2023-06-16 06:55:29 <veltas      > KipIngram: Have you kept up with this Reddit drama?
2023-06-16 07:40:09 <drakonis    > ah yes, reddit vs its userbase
2023-06-16 07:40:38 <drakonis    > they're strikebreaking by forcing the private and restricted subreddits into being public and replacing the moderators with scabs
2023-06-16 07:40:50 <drakonis    > its...
2023-06-16 07:40:57 <drakonis    > fucked up, really.
2023-06-16 07:42:06 <drakonis    > nothing good will ever come out of it
2023-06-16 07:44:58 <KipIngram   > veltas
2023-06-16 07:45:07 <KipIngram   > : No, not really.
2023-06-16 07:45:10 <veltas      > KipIngram
2023-06-16 07:45:16 <veltas      > Fair enough
2023-06-16 07:45:44 <KipIngram   > I assume it's similar to what went on with irc a while back.
2023-06-16 07:47:25 <veltas      > Similar in the sense of owner vs community
2023-06-16 07:47:31 <veltas      > I don't think any of the details are the same
2023-06-16 07:47:39 <veltas      > Well owner vs moderators, really
2023-06-16 07:48:13 <KipIngram   > I haven't "felt" any change in my own work moderating.
2023-06-16 07:48:27 <veltas      > I don't think it will have the same result as with Freenode
2023-06-16 07:50:44 <KipIngram   > I just don't seem to be bothered as much as a lot of other people by the fact that lots of things are owned.
2023-06-16 07:51:04 <KipIngram   > Guess I'm just a product of an earlier era
2023-06-16 07:51:30 <veltas      > It's different in every way. Reddit has always been commercial, Freenode was only meant to be owned by a sort of 'custodian'. Freenode's infrastructure was actually run by the mods, Reddit's infrastructure is run by paid employees
2023-06-16 07:53:16 <veltas      > I will say that it's also a known thing that Reddit mods aren't representative of the larger community, mods are often at odds with their subreddit's subscribers
2023-06-16 07:57:34 <KipIngram   > So far I seem to have managed not to land there, though I don't think you can do that job without a person here and there getting frosted at you.
2023-06-16 07:57:52 <KipIngram   > Sometimes community members behave badly, and they don't like it when they're called out on it.
2023-06-16 07:58:14 <veltas      > There's a type of person that desires power and will work for free
2023-06-16 07:58:14 <KipIngram   > I've tried to be very  "interactive and transparent," though.
2023-06-16 07:58:24 <veltas      > And they are often the surviving mods of big subreddits
2023-06-16 07:58:48 <veltas      > And totally at odds with their community, using their 'power' to direct the community in their minority vision
2023-06-16 07:59:07 <KipIngram   > I guess so.  I just want it to be a smoothly operating community.  I've had a lot of fun there and I want others to as well.
2023-06-16 07:59:25 <KipIngram   > When the opening came up, I realized that I cared about the place enough to want to help.
2023-06-16 08:00:14 <KipIngram   > There are come community members, though, who seem to want it to be a very locked down, carefully controlled place.
2023-06-16 08:00:28 <KipIngram   > Like a "curated literary forum" or something.
2023-06-16 08:00:49 <KipIngram   > I figure it's a public community, and not meant to be for a specific subset of interests.
2023-06-16 08:01:24 <KipIngram   > I do know the kind of person you were referring to up there, though - I think you see that in some police officers.
2023-06-16 08:01:28 <KipIngram   > Power seekers.
2023-06-16 08:01:38 <KipIngram   > People who enjoy *weilding* authority.
2023-06-16 08:05:27 <veltas      > The best mods are always the ones you can't tell are there
2023-06-16 08:05:47 <veltas      > But they are there and are removing viagra ads and incredibly low quality abrasive posts
2023-06-16 08:06:12 <veltas      > Or stuff that's wildly off-topic
2023-06-16 08:06:40 <KipIngram   > Right, or in general a rules violation.  We don't have many rules.
2023-06-16 08:07:10 <KipIngram   > Be nice, don't pirate, don't post spoilers, don't karma game.
2023-06-16 08:07:13 <KipIngram   > That's it.
2023-06-16 08:07:17 <veltas      > I dislike Reddit's guidelines for communities because it recommends being very up-front about rules. In my opinion you shouldn't do that, if you talk rules it almost incites rule-breaking, and then people who think they're justified as long as they are within the rules
2023-06-16 08:07:38 <KipIngram   > My most common action is getting accidental spoiler violations fixed.
2023-06-16 08:07:57 <KipIngram   > And I usually work with the poster to FIX it, rather than just toss it and move on.
2023-06-16 08:08:09 <KipIngram   > It's very rare for someone to deliberately spoil.
