[X] My Assistant

I wonder how many and what "services" (can't think of another word on short notice) you can get directly from google search, besides "weather" or "random"

I wonder how many and what "services" (can't think of another word on short notice) you can get directly from google search, besides "weather" or "random"

I'd probably call them "widgets," or "applets," or maybe "tools." I think Google might actually term them "toys."

There's also a unit converter, a "do a barrel roll" thing, "the answer to the ultimate question of life the universe and everything," and pac-man.

Additionally, there are a lot listed here:
[en.wikipedia.org] https://en.wikipedia.org/wiki/List_of_Google_Easter_eggs

Lol, since I last watched it they changed the title on "Too Many Cooks."

This post has been edited by dragontamer8740: Oct 8 2020, 18:57

Yes, but the BIOS wasn't handling them on its own (to the extent of actually changing brightness levels); just generating ACPI events that nothing was doing anything with.

just happened to talk with someone else about something similar like a couple days ago, and I remember something about you building your own kernels as well. what made it work there was setting the acpi_backlight=intel kernel parameter, while it worked ootb on the arch kernel I was using. may or may relate to your case(?)


on another unrelated topic,
someone finally found out what caused the bug I was having in firefox and its only a one line patch(yay), but it wasnt a dev just someone else with the same issue(less yay). so now im back at the awful hellscape that is having to sit through 80 minute+ firefox compile times each time I want to follow a version bump.

[pbs.twimg.com] https://pbs.twimg.com/media/Ej0d4eiXYAEMTn3...&name=large


The 5600X is apparently 300$, derp.

just happened to talk with someone else about something similar like a couple days ago, and I remember something about you building your own kernels as well. what made it work there was setting the acpi_backlight=intel kernel parameter, while it worked ootb on the arch kernel I was using. may or may relate to your case(?)

I know about all of that stuff, it didn't help on this particular machine. BTW this system is using a stock debian kernel still.

On my Thinkpad I have been using acpi_backlight=video "acpi_osi=!Windows 2012" instead and that worked for whatever reason, and has since I got the machine years ago. This other laptop is not a thinkpad though (some random single-core Celeron Toshiba from around 2010), and seems to like misbehaving. "intel" did not do what I wanted on the thinkpad, but was somewhat close.

In any case my scripts work fine and have a lot less overhead than a power managing daemon.

someone finally found out what caused the bug I was having in firefox and its only a one line patch(yay), but it wasnt a dev just someone else with the same issue(less yay). so now im back at the awful hellscape that is having to sit through 80 minute+ firefox compile times each time I want to follow a version bump.

What is this bug again?

[pbs.twimg.com] https://pbs.twimg.com/media/Ej0d4eiXYAEMTn3...&name=large
The 5600X is apparently 300$, derp.

Link doesn't seem to be working right now. And what "tier" is that one? I don't follow CPU's closely anymore.

This post has been edited by dragontamer8740: Oct 9 2020, 08:15

Yeesh, connecting 72VDC to something with large input filter caps makes a non-trivial spark. My first thought was "I hooked it up wrong and nuked my new toy" before the screen lit up.
(IMG:[i.postimg.cc] https://i.postimg.cc/JrZjT2Dq/20201008-172632.jpg)
The lack of a power switch on the radio and the sparks it makes when hotplugging tell me I should put more effort into making the power supply clean looking and easy to interact with. I may as well have the round connector for the radio be on the back panel, though I'm sure MIL-5105 connectors with silver contacts resist arcing better than banana plugs.
I suppose one of the adaptations inside this model is a brownout detector to power it down cleanly- the radio it's based on was typically configured for ignition sense, or sometimes a power button (it always had battery power applied).

I know about all of that stuff, it didn't help on this particular machine. BTW this system is using a stock debian kernel still.

On my Thinkpad I have been using acpi_backlight=video "acpi_osi=!Windows 2012" instead and that worked for whatever reason, and has since I got the machine years ago. This other laptop is not a thinkpad though (some random single-core Celeron Toshiba from around 2010), and seems to like misbehaving. "intel" did not do what I wanted on the thinkpad, but was somewhat close.

In any case my scripts work fine and have a lot less overhead than a power managing daemon.

in the end, whatever works out. as all my stuff is 'enriched' with systemd that handles most of it so I'm not entirely sure what it'd look like without it.

