terrible programmers: my god, it's full of chars

The Something Awful Forums > Discussion > Serious Hardware/Software Crap > YOSPOS > terrible programmers: my god, it's full of chars

«‹›1446 »

JawnV6: Jul 4, 2004; So hot ...

Arcsech posted:

maybe i have stockholm syndrome for bit-twiddling c, but that actually isn't too bad to read to figure out what its doing

but if you dont have a comment explaining why its doing that, then gently caress you

i ended up trashing it in favor of using a ring buffer and not tracking per-block usage, but it made it into version control

Arcsech posted:

also use #defines for the 3 and 0x7 so its clearer about what those actually represent

and avoid the type system!?!?

it probably makes more sense on a few lines as

code:

uint8_t byte_index = (block >>3);
uint8_t bit_index = (block&0x7);
block_used[byte_index] |= (1<<bit_index);

or just to be terrible about it

code:

 typedef struct block_index {
    unsigned int bit_index:3;
    unsigned int byte_index:5;
} block_index_t;
block_index_t bInd = (block_index_t) block;
block_used[bInd.byte_index] |= (1<<bInd.bit_index);

# ? Dec 2, 2015 20:46

Adbot: ADBOT LOVES YOU

# ? May 11, 2024 22:55

Jeffrey of YOSPOS: Dec 22, 2005; GET LOSE, YOU CAN'T COMPARE WITH MY POWERS

My paranoid rear end would put __attribute__ ((__packed__)) on that struct, dunno if it's really needed.

# ? Dec 2, 2015 20:51

Zaxxon: Feb 14, 2004; Wir Tanzen Mekanik

Jeffrey of YOSPOS posted:

Improvements are pretty ugly too:
code:
block_used[(block >> BLOCK_FIELD2_SHIFT) & BLOCK_FIELD2_MASK]  |= ((1 << (block >> BLOCK_FIELD1_SHIFT) & BLOCK_FIELD1_MASK);
You can do more tricks to make it smaller if you do this all with macros or something, but it's ugly either way.

I think normally the nice thing to do would be to make (block >> BLOCK_FIELD2_SHIFT) & BLOCK_FIELD2_MASK type things local variables so you get

code:

uint8_t block_addr = block>>3;
uint8_t block_set_bit = 1<<(block & 0x07);

block[block_addr] |= block_set_bit;

# ? Dec 2, 2015 20:51

Zaxxon: Feb 14, 2004; Wir Tanzen Mekanik

JawnV6 posted:

i ended up trashing it in favor of using a ring buffer and not tracking per-block usage, but it made it into version control

and avoid the type system!?!?

it probably makes more sense on a few lines as
code:
uint8_t byte_index = (block >>3);
uint8_t bit_index = (block&0x7);
block_used[byte_index] |= (1<<bit_index);
or just to be terrible about it
code:
 typedef struct block_index {
    unsigned int bit_index:3;
    unsigned int byte_index:5;
} block_index_t;
block_index_t bInd = (block_index_t) block;
block_used[bInd.byte_index] |= (1<<bInd.bit_index);

yeah do that first one.

# ? Dec 2, 2015 20:53

NihilCredo: Jun 6, 2011; iram omni possibili modo preme:
plus una illa te diffamabit, quam multæ virtutes commendabunt

Shaggar posted:

I did a powershell script for a thing and powershell is still kind of dumb because why didn't they just use c# syntax, but the one thing I do like is you can sign scripts and require signatures for scripts to run. that's cool and way better than Linux where the bad scripting language is not just the UI, but its unrestricted.

the new visual studio update has the c# interpreter in it. I'm actually not sure if you can run c# script files with it but I presume so since it seems a straight clone of f#'s fsi.exe

of course neither are shipped with windows unlike power shell. this may or may not be a showstopper fit you. sadly it is for me

# ? Dec 2, 2015 22:24

sarehu: Apr 20, 2007; (call/cc call/cc)

JawnV6 posted:

code:
block_used[(block>>3)] |= (1<<(block&0x7));
this is the kind of thing people hate, right?

No it's good.

JawnV6 posted:

code:

uint8_t byte_index = (block >>3);
uint8_t bit_index = (block&0x7);
block_used[byte_index] |= (1<<bit_index);

Now your byte index has been truncated to an 8-bit value.

