Bay 12 Games Forum

Please login or register.

Login with username, password and session length
Advanced search  

Show Posts

This section allows you to view all posts made by this member. Note that you can only see posts made in areas you currently have access to.

Topics - methylatedspirit

Pages: [1]
1
General Discussion / A think about the Great Disorder
« on: March 26, 2021, 11:44:54 am »
You know the Great Disorder, which causes post links in threads where greatorder posted (and had deleted a while ago) to be misaligned forwards by some thread-dependent amount? I just realized that pages, despite being indexed by post number, are not affected by this glitch. I lack the knowledge to understand why SMF behaves like this, but I think it's neat to point out.

This is a link pointing to page 10369 (at 15 posts per page) of the WTF thread.
http://www.bay12forums.com/smf/index.php?topic=68850.155520

This is a link that is supposed to point to the 155520th reply (10369 × 15 - 15, assuming 15 PPP) of the WTF thread. Instead, it launches you straight to the next page (as of this writing). Sorry MrRoboto, I'll be using your post as a test subject throughout. I'll replace it if you don't want me to use it.
http://www.bay12forums.com/smf/index.php?topic=68850.msg8262806#msg8262806

Observe the ends of the URLs. The one that you get for a page is indexed relative to the current total number of posts. The one for a post is indexed relative to something else. I have no clue what it means, but it sure as hell is larger. Global post counter? If so, there are a lot of posts I don't see, since the forum itself shows 4.8 million posts, and that number is 8.3 million.

If I run with the hypothesis that the 8.3 million figure is a sequential index of every post in chronological order, regardless of deletions or hidings, it's even more odd that the Great Disorder affects only quote links.



There's a possible workaround here by using the page link formatting instead of the quote format. Notice how a typical quote is formatted.
Code: [Select]
[quote author=MrRoboto75 link=topic=68850.msg8262806#msg8262806 date=1616638054]
[/quote]

It turns out that the link= flag is very easy to work with. The link it generates is (assuming there isn't a different website specified?) "bay12forums.com/smf/index.php?", and then after that, it's concatenated with whatever is after the link= flag.  I'm sure you can play some very funny tricks with that.

My workaround involves replacing the msg-followed-by-numbers thing with the post number, just like how page links are formatted. Just replace it with the trailing number of the URL of the page you found the post on. In my example, this would be
Code: [Select]
[quote author=MrRoboto75 link=topic=68850.155520 date=1616638054]
[/quote]

And if I release it from its [code] cage, it produces:

Which would get you at least within the vicinity of the intended post, rather than being flung forwards a hundred pages like you would normally.

I don't think it's a very practical workaround, but I suppose if there's a particular need for vaguely-accurate quote linking within the subset of affected threads...

2
I read this in a different thread:
Hardware support in Mint is amazing compared to Hurd.

The madlad in me read that as: "Install Hurd on real hardware, entirely for the hell of it". Since it's getting to be a bit too long for the original thread anyway, I've decided to start a new thread.

The plan is to install a Hurd-based distro on real hardware. I've decided on Debian Hurd, mostly because it seems to at least support a GUI out of the box. Since I enjoy not destroying my precious Windows 10 install, I'll go the unconventional route of installing to another USB stick. I only happen to have 16 GB sticks around, so I'll be a bit constrained.

I don't have a capture card, so you'll be seeing pictures taken with my phone. Some pictures were taken out of order, since I completely forgot about documenting the experience at first. Certain bits of experimentation with the BIOS settings removed, because I think it would've been too dull.

Spoiler: System Specs (click to show/hide)

Using the 20210219-23:25 nightly build
(It's new hardware, I figured that I'd get better results on nightly builds.)

Let's see, Debian Hurd only works on i386 (x86, non-64-bit) systems. How about I try with my current settings anyway?



No. Disabling Secure Boot does the same. Legacy boot?



Yes! Enter. I'm at GRUB. Every time I get to GRUB in this build, the computer beeps twice at me.



I'll select graphical install, because I'm a total noob. And...