2023-06-16 08:08:28 <KipIngram   > Usually they just don't know how they're supposed to flair the post.
2023-06-16 08:09:19 <veltas      > Whenever I've had to moderate the more 'invisble' the moderation is, the better. When people complained about content I tended to say "don't look at it if you don't like it" because otherwise the work increases ten-fold, people start to *expect* moderation on everything
2023-06-16 08:09:23 <KipIngram   > Sometimes there are trolls, deliberately picking fights with people, but even the "be nice" violations you can see how came about.
2023-06-16 08:09:27 <veltas      > And people start pushing those rules
2023-06-16 08:09:38 <KipIngram   > Someone says something minor someone doesn't like, the other person responds, and it just escalates.
2023-06-16 08:09:51 <KipIngram   > No one can let the other one have the last word, and after a while it turns ugly.
2023-06-16 08:09:53 <KipIngram   > Human nature.
2023-06-16 08:10:01 <veltas      > Yeah but... just ignore it
2023-06-16 08:10:22 <veltas      > And in real life you will have much more offensive situations to deal with, it helps to be exposed so you can learn tolerance
2023-06-16 08:10:23 <KipIngram   > Well, right - I usually respond to both people and just ask them to drop it.
2023-06-16 08:10:46 <veltas      > I would allow the argument, after all it's Reddit. The post would get buried in the thread
2023-06-16 08:10:53 <veltas      > And you can ignore the whole thing trivially
2023-06-16 08:10:59 <KipIngram   > It really is funny how the anonymity of that kind of thing will get behavior out of people they probably wouldn't do in a face to face situation.
2023-06-16 08:11:18 <veltas      > I think the internet provokes this because people reflect their environment
2023-06-16 08:11:22 <KipIngram   > Well, arguments are fine.
2023-06-16 08:11:38 <KipIngram   > It's when they start talking about one another instead of about the book series that I usually step in.
2023-06-16 08:11:56 <KipIngram   > An argument about the books?  No problem.
2023-06-16 08:12:05 <KipIngram   > Like you said - it's Reddit.
2023-06-16 08:12:18 <veltas      > I have a thick skin but I'm also a bit prickly and it's probably from being exposed to the internet. I can take a lot, but maybe I'm a bit colder than most people. I'm conditioned to this
2023-06-16 08:12:24 <veltas      > But this has been useful in real life too
2023-06-16 08:12:34 <veltas      > I wonder about the kids today growing up in the hyper-censored internet
2023-06-16 08:12:44 <KipIngram   > But I've looked at Stack Exchange over the years, and I feel like there's more picking on people than helping, and I just don't want to see that happen to this community.
2023-06-16 08:12:56 <veltas      > Where there's a quick judgement and execution awaiting every "wrong-thinker"
2023-06-16 08:13:22 <veltas      > Stack Exchange was like that almost immediately
2023-06-16 08:13:25 <veltas      > It's the same as Wikipedia
2023-06-16 08:13:33 <KipIngram   > Yeah, having a skin of some thickness is important.
2023-06-16 08:13:55 <unjust      > i see the same pattern of 'picking on' instead of 'helping' in most computing related channels on irc as well
2023-06-16 08:14:06 <veltas      > The issue is you have a community run by unpaid volunteers, you end up attracting people who couldn't be paid to do anything
2023-06-16 08:14:16 <unjust      > ok maybe not most, but many
2023-06-16 08:14:29 <veltas      > ##C is like that, but there are some good people as well
2023-06-16 08:14:35 <veltas      > ANd you will usually learn something
2023-06-16 08:15:00 <unjust      > i stick to #c-jeopardy for anything C related
2023-06-16 08:15:09 <veltas      > I thought caze was an asshole, and he is a bit really. But he's also taught me stuff. Unfortunately lots of people would just never learn anything from him because they can't handle his approach.
2023-06-16 08:16:27 <veltas      > And then there are some people in there who enjoy being difficult with people asking questions for a sense of superiority, I won't name names!
2023-06-16 08:16:41 <veltas      > You can find out from the logs, they don't hide it well
2023-06-16 08:18:52 <unjust      > i like this channel because it is generally the total opposite of that, and there hasn't been anyone peddling the Not-Invented-Here nonsense. not that i've noticed, at least
2023-06-16 08:19:54 <KipIngram   > There are always some good folks.
2023-06-16 08:20:24 <veltas      > There's been arguments in here but most of the content in here is people asking questions and getting genuine help, and sharing their work/ideas
2023-06-16 08:20:27 <KipIngram   > That sense of superiority is the thing I notice a lot.
2023-06-16 08:20:41 <KipIngram   > Can't just help someone - gotta show them all the ways they're ignorant and misguided.
