mathiasbynens / small Goto Github PK

Smallest possible syntactically valid files of different types

Home Page: https://mathiasbynens.be/notes/minimal-html

Makefile 0.16% Ada 2.15% C++ 0.96% C# 2.39% Eiffel 0.48% Fortran 0.96% Haskell 1.84% HTML 70.79% Inform 7 2.95% Java 2.63% Pascal 0.88% C 0.80% Go 2.08% Batchfile 0.08% COBOL 7.98% Scala 0.64% DIGITAL Command Language 0.08% Assembly 0.64% Mercury 0.96% Rich Text Format 0.56%

small's Introduction

Smallest possible […] file

This repository aims to collect the smallest possible syntactically valid files in different programming/scripting/markup languages.

It all started with a blog post on the smallest possible HTML/XHTML files…

Pull requests welcome!

– Mathias

License

To the extent possible under law, the author has waived all copyright and related or neighboring rights to this work.

File types

Audio

Documents

Executables

Graphics

Languages

Markup

JavaScript Object Notation .json (with padding .p)
Markdown .md
Scalable Vector Graphics .svg
toml
YAML Ain't Markup Language .yaml/.yml
Extensible Hypertext Markup Language .xhtml (basic-1.0) (1.0-frameset) (1.0-strict) (1.1) (basic-1.1) (5)
Extensible Markup Language XML 1.0 (valid) 1.1 (1.1-valid)
Hypertext Markup Language (2.0) (3.2) (4.0-strict) (4.01-strict) (4.01-frameset) (4.01-transitional) (5) iso-html

Video

Unsorted

chicken
crystal
Cascading Style Sheets .css
eiffel
elixir
Git commit
groovy
haskell_loop
haskell_term
i
International Color Consortium profile .icc
inform-6
intercal
jess
malbolge
cache manifest .appcache
opa
promela
scala
story
tads-3
typescript
vertex-shader
whitespace

small's People

Contributors

Stargazers

Watchers

Forkers

timmywil arian sigvef leecade aqua-ye gweax rbirkby kalifg chevett c0nspiracy alokmenghrajani haasn zcorpan retord jesusfuenlabrada bjornd subzey sshilko derekliang robbert unforeseenocean mems plujon phwd tetyys fuzzstati0n shritesh waldyrious lastguest statico peckjon wyding 1ternal rgbkrk agriffis kanaka sleepwalker mbje-saxo lafelabs oesterle credil pentestingtools piratetone ikarius6 nnamon smt923 m1a0rz sazoji heydon jeffsee55 mrkalach egocks fulldecent omar2002 unichat rmm5t sdkrup dmak subinacls marto1 creack saibotshamtul sky0014 metamaniac mauriciogg binji mehrdad-shokri chenadamyu spiderhands tgxworld liquorvicar gergoerdi korvinszanto linagirl ustc-zzzz vitor-aguiar martinleblanc tomedharris lallouslab hooriza 3dsexploits the-xentropy hhravn 0xrustlang lynn 0xbadca7 cittyinthecloud etiennelareau thestinkyskunk macromachine lewisje jesterjunk david-russo aanelson rainerknabenbauer pwnfuzzsec dominik-bln psaxton alphare aaronportnoy

small's Issues

Add hocon.conf

According to official specification:

Empty files are invalid documents

A valid file of HOCON format can be empty.

Request: png

I see the smallest transparent png. But is that necessarily the smallest png?

Problems with C file.

Empty string is the smallest valid c code - it has no executable code and no data, but it remains valid. You can compile it to an object file (.o) and a shared library (.so on UNIX). So the c file in this repo should be an empty file, as this is exactly the smallest valid file of this type.
What you can't do with it, is to link it as a program, as that requires exactly one main function. That is what you're aiming for in the current c file. But there are a few problems with this file (int main;):

As mentioned above, this is not the smallest possible c file.
The file produces undefined behavior. Normally, main is a function. You've declared main as an int variable. In such case, the C standard AFAIK doesn't specify what happens, although it's permitted (i.e. it's undefined behavior). What a c compiler will typically do in this case, is to put the contents of main in an executable section and call it. This actually allows to run any machine code by assigning an array to main. Anyway, in this case, calling main causes the computer to run machine code consisting of one or more zero bytes (depending on the size of int), which e.g. on my linux x86-64 system causes a segmentation fault. To fix this, you can make main an empty function, by adding {} (returning 0 from main is implicit).
The type (int) is not needed, int is an implicit type used in c when none is given. Also, the ending semicolon is not needed as well.
Given the above, I propose the following:
Make c.c an empty file.
Add another file, maybe called c_program.c, with the following contents: main(){}.