What is this bug again?
Link doesn't seem to be working right now. And what "tier" is that one? I don't follow CPU's closely anymore.

its the one where using multiple keyboards causes lag [bugzilla.mozilla.org] https://bugzilla.mozilla.org/show_bug.cgi?id=1640535. it turned out to be blamable on the bidirecitonal keyboard check. but the real issue remains that this reset call is made each time you press a key on another keyboard. so while I'm glad someone found a fix, I dont think this will make it any more likely to be fixed in the future upstream at this state sadly enough.

Yeesh, connecting 72VDC to something with large input filter caps makes a non-trivial spark. My first thought was "I hooked it up wrong and nuked my new toy" before the screen lit up.
The lack of a power switch on the radio and the sparks it makes when hotplugging tell me I should put more effort into making the power supply clean looking and easy to interact with. I may as well have the round connector for the radio be on the back panel, though I'm sure MIL-5105 connectors with silver contacts resist arcing better than banana plugs.

Try a europlug (type E, actually).
I love europlugs.

For the plug itself, use a CEE 7/6. For the receptacle, use CEE 7/5 (type E, french). I personally am fond of CEE 7/5 (french) because when working with AC you always know which line is neutral (and that both are not actually live) and they prevent plugging in upside down.
In the case of direct current, of course, this means you can't mix up the terminals. Type F (the german plug) allows plugging in either way, which is why it wouldn't be as good for your case.

Also, consider adding bleeder resistors to the filtering caps, if there aren't any.

in the end, whatever works out. as all my stuff is 'enriched' with systemd that handles most of it so I'm not entirely sure what it'd look like without it.

This happened when it was running systemd as well; I just didn't bother to fix it until after the sysvinit switch (priorities).
I should specify, [sourceforge.net] acpid is running, but I meant something like gnome-power-manager.

I just cleaned the script up a bunch.

CODE

#!/bin/sh
# Perform brightness adjustments whenever ACPI brightness events occur.
BACKLIGHT_DEV='/sys/class/backlight/acpi_video0'
MAXBRIGHT="`cat "$BACKLIGHT_DEV"'/max_brightness'`"
CURBRIGHT="`cat "$BACKLIGHT_DEV"'/brightness'`"
setbright()
{
  if [ "$1" -ge 0 ] && [ "$1" -le "$MAXBRIGHT" ]; then
    echo "$1" > "$BACKLIGHT_DEV"'/brightness'
  fi
}
adjbright()
{
  setbright "$(expr "$CURBRIGHT" '+' "$1")"
  true
}
in="$1";
inStart="${in%% *}"
inDirection="${inStart#video/brightness}"
case "$inDirection" in
  "up")
    adjbright '1'
   ;;
  "down")
    adjbright '-1'
   ;;
esac

its the one where using multiple keyboards causes lag [bugzilla.mozilla.org] https://bugzilla.mozilla.org/show_bug.cgi?id=1640535. it turned out to be blamable on the bidirecitonal keyboard check. but the real issue remains that this reset call is made each time you press a key on another keyboard. so while I'm glad someone found a fix, I dont think this will make it any more likely to be fixed in the future upstream at this state sadly enough.

Yeah I kind of doubt it will be, too. I pretty much have no faith in Mozilla to do the right thing anymore.

This post has been edited by dragontamer8740: Oct 9 2020, 23:33

The PS5 is compatible with 99.9% of all PS4 games. Wow. So long PS4.

The PS5 is compatible with 99.9% of all PS4 games. Wow. So long PS4.

Good; that might make them actually affordable in a couple years.
Not sure I'll buy one though, unless there's a bootloader exploit and a linux/BSD with Xorg for it.

work

I pretty much have no faith in Mozilla to do the right thing anymore.

I get worried that at some point, FireFox market share will shrink enough that the greedy executive managers will vote to loot the company by selling the company to whomever will buy its assets. That way, the company board of directors will have extra funds to vote each executive manager their [en.wikipedia.org] Golden Parachute in addition to funding their escrow cushy pension plans as everyone is laid off.

I just modified MComix3 to be able to use the Lanczos resampling algorithm from PIL/Pillow (and, as a result, also the other resamplers that PIL supports). Anyone want it?
 PIL_resampling.txt ( 4.3k ) Number of downloads: 4
(Save as a .patch file; patch with `patch -p1 < PIL_resampling.patch`)

I might upstream it later, but right now I have enough objections to some stupid behavioral changes the MComix3 fork owner has made to the program (without opt-outs) that I might just maintain my own derivative fork instead.