2023-06-16 08:22:17 <veltas      > It's especially frustrating in ##C because most of the channel joiners are just trying to finish homework, find the right syntax for something, or fix a bug; really early in their learning, and they get lectured on standardese that won't help them unless they join the C language committee 10+ years down the line
2023-06-16 08:23:58 <veltas      > Knowing the C standard more intricately, as I have learned solely from being exposed to that IRC channel, has not actually helped me at all. 95% of professional C programmers have no idea about standardese anyway and they are the ones I have to work with
2023-06-16 08:24:08 <veltas      > And it's not really a defect that they don't know
2023-06-16 08:25:52 <KipIngram   > That sounds pretty right to me.
2023-06-16 08:26:13 <KipIngram   > When I'm messing with C I just want to get something working, on the particular machine sitting in front of me.
2023-06-16 08:26:21 <unjust      > it's good to have some familiarity with the standards though, but i don't think it's helpful to spend time getting to the language lawyer level unless that holds some appeal to you.
2023-06-16 08:27:47 <KipIngram   > Most of the disagreement about content in my sub comes from a group of people who only want to have "serious plot topics" discussed.
2023-06-16 08:27:55 <unjust      > knowing more or less what is (and isn't) locale-specific, unspecified and undefined behaviour helps to write portable code, and helps to avoid anything too unexpected down the line when things change in terms of (1) compiler version, (2) compiler itself, (3) architecture
2023-06-16 08:27:58 <KipIngram   > They just don't like the "light fare" that some others bring in.
2023-06-16 08:28:08 <KipIngram   > I figure it's a public community - there's a place for both.
2023-06-16 08:28:41 <KipIngram   > And then there's one particular character that seems to have elicited a group of staunch supporters and a group of "haters."  And they go at it sometimes.
2023-06-16 08:29:23 <KipIngram   > I tell my wife that's a "nerds vs. jocks" thing.  It seems pretty obvious to me.
2023-06-16 08:29:45 <KipIngram   > The character is basically a nerdy character, but he's well and truly "escaped the nerd box" in the series, and there are people that bothers, apparently.
2023-06-16 08:29:57 <veltas      > unjust: I told myself that too, but really "what GCC does" is more useful than the standard for portability and practicality
2023-06-16 08:30:02 <veltas      > Or whatever compiler you use
2023-06-16 08:30:19 <KipIngram   > ^ I agree with that.
2023-06-16 08:30:37 <KipIngram   > Probably my biggest interest in gcc is a thing that is NOT standard.
2023-06-16 08:30:51 <KipIngram   > I like that pointer to label thing, because you can write a Forth with it.
2023-06-16 08:31:03 <veltas      > The standard doesn't allow signed numbers to overflow
2023-06-16 08:31:13 <unjust      > i also agree with that. and there are some gcc extensions that are definitely nice to have that the standard doesn't
2023-06-16 08:31:30 <veltas      > It's not even the extensions, it's the implementation choices GCC makes
2023-06-16 08:31:45 <veltas      > Many of those choices apply to all platforms GCC builds to
2023-06-16 08:32:05 <veltas      > And the C standard doesn't specify it, so we look to a sensible benchmark, which GCC is
2023-06-16 08:36:25 <unjust      > sure. obviously the standards can't be all encompassing, and there are always going to be things that are deliberately omitted to grant some freedoms to compiler (or interpreter) implementors. i'm also definitely not saying the standards are the be-all and end-all of C, but i am saying that having an awareness of the abstract machine that they define can be helpful to have a better understanding of the language itself
2023-06-16 08:45:07 <veltas      > I think I got a better understanding of C from learning about its history than from reading the spec. And actually knowing historical C, like first edition K&R, makes it easier to understand the standard
2023-06-16 08:45:12 <veltas      > Rather than the other way around
2023-06-16 08:46:20 <unjust      > it's great that there are many paths to the same end
2023-06-16 08:46:35 <veltas      > The places where this matter are not huge. Like I saw someone wondering why printing a signed byte as hex gives them "FFFF..." when it's only a byte
2023-06-16 08:46:57 <veltas      > And you can understand that via the standard or via the historical understanding, but either way this stuff doesn't come up much IMO
2023-06-16 08:47:46 <KipIngram   > First edition K&R is nice.
2023-06-16 08:48:12 <unjust      > i guess a lot of the issues rear their heads when you have code bases slapped together over a long period of time by people with varying skill levels, and no effort has been made at refactoring or addressing technical debt
2023-06-16 08:48:50 <KipIngram   > That bit about the signed byte - that's more about numbers and how to convey information to the reader than about what the bits actually are.
2023-06-16 08:48:59 <KipIngram   > Those extra 1 bits are just "implied."
2023-06-16 08:49:24 <KipIngram   > I guess it's about how they choose to print NUMBERS.