JSON file isn't the smallest

EDIT: nvm, turns out I was wrong.

html5.html is invalid

<!DOCTYPE html> isn't valid HTML5 courtesy of https://html5.validator.nu/

At a minimum, the title element is required, and it must have a value:

<!DOCTYPE html><title>_</title>

is the smallest valid HTML5 document that I'm aware of.

MP4 + HTML5 = MEDIA_ERR_SRC_NOT_SUPPORTED

When trying to use the mp4 file in a HTML5 video player, it returns a MEDIA_ERR_SRC_NOT_SUPPORTED error. Here's a couple of test pages (check the console):

Using RawGit (which appears to use wrong Content Type):
http://codepen.io/sampotts/pen/RRxwxZ

Then tried a different CDN with correct Content Type but same error:
http://codepen.io/sampotts/pen/OXzJOG

WebP

Smallest possible WebP image.

Seen it done with gif. Would be cool to see how small WebP could be.

TIFF is not valid

TIFF added in #86 is not entirely valid, and both libtiff and ImageMagick complain about it.

$ tiffinfo tiff.tif
TIFFReadDirectory: Warning, TIFF directory is missing required "StripByteCounts" field, calculating from imagelength.
$ magick identify tiff.tif
identify: TIFF directory is missing required "StripByteCounts" field, calculating from imagelength. `TIFFReadDirectory' @ warning/tiff.c/TIFFWarnings/956.
identify: improper image header `tiff.tif' @ error/tiff.c/ReadTIFFImage/1349.

As I read the standard, it should add the StripByteCounts field, and terminate the IFD with a zero long to indicate no more IFDs.

mp3 format is not correct

➜  Desktop ffprobe -i mp3.mp3
ffprobe version 4.0.1 Copyright (c) 2007-2018 the FFmpeg developers
  built with Apple LLVM version 9.1.0 (clang-902.0.39.2)
  configuration: --prefix=/usr/local/Cellar/ffmpeg/4.0.1 --enable-shared --enable-pthreads --enable-version3 --enable-hardcoded-tables --enable-avresample --cc=clang --host-cflags= --host-ldflags= --enable-gpl --enable-libmp3lame --enable-libx264 --enable-libxvid --enable-opencl --enable-videotoolbox --disable-lzma
  libavutil      56. 14.100 / 56. 14.100
  libavcodec     58. 18.100 / 58. 18.100
  libavformat    58. 12.100 / 58. 12.100
  libavdevice    58.  3.100 / 58.  3.100
  libavfilter     7. 16.100 /  7. 16.100
  libavresample   4.  0.  0 /  4.  0.  0
  libswscale      5.  1.100 /  5.  1.100
  libswresample   3.  1.100 /  3.  1.100
  libpostproc    55.  1.100 / 55.  1.100
[mp3 @ 0x7fb4d7002a00] Format mp3 detected only with low score of 1, misdetection possible!
[mp3 @ 0x7fb4d7002a00] Invalid frame size (72): Could not seek to 72.
mp3.mp3: Invalid argument

MP3 file no longer work - here's a new one

The one from #57 doesn't work anymore (DEMUXER_ERROR_COULD_NOT_OPEN: FFmpegDemuxer: open context failed in browser and Invalid frame size (72): Could not seek to 72 in ffmpeg).

Here's a new one with 76 bytes: mp3.zip

data:audio/mpeg;base64,/+MYwAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA/+MYwA==

`gif.gif` not rendering

I think it's simply because it's saved incorrectly.

This may also prove to be a useful resource: http://probablyprogramming.com/2009/03/15/the-tiniest-gif-ever

No video formats: QuickTime(QTTF) mov, avi, ogg, flv, wmv

TBD

Curious about dashes dude

Some files are like... html-9001 and others are like html5

So, like, what's the deal?

Measuring in bytes is not fully correct file must have minimal information possible too

for example https://github.com/mathiasbynens/small/blob/master/scala.scala example

class m

have amount of information ie shannon enthropy 2.52bits

so better will be

class s

it has much more less information 2.13 so better

Imagine you need to compress all that files, files with less information will be better compressed, so will have a less size.

PDF seems wasteful

Is that really the smallest? It has words!

Copy-ability

Is providing base64 equivalents a good idea?