It really does make images look sharper when downscaled to fit.

Sample comparing the old "hyperbolic" option with "lanczos" (animated PNG; should work in newish browsers as well as every version of Firefox released since over a decade ago):

(non-magnified:)


Ignore the ugly fonts; that's my fault but not something caused by the patch.
I don't find it slow on my desktop but know from experience that Lanczos is still slower than a lot of the other algorithms in this list, even if not noticeably slow on a beefy machine. It's actually hardly noticeable on my less-beefy ones, either. Maybe on a 486 or Pentium you'd really feel it.
Also, I fixed the duplicate entry in that list in the patch above. It's actually Hamming.


This post has been edited by dragontamer8740: Oct 12 2020, 08:02

I get worried that at some point, FireFox market share will shrink enough that the greedy executive managers will vote to loot the company by selling the company to whomever will buy its assets. That way, the company board of directors will have extra funds to vote each executive manager their [en.wikipedia.org] Golden Parachute in addition to funding their escrow cushy pension plans as everyone is laid off.

can leave out the 'at some point' mozilla has been going in bad directions for a while now and honestly it seems like they're trying to move away from the browser as their main project to focus on other nonsense.
(IMG:[calpaterson.com] http://calpaterson.com/assets/mozilla-boss-pay.png)
[calpaterson.com] http://calpaterson.com/mozilla.html

For better or for worse, the chromium engine seems to have ursurped gecko as the dominant web browser family. I think 2 of the 3 browsers on this phone (I like options, and each one renders stuff differently ofc, none are Google Chrome) are chromium based. I've stuck with firefox based stuff for linux and windows.
I think android's proliferation may have been the nail in firefox's coffin.

For better or for worse, the chromium engine seems to have ursurped gecko as the dominant web browser family. I think 2 of the 3 browsers on this phone (I like options, and each one renders stuff differently ofc, none are Google Chrome) are chromium based. I've stuck with firefox based stuff for linux and windows.
I think android's proliferation may have been the nail in firefox's coffin.

I don't agree; Gecko runs rather well on Android, and FF android is probably the most configurable/customizable browser available for android right now. Definitely the most extensible (I have my user scripts on it, and an ad blocker; I can also change the user agent it reports in about:config).

Chromium is preinstalled and people are lazy, its why IE still exists.

Chromium is preinstalled and people are lazy, its why IE still exists.

I'm arguing that it's not what's been the final nail in Mozilla's coffin. It's nails have been constant mismanagement and alienation of users that would otherwise be actively vouching for people to switch. If I were to tell someone to try FIrefox, these days it'd be with many reservations and much hesitation.

I remember around 6-7 years ago (2013-2014), someone I knew peripherally mentioned he switched to Firefox (from Chrome) because he had realized that Chrome was holding him back; there was simply no reason to use the inferior (less extensible/powerful) product. At that point, there were far more powerful addons available at that point for Mozilla's products, since it's before they had even announced their intent to kill off all XUL-based software. He wasn't in a CS career (although he was definitely more technically capable than a lot of others I knew); he just realized that Firefox had more features and greater configurability. Back around 2010 (and still for the most part in 2014) I was able to recommend people use Mozilla products with little reluctance, since I knew they were getting an upgrade for doing so.

The rapid release cycles also were a shot in the foot; they annoyed corporate users and end-users, and unlike Chromium rapid releases they regularly introduced changes that broke users' existing workflows for little benefit. I am still mad about Australis (29), even though 57 just made everything so much worse.

Instead of aiming at leveraging what makes their codebase unique, their management seems to be under the impression that "security" and "privacy" are the only important thing for a browser when someone is looking at options. To that end, they have just decided to dumb-down the browser and reduce maintenance costs by destroying its versatility. All while their executives pocket the savings incurred.

This post has been edited by dragontamer8740: Oct 12 2020, 07:02

It doesn't use normal error message screens anymore, it shows a little picture like I'm an idiot with zero knowledge of the net and browsers.

It doesn't use normal error message screens anymore, it shows a little picture like I'm an idiot with zero knowledge of the net and browsers.

which one is this?
Firefox?

I haven't updated it in at least a year. Sounds like I should continue not doing that.

Today I traced a ground loop on a 24-channel theater mixer board.

This post has been edited by dragontamer8740: Oct 12 2020, 21:23