usable to sounding
basically great and quite loud (I had to turn it down!)

Somebody claiming to be Stephen Paul Weber wrote:
>>polkit is using 0.8% of ram on my sxmo install. How did you install
>>sxmo? Are you running any scripts that may be affecting polkit?
>
>hmm... I installed using the image. The only thing always 
>running is the dino im client, and usually mutt.

Ok, installing htop shows the same results as you. After deeper 
investigation it seems a huge amount of the ram is going to mutt, which 
makes way more sense given the size if my mailbox.

>polkit is using 0.8% of ram on my sxmo install. How did you install
>sxmo? Are you running any scripts that may be affecting polkit?

hmm... I installed using the image. The only thing always running is the 
dino im client, and usually mutt.

I haven't really had RAM or performance problems at all yet, so this is more of
a curiosity: for others running sxmo, if you run top and sort by RAM usage (m)
do you see polkit taking 50%+ of the RAM?  free confirms over half the RAM is in
use, showing 942 MB of RAM available for use (out of 1928 MB total).

apk search -r shows that polkit is only present as a depedency of modemmanager.
Obviously I want the modem to work, but it seems incredibly heavy for just
that... but maybe I'm reading the stats wrong?  Or maybe something is strang on
my device in particular?  Just wondering what others see.

This way we catch all load events and keep the property in sync.

Closes https://todo.sr.ht/~mil/sxmo-tickets/71
---
 surf.c | 3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)

diff --git a/surf.c b/surf.c
index 000148d..63646a0 100644
--- a/surf.c
+++ b/surf.c
@@ -1312,17 +1312,16 @@ void
 loadchanged(WebKitWebView *v, WebKitLoadEvent e, Client *c)
 {
[message trimmed]

Replace the fallback which prepends "http://" to check via regex if this is a
plausible URL, and if not, search using non-JS DuckDuckGo instead.

Closes https://todo.sr.ht/~mil/sxmo-tickets/3
---
 surf.c | 15 ++++++++++++++-
 1 file changed, 14 insertions(+), 1 deletion(-)

diff --git a/surf.c b/surf.c
index a58aae1..000148d 100644
--- a/surf.c
+++ b/surf.c
@@ -538,10 +538,19 @@ newclient(Client *rc)
 void
[message trimmed]