Using 0x7 is nauseating though.

# ? Dec 2, 2015 22:29

Jeffrey of YOSPOS: Dec 22, 2005; GET LOSE, YOU CAN'T COMPARE WITH MY POWERS

please don't use bitwise operators with decimal numbers

# ? Dec 2, 2015 23:06

Soricidus: Oct 21, 2010; freedom-hating statist shill

Arcsech posted:

maybe i have stockholm syndrome for bit-twiddling c, but that actually isn't too bad to read to figure out what its doing

but if you dont have a comment explaining why its doing that, then gently caress you

also use #defines for the 3 and 0x7 so its clearer about what those actually represent

#define SHIFT_VALUE_THREE 3
#define MASK_VALUE_SEVEN 0x7

# ? Dec 2, 2015 23:13

Shaggar: Apr 26, 2006

NihilCredo posted:

the new visual studio update has the c# interpreter in it. I'm actually not sure if you can run c# script files with it but I presume so since it seems a straight clone of f#'s fsi.exe

of course neither are shipped with windows unlike power shell. this may or may not be a showstopper fit you. sadly it is for me

Windows doesn't ship with updated PowerShell either. I had to install PowerShell 4 to get a cmdlet i wanted. Idk why it's not distributed via windows update.

# ? Dec 2, 2015 23:15

DONT THREAD ON ME: Oct 1, 2002; by Nyc_Tattoo; Floss Finder

i tried to do some stuff with node today because i wanted to demonstrate that some json-schema poo poo wasn't working, and i wanted to use like, a well vetted json-schema library to show it (rather than the one i was using )

turns out all the js json schema validators on node are like, kinda poo poo?

i found a library that would asynchronously pull in missing schemas so i used that one.

it failed because the library used jquery to do its async stuff.

jquery requires a window.document to work (apparently)

a library, available in node, uses jquery to do async stuff, and cant, because node does not have a browser window. shouldn't node have all kinds of much better things than jquery to do async things??

i just dont get it. where does browser JS end and node js begin?

# ? Dec 2, 2015 23:23

pseudorandom name: May 6, 2007

Shaggar posted:

Windows doesn't ship with updated PowerShell either. I had to install PowerShell 4 to get a cmdlet i wanted. Idk why it's not distributed via windows update.

it does get distributed via windows update, its just bundled with some other stuff

you can read about it in KB3035583

# ? Dec 2, 2015 23:28

zeekner: Jul 14, 2007

MALE SHOEGAZE posted:

i just dont get it. where does browser JS end and node js begin?

people use NPM and webpack to do node-style development targetting browsers, so there are a ton of packages that are browser-only on npm

no way to tell unless they specifically state it in their docs

lol, js docs

# ? Dec 2, 2015 23:31

sarehu: Apr 20, 2007; (call/cc call/cc)

pseudorandom name posted:

KB3035583

Trigger warning please

# ? Dec 2, 2015 23:54

JawnV6: Jul 4, 2004; So hot ...

sarehu posted:

Using 0x7 is nauseating though.

0b111?

# ? Dec 2, 2015 23:57

Maluco Marinero: Jan 18, 2001; Damn that's a
fine elephant.

Uncomfortable Gaze posted:

lol, js docs

# ? Dec 3, 2015 00:16

Jerry Bindle: May 16, 2003

JawnV6 posted:

0b111?

i wish c had binary literals

# ? Dec 3, 2015 00:16

Bloody: Mar 3, 2013

JawnV6 posted:

code:
block_used[(block>>3)] |= (1<<(block&0x7));
this is the kind of thing people hate, right?

code:

block_used(block >> 3)] |= 1 << (block & 0x07);

# ? Dec 3, 2015 00:23

MrMoo: Sep 14, 2000

Barnyard Protein posted:

i wish c had binary literals

apparently GCC supports them.

# ? Dec 3, 2015 00:38

qntm: Jun 17, 2009

Barnyard Protein posted:

i wish c had binary literals

it boggles my mind that there exists a single programming language in the whole world which doesn't support binary literals, let alone that C of all languages has this problem

# ? Dec 3, 2015 00:41

Jerry Bindle: May 16, 2003

oh awesome! thanks! sometimes wishes do come true

# ? Dec 3, 2015 00:42

suffix: Jul 27, 2013; Wheeee!