Blank screen. I'll be seeing this thing a lot of times. To save on bandwidth (and because I don't know which blank screen belongs to which), I'll replace all future instances of a blank screen with a 1536x864 #000000 black image. What if I enable the Intel iGPU, since the only good Nvidia drivers are proprietary?



Again? What about pseudo-graphical mode?



No. Text mode?



Okay, there's something. I'll just power through and select the defaults, hitting Enter at every prompt until I encounter a problem. Here's the first, and pretty much fatal problem:



How does an installer not see the damn thing it booted off off? Uh, sure, load drivers from removable media. (I had another Debian Hurd USB plugged in while taking the next photo, hoping that it would work there, but no, whatever drivers it wants, it sure as hell can't see)



This is pretty much a fatal error. I suppose it doesn't have the drivers to even read off its own USB, so it just stops there. If I answered no to loading drivers, I get this:



In either case, if you continue, you'll see:


(Yes, the timeline's getting a bit fudged. Blame me.)

I suppose the only thing I could do at this point is to abort installation.



It can't even shut down right, it seems, or I'm just really impatient. It was just stuck here, and eventually, I just put it out of its misery and shut it down manually.



That was a bust.

Using the stable 20190705 image

I booted back into Windows (and reverted my BIOS settings) I used Rufus to burn another Debian Hurd USB, but this time using the stable 20190705 build. Maybe this one's actually gonna work. I went back to BIOS and changed the settings back to legacy boot and iGPU enabled.



At least it's seeing it. "Enter" as usual.



Holy crap, that is some tiny text! I'll just go into "Debian GNU/Hurd", just because.



I'm pretty sure "start ext2fs: ext2fs: device hd0s2: No such device or address" is low-level-speak for "I can't find the USB drive that I need to begin installing Debian!". It just hangs here, since it's exactly as if I pulled out the USB stick out in GRUB. It was plugged in the whole time, mind you.

I dunno, try again and go to advanced options?



Recovery mode? Maybe it loads fallback drivers or something.



I'm recycling images, but believe me, it's the same exact text in both instances. It just crashes when trying to find the USB install media.

Conclusion

Ziusudra, you're right. GNU Hurd's hardware support sucks. It doesn't even support the USB driver that's needed to read the install data in the first place. If anyone has any ideas of other Hurd-based distros that would work, let me know.

3
Creative Projects / Extended Databending
« on: February 07, 2021, 09:52:17 am »
These are some writeups I made about databending, or rather, new techniques that I haven't seen anyone publish yet. It's all posted on GitHub, which probably isn't the best way to do it, but I just wanted to get the info out there. This is pretty much all I know for now.

(Table of Contents)

Databending with Audacity: What I do Differently/Required Reading
|
-------->
Addendum: What happens if you use different audio encodings while importing?

Databending with Audacity: The Effect of Paulstretch on Planar RGB Images

Databending with Audacity: FFmpeg as an Intermediary for Images

Databending with FFmpeg: (Ab)using Lossy Audio Compression to Databend Images
|
-------->
Addendum: Automated image to audio 'bending

Databending with FFmpeg: Misinterpreting Audio as Video for Fun and Profit

Of course, all I'm doing here is based on other people's work, and none of this would be possible without them. The ones I could remember were:

Databending using Audacity by Antonio Roberts
Databending using Audacity Effects by questionsomething (this one inspired the Paulstretch piece, and it's the same technique down to using planar RGB, actually)

4
Life Advice / I want to throw some info out there into the world
« on: February 05, 2021, 05:16:15 am »
(I've asked this before on a different thread, but I think it's too "big" of a question for said thread)

Yeah. See, I did some databending, and I noticed that what I'm doing hasn't yet been detailed online, at least from what I can tell. My techniques are new, as far as I know. Now, I'm not exactly a scientist, but I do like the idea of documenting my methods so that others may make use of them. I'm not an artist. I don't see myself as having any artistic ability, and all work that resembles it is just a result of a repeatable procedure.

