Public
- Public
- Network
- Groups
- Featured
- Popular
- People

Conversation

Notices

Embed this notice
:umu: :umu: (a1ba@suya.place)'s status on Friday, 11-Apr-2025 01:14:20 JST :umu: :umu:

big indian
little indian

sir please swap bytes

In conversation about 2 months ago from suya.place permalink
- Haelwenn /элвэн/ :triskell: likes this.
- Embed this notice
  Haelwenn /элвэн/ :triskell: (lanodan@queer.hacktivis.me)'s status on Friday, 11-Apr-2025 01:31:01 JST Haelwenn /элвэн/ :triskell:
  in reply to
  
  @a1ba Meanwhile me who's read https://commandcenter.blogspot.com/2012/04/byte-order-fallacy.html
  
  DO NOT SWAP
  It's just encoding/decoding, and you're using a high-level language wtf.
  In conversation about 2 months ago permalink
  Attachments
  1. No result found on File_thumbnail lookup.
    
    The byte order fallacy
    
    Whenever I see code that asks what the native byte order is, it's almost certain the code is either wrong or misguided. And if the native b...
- Embed this notice
  :umu: :umu: (a1ba@suya.place)'s status on Friday, 11-Apr-2025 01:31:02 JST :umu: :umu:
  in reply to
  
  DO NOT MEMCPY
  
  In conversation about 2 months ago permalink
- Embed this notice
  Haelwenn /элвэн/ :triskell: (lanodan@queer.hacktivis.me)'s status on Friday, 11-Apr-2025 02:30:35 JST Haelwenn /элвэн/ :triskell:
  in reply to
  - [SUYA] ひみつ :blabcatverified:
  @ivesen @a1ba Nothing wrong with that one, it's decoding.
  And BSD got {be,le}{16,32,64}{dec,enc} as well.
  
  The one that's wrong is swapping, specially mindlessly because then instead of just "okay, it happens here" you end up having to hope for even/odd amounts of swapping.
  
  In conversation about 2 months ago permalink
- Embed this notice
  [SUYA] ひみつ :blabcatverified: (ivesen@miniwa.moe)'s status on Friday, 11-Apr-2025 02:30:36 JST [SUYA] ひみつ :blabcatverified:
  in reply to
  - Haelwenn /элвэн/ :triskell:
  @lanodan @a1ba me, looking at that blog post:
  what's wrong with ntohl()? (granted, I'm not sure a 64bit equivalent exists)
  
  In conversation about 2 months ago permalink
- Embed this notice
  Haelwenn /элвэн/ :triskell: (lanodan@queer.hacktivis.me)'s status on Friday, 11-Apr-2025 02:45:24 JST Haelwenn /элвэн/ :triskell:
  in reply to
  - [SUYA] ひみつ :blabcatverified:
  @a1ba @ivesen Yeah compiler will optimise it, that said even just defining a macro for it is probably a good idea so the code is more readable and less chances of making a typo/off-by-one (at least doing it ifdef-free makes it easy to test).
  
  And well if compiler doesn't optimises it… I feel like modern CPUs will :D
  
  In conversation about 2 months ago permalink
- Embed this notice
  :umu: :umu: (a1ba@suya.place)'s status on Friday, 11-Apr-2025 02:45:25 JST :umu: :umu:
  in reply to
  - Haelwenn /элвэн/ :triskell:
  - [SUYA] ひみつ :blabcatverified:
  @lanodan @ivesen there is always code that already exists and there is code that has to be written. For the first, swapping bytes in place is the way to go (especially that there aren't many big endian systems these days, anyway). For the latter... just do byte shifting thing, compiler will optimize it into single load anyway (though some people claim not all compilers do this)
  
  In conversation about 2 months ago permalink
- Embed this notice
  Haelwenn /элвэн/ :triskell: (lanodan@queer.hacktivis.me)'s status on Friday, 11-Apr-2025 03:15:18 JST Haelwenn /элвэн/ :triskell:
  in reply to
  - [SUYA] ひみつ :blabcatverified:
  @ivesen @a1ba Huh.
  Meanwhile here it's optimising it: https://godbolt.org/z/Tfsqe4jc8
  In conversation about 2 months ago permalink
  Attachments
  1. Domain not in remote thumbnail source whitelist: raw.githubusercontent.com
    
    Compiler Explorer - C++ (x86-64 gcc 13.2)
    
    from Matt Godbolt
    
    int main() { uint8_t data[4]; ssize_t got = read(1, data, sizeof(data)); assert(got > 0); uint32_t little = (data[0]<<0) | (data[1]<<8) | (data[2]<<16) | (data[3]<<24); uint32_t big = (data[3]<<0) | (data[2]<<8) | (data[1]<<16) | (data[0]<<24); printf("little:%d big:%d", little, big); }
- Embed this notice
  [SUYA] ひみつ :blabcatverified: (ivesen@miniwa.moe)'s status on Friday, 11-Apr-2025 03:15:20 JST [SUYA] ひみつ :blabcatverified:
  in reply to
  - Haelwenn /элвэн/ :triskell:
  @lanodan @a1ba seems the BE->LE method suggested by the blog in question isn't optimized, while the byteswap that's argued against is.
  will this ever matter? :blobshrug:
  
  https://godbolt.org/z/bYcvcvMvh
  In conversation about 2 months ago permalink
  Attachments
  1. Domain not in remote thumbnail source whitelist: raw.githubusercontent.com
    
    Compiler Explorer - C
    
    from Matt Godbolt
    
    int func(char *data){ int i = (data[3]<<0) | (data[2]<<8) | (data[1]<<16) | (data[0]<<24); return i; } int func2(char *data) { int i = *((int*)data); /* swap the bytes */ i = ((i&0xFF)<<24) | (((i>>8)&0xFF)<<16) | (((i>>16)&0xFF)<<8) | (((i>>24)&0xFF)<<0); return i; } int func3(char *data) { int i = *((int*)data); return ntohl(i); }

Public

Conversation

Notices

Feeds