JawnV6 posted:

code:
block_used[(block>>3)] |= (1<<(block&0x7));
this is the kind of thing people hate, right?

code:

/*
 * Mark a block as used in the used block bitmap.
 *
 * Passing a block_num >= MAX_BLOCK_NUM results in undefined behaviour, the caller must check this!
 */
inline void mark_block_as_used(uint8_t used_block_bitmap[], size_t block_num) {
        assert(block_num < MAX_BLOCK_NUM);
        // Used blocks are stored as a bitmap, so we divide block_num with eight to get the byte,
        // then take block_num modulo eight to find the bit position to set.
        used_block_bitmap[block_num >> 3] |= 1 << (block_num & 0x7);
}

# ? Dec 3, 2015 00:59

Soricidus: Oct 21, 2010; freedom-hating statist shill

Assertion Failed: Abort=Quit, Retry=Debug, Ignore=Continue

idk if this is vs or .net or windows or what but lmfao at whatever it is

# ? Dec 3, 2015 01:10

Bloody: Mar 3, 2013

JawnV6 posted:

0b111?

4'b0111?

# ? Dec 3, 2015 01:10

Shaggar: Apr 26, 2006

pseudorandom name posted:

it does get distributed via windows update, its just bundled with some other stuff

you can read about it in KB3035583

I had the machine up to date and still had to install it manually.

# ? Dec 3, 2015 01:26

JawnV6: Jul 4, 2004; So hot ...

Barnyard Protein posted:

i wish c had binary literals

i tested that before i posted, WORKSFORME on my dodgy gcc shard that doesn't respect falter cases

Bloody posted:

4'b0111?

psh, i wish i could still specify things outside these crude 8-bit boundaries

although i am using a part with 24bit results

# ? Dec 3, 2015 01:45

Luigi Thirty: Apr 30, 2006; Emergency confection port.

i updated vs2015 to update 1 and now it doesn't work! it just freezes on startup

# ? Dec 3, 2015 02:31

Plorkyeran: Mar 22, 2007; To Escape The Shackles Of The Old Forums, We Must Reject The Tribal Negativity He Endorsed

Barnyard Protein posted:

i wish c had binary literals

just use c++14

# ? Dec 3, 2015 04:12

fritz: Jul 26, 2003

JawnV6 posted:

0b111?

# ? Dec 3, 2015 04:29

triple sulk: Sep 17, 2014

Luigi Thirty posted:

i updated vs2015 to update 1 and now it doesn't work! it just freezes on startup

are you opening a project right from startup or is it just the home page

# ? Dec 3, 2015 06:41

eschaton: Mar 7, 2007; Don't you just hate when you wind up in a store with people who are in a socioeconomic class that is pretty obviously about two levels lower than your own?

JawnV6 posted:

or just to be terrible about it

code:

 typedef struct block_index {
    unsigned int bit_index:3;
    unsigned int byte_index:5;
} block_index_t;
block_index_t bInd = (block_index_t) block;
block_used[bInd.byte_index] |= (1<<bInd.bit_index);

you can't just cast that

do something like

code:


typedef union block_index {
    uint8_t raw;
    struct {
    unsigned int bit_index:3;
    unsigned int byte_index:5;
    } cooked;
} block_index_t;
block_index_t bInd;
bInd.raw = block;
block_used[bInd.cooked.byte_index] |= (1<<bInd.cooked.bit_index);

# ? Dec 3, 2015 09:38

redleader: Aug 18, 2005; Engage according to operational parameters

Shaggar posted:

I did a powershell script for a thing and powershell is still kind of dumb because why didn't they just use c# syntax, but the one thing I do like is you can sign scripts and require signatures for scripts to run. that's cool and way better than Linux where the bad scripting language is not just the UI, but its unrestricted.