Philosophy aside, where can I write about these new techniques, such that they will be indexed by search engines and turn up in search results? I'm not even saying they're good, they're just new. I don't want publicity, I just want to put them somewhere online for people to see and hopefully make their own work with those techniques. I want to be as anonymous as possible, but I don't think I could pay to host a web server to host a blog on if it's only a few things I want to share.

Also, I suck at actually finding good source material. I don't need direct links to pictures, per se, just a list of keywords that yield interesting pictures. You can give me specific images if you want, but I won't guarantee that I'll actually use them.

5
A recent Anandtech article revisited the Intel Broadwell-DT processors using their current testing methodology. These tests include the DF 44.12 worldgen tests, testing worldgen at 65x65 250 years, 129x129 550 years, and 257x257 550 years. Further details can be found in the linked article. The interesting thing is that the Broadwells tested have 128 MB of eDRAM acting as L4 cache. Combined with the DF tests, I'd like to bring up a point:




(Source)

The CPUs with L4 cache are the i7-5775C and the i5-5675C in light blue and blue respectively. It would seem that with the cache, even the i5-5675C is able to outperform an i7-6700K in all three tests, which is a year newer than the i5. It manages to offset the RAM speed disadvantage, considering that the 5675C ran DDR3-1600, and the 6700K ran DDR4-2133.

Consider how close these CPUs are to the top. The i7 is impressive, managing to trail slightly behind the i5-10600K in the first and third test, and actually beating it in the second test. Keep in mind, this i7-5775C is 5 years older than the i5-10600K, and it runs on an older architecture than the i5-10600K (the former runs on Haswell, the latter on Skylake). Not only that, the i7-5775C can boost up to 3.7 GHz, and the i5-10600K boosts up to 4.8 GHz, so even with the clock speed disadvantage, it manages to do very well.

That's all I can really say about this. It's an interesting point, if anything.

6
DF General Discussion / What is the oldest CPU that can run DF?
« on: September 11, 2020, 09:02:42 am »
(I asked this question somewhere else on this forum, but there was no response there, so I'm asking somewhere more relevant. A cursory search, both on Google and on this forum yields old (Pentium 4-era, oldest being Pentium III-era as in the link below), but even that one doesn't confirm that it's the oldest CPU capable of running DF. No-one seems to have asked what I'm asking.)

Exactly what the subject says. I've seen a 700 MHz Coppermine (Pentium III-era) Celeron with 180 MB of RAM running DF, so now I'm wondering, how low and old can you go, CPU-wise? Pentium II? Pentium MMX? OG Pentium? 486? What instruction sets are needed to run DF? This isn't a question of how well it will run, this is more about what will run DF, as badly as it would.

Since I want this question to have some meaning, I have to specify some rules: 47.04, any variant (32/64-bit, legacy/SDL), any OS, any amount of RAM (try to max out the RAM to rule out out-of-memory crashes), and screenshots of DF running on the system (title screen, worldgen, embark) are highly recommended. "Running" here means that it is able to do worldgen (any parameters, including advanced), and it's able to do an embark (any size, any bought items). For bonus points, record how long worldgen takes and the FPS after embarking.

Summary (based on using virtual machines; if possible, confirm with real hardware and post it here):

32-bit Windows DF requires SSE2, thus making the Willamette Pentium 4s (2000) the oldest CPUs capable of running Windows DF.
32-bit Linux DF requires SSE, making the Katmai Pentium IIIs (1999) the oldest CPUs capable of running Linux DF.
The oldest x86 Macs have Yonah Core Solos/Duos, which has SSE3. Assuming that Mac DF doesn't need SSSE3 or later (I don't see why this would be the case), this would make the Yonah Core Solos/Duos the oldest CPUs capable of running Mac DF. I would still like conformation with VMs or real hardware, just to make sure.

7
So, I have a fort with 7 caged troglodytes, some of them male, some of them female.
Is there a good use for these troglodytes that isn't "build cages, connect to lever, pull the lever, use military to kill them all"?

Pages: [1]