2023-06-16 08:49:45 <unjust      > especially where assumptions were made about what seems to work on a particular target, and the compiler didn't complain about it and things sort of worked ... until someone else comes along and gets the compiler to be fussy about warnings, errors and pedantic behaviour - and now you're compiling the code for a target machine that has a different machine word size.
2023-06-16 08:50:32 <veltas      > I've worked with lots of legacy code and I don't think I've ever seen a situation where standardese would have helped. In fact one of the common issues I see with legacy code is a pre-standard compatibility issue
2023-06-16 08:51:25 <unjust      > i've worked with a bunch of legacy C systems, and this is a recurring problem i've seen in my experience - and i'm not defending anyone spouting standardese, i'm just saying there is some value in being aware of their contents
2023-06-16 08:51:35 <veltas      > Where the standard decided that C would not allow an lvalue to remain an lvalue after a cast, and so a lot of "(volatile u32)reg = ..." had to be rewritten
2023-06-16 08:51:55 <veltas      > Lots of pre-standard compilers had more a more intelligent volatile
2023-06-16 08:54:10 <veltas      > unjust: What issue relating to the abstract machine comes up though?
2023-06-16 09:16:22 <unjust      > one i've seen repeatedly is: serialisation/deserialisation of structures to/from stream/file/socket (ie. memcpy, write, or fwrite (even seen code using strcpy for this)) with no regard for padding, alignment, or endianness - which appears fine until you have a new system (compiler and/or machine) in the mix and then sizing and alignment expectations don't match between producer and consumer
2023-06-16 09:34:57 <Zarutian_iPad> 
ACTION reads https://nathanotterness.com/2021/10/tiny_elf_modernized.html

2023-06-16 09:55:43 <veltas      > Zarutian_iPad: And yet the tiny ELF file might take up more memory than a padded ELF file, because they are padded to allow paging the whole thing back to the ELF on disk while it's not being used
2023-06-16 09:56:29 <veltas      > Although I doubt it really makes any difference but that's the first thing that came to mind, given this is one of the motivations for ELF
2023-06-16 09:57:55 <Zarutian_iPad> yebb but I like this article as an link resource to the ELF binary spec and x86_64 Linux syscall numbers searchable database
2023-06-16 09:58:39 <veltas      > Yes true
2023-06-16 09:59:00 <veltas      > It's useful to me, I might need to write some simple ELF generation soon
2023-06-16 09:59:13 <veltas      > (Working on a simple compiler)
2023-06-16 09:59:54 <Zarutian_iPad> semi-goal of mine is to write an ELF executable that has two segments one 1k .text executable and one big .data one for forth memory and program (plus I would restrict the syscall interface to the CloudABI subset)
2023-06-16 10:00:10 <unjust      > i'm impressed that it doesn't segfault with that crazy program header
2023-06-16 10:00:38 <unjust      > Type           Offset   VirtAddr           PhysAddr           FileSiz  MemSiz   Flg Align
2023-06-16 10:00:53 <unjust      > LOAD           0x000000 0x0000000000028000 0x050fff313cb0050f 0x000078 0x000078 R E 0x77202c6f6c6c6548
2023-06-16 10:07:29 <unjust      > i noticed that if the segment associated with my .text section in the ELF executable i generated from a script was any less than my system's 'Hugepagesize' (2048 bytes according to my linux/x86-64 kernel) it'd crash)
2023-06-16 11:11:14 <KipIngram   > I really don't have a firm grip on the whole point of hugepages.
2023-06-16 11:11:26 <KipIngram   > That is, why they seem to be treated separately from other RAM allocations.
2023-06-16 11:20:31 <drakonis    > https://lwn.net/Articles/374424/ you've seen this, yeah?
2023-06-16 11:24:39 <KipIngram   > Nope - thanks.
2023-06-16 11:27:45 <drakonis    > might be somewhat out of date given its age
2023-06-16 11:28:40 <drakonis    > 13 years is a lot of time in linux years
2023-06-16 12:59:22 <KipIngram   > Indeed, but it's still good background knowledge.
2023-06-16 17:46:51 <veltas      > KipIngram: https://blog.nelhage.com/post/transparent-hugepages/
2023-06-16 17:48:47 <drakonis    > that's for the more recent usage of hugepages in linux
2023-06-16 17:52:28 <veltas      > Yes but this article explains why hugepages were originally handled separately
2023-06-16 17:52:48 <veltas      > It seems like they never solved that problem and just made it automatic later without proper understanding
2023-06-16 17:53:11 <veltas      > But it gives a good reason why they were originally specified differently
2023-06-16 17:53:41 <veltas      > The other reason is it would have increased kernel memory usage for your hundreds of processes that don't use or need any huge pages