for a language designed around piping objects around, it's baffling that powershell has no convenient way of creating a custom object to sling through the pipeline. i'd really, really like the ability to jam a few fields into some ad hoc object, a la c# or js, but in powershell creating an object is just gross

i really wanted to like powershell, but there are so many annoying gotchas (how do you gently caress up multidimensional lists so badly?) that i just can't. it's so, so awkward in practice. powershell should just have been a more dynamically-typed c# with some additional linqy stuff and syntax to make typical scripty tasks easier

# ? Dec 3, 2015 11:14

~Coxy: Dec 9, 2003; R.I.P. Inter-OS Sass - b.2000AD d.2003AD

Soricidus posted:

Assertion Failed: Abort=Quit, Retry=Debug, Ignore=Continue

idk if this is vs or .net or windows or what but lmfao at whatever it is

it's an assertion; it's meant to be a low level failsafe for programmers and it doesn't get included in release builds

# ? Dec 3, 2015 12:08

Soricidus: Oct 21, 2010; freedom-hating statist shill

~Coxy posted:

it's an assertion; it's meant to be a low level failsafe for programmers and it doesn't get included in release builds

yes? that doesn't really excuse a crazy interface where the buttons have incorrect labels and the correct labels are given in the dialog title instead

# ? Dec 3, 2015 12:48

Luigi Thirty: Apr 30, 2006; Emergency confection port.

triple sulk posted:

are you opening a project right from startup or is it just the home page

the start page

I get that Windows error sound and it acts like there's a modal dialog up but nothing appears

# ? Dec 3, 2015 15:39

Shaggar: Apr 26, 2006

redleader posted:

for a language designed around piping objects around, it's baffling that powershell has no convenient way of creating a custom object to sling through the pipeline. i'd really, really like the ability to jam a few fields into some ad hoc object, a la c# or js, but in powershell creating an object is just gross

i really wanted to like powershell, but there are so many annoying gotchas (how do you gently caress up multidimensional lists so badly?) that i just can't. it's so, so awkward in practice. powershell should just have been a more dynamically-typed c# with some additional linqy stuff and syntax to make typical scripty tasks easier

you can create maps like $aThing = @{ SomeProp="fffff", whatever="guuuuhhhhh" } and then $aThing.SomeProp or $aThing.whatever. I haven't done much diving into the syntax cause I was doing something pretty simple.

# ? Dec 3, 2015 15:39

Shaggar: Apr 26, 2006

I would guess that maybe

$butts = @{ boners=@{ whatever="FFFF"} } might work too? idk.

# ? Dec 3, 2015 15:40

Baxate: Feb 1, 2011

MALE SHOEGAZE posted:

i tried to do some stuff with node today because i wanted to demonstrate that some json-schema poo poo wasn't working, and i wanted to use like, a well vetted json-schema library to show it (rather than the one i was using )

turns out all the js json schema validators on node are like, kinda poo poo?

i found a library that would asynchronously pull in missing schemas so i used that one.

it failed because the library used jquery to do its async stuff.

jquery requires a window.document to work (apparently)

a library, available in node, uses jquery to do async stuff, and cant, because node does not have a browser window. shouldn't node have all kinds of much better things than jquery to do async things??

i just dont get it. where does browser JS end and node js begin?

doesn't node like fire up a whole freaking instance of chrome in order to do things
last time i tried to install node with homebrew it would like try to compile chrome from source and i'm like yeah... nope ^C

# ? Dec 3, 2015 16:17

Luigi Thirty: Apr 30, 2006; Emergency confection port.

http://youtu.be/FuMUYFFSV7Y

behold the power of the laserdisc

now I want to know what that ran on in 1982

e: I can find out if I want to spend 15 dollars http://spie.org/Publications/Proceedings/Paper/10.1117/12.935588

Luigi Thirty fucked around with this message at 16:33 on Dec 3, 2015

# ? Dec 3, 2015 16:30

triple sulk: Sep 17, 2014

swift got open sourced it seems like (or will be)

# ? Dec 3, 2015 16:46

Adbot: ADBOT LOVES YOU

# ? May 11, 2024 22:55

leftist heap: Feb 28, 2013; Fun Shoe

i like swift. i hope the OS version takes off but i have my doubts

# ? Dec 3, 2015 16:49

The Something Awful Forums > Discussion > Serious Hardware/Software Crap > YOSPOS > terrible programmers: my god, it's full of chars

«‹›1446 »