HTML5 file has syntax errors

The current content of the HTML5 (<!DOCTYPE html>) has syntax errors:

This means that there must be present a <title> element in the document, being this the final code:

<!DOCTYPE html><title>x</title>

The same validates properly:

This was verified using https://html5.validator.nu/

java

an empty .java is compilable by javac, or even class A{} in A.java is compilable.

Smaller versions?

C - main; - who cares about result? It's syntactically valid...
zip - data:application/zip;base64,UEsFBgAAAAAAAAAAAAAAAAAA - 18 bytes

Small project based on "small"

This class responds with a proper 404 content
https://github.com/szepeviktor/wordpress-plugin-construction/blob/master/404-adaptive.php
uses a lot of "small".

Thanks.

AVIF

AV1 Image File Format (AVIF).

A 307 byte version: AAAAIGZ0eXBhdmlmAAAAAGF2aWZtaWYxbWlhZk1BMUEAAADybWV0YQAAAAAAAAAoaGRscgAAAAAAAAAAcGljdAAAAAAAAAAAAAAAAGxpYmF2aWYAAAAADnBpdG0AAAAAAAEAAAAeaWxvYwAAAABEAAABAAEAAAABAAABGgAAABkAAAAoaWluZgAAAAAAAQAAABppbmZlAgAAAAABAABhdjAxQ29sb3IAAAAAamlwcnAAAABLaXBjbwAAABRpc3BlAAAAAAAAAAEAAAABAAAAEHBpeGkAAAAAAwgICAAAAAxhdjFDgSAAAAAAABNjb2xybmNseAACAAIABoAAAAAXaXBtYQAAAAAAAAABAAEEAQKDBAAAACFtZGF0EgAKBzgADtAgIGkyDB/wYAALAAAAhPPLwA==

But there are probably way smaller sizes possible.

XMP

The smallest possible XMP (metadata file) is

<x:xmpmeta xmlns:x="adobe:ns:meta/"/>

(validated with xmp validator)

Smallest possible Ogg/WebM files

Suggested by @zcorpan:

What's the smallest possible Ogg/WebM file? @foolip suggested 0 kbps Vorbis silence or a black or green Theora or VP8 frame.

WebM
Ogg

Can't play mp4.mp4 with html5

I've setup a test page: http://filibuster.herokuapp.com/mobile-video/blank.html

I've copied mp4.mp4 to small.mp4. If you open the test page you will see that big.mp4 loads and can play no problem. However, there is an error event while loading small.mp4 and you cannot right click to play small.mp4. I'm pretty sure I am doing it right. Agree/disagree?

Any ideas? Thanks!

@kalifg

license?

It would be helpful if this repository had a license! I realize for some of the files (like the python example...) this is a ridiculous thing to request, but for the others, it would be nice to know what the terms of re-use are. Otherwise it's hard to know if/how I can use these files in my own projects.

Add SWF

https://imgur.com/a/MtQZv#6

via https://twitter.com/webtonull/status/672358982061789184

Add diff

i thought what the smallest possible (git) diff would be

the answer would fit greatly in this repo

c.c is invalid

Exporting an int as 'main' is not a legal C program. This will crash with a bus error on OS X, SEGV on Debian 8 and FreeBSD 10.

Add GIF, PNG, JPG…

GIF: http://programming.arantius.com/the+smallest+possible+gif and http://www.perlmonks.org/?node_id=7974

Minimal PDF

Not the smallest, but based on Adobe PDF specification "minimal PDF file" (aka "Hello World" PDF): Minimal PDF

Windows batch file is not valid

Request: .ico file

Apparently the smallest is 70B consisting of a 1x1 pixel image

https://jroatch.xyz/2015/blog/1x1-favicon-ico-generator

JPEG XL

@jonsneyers (who worked on the spec) has been tweeting small JPEG XL artworks recently. If anyone knows what the smallest possible valid JPEG XL document looks like, it's probably him. :)

Don't include empty files

Wouldn't it be better to include a list in the README of formats/languages for which the minimal file is an empty file?

WAV question

@angea https://code.google.com/p/corkami/source/browse/trunk/src/PoCs/posters/mini.wav is 52 bytes, but https://github.com/mathiasbynens/small/blob/master/wav.wav is only 44 bytes. Do you see anything wrong with the latter file?

See #50. cc @mgeier

Smallest possible tiff is 46 bytes

Please check this post which claims that smallest possible TIFF is 46 bytes.

Smallest possible *.json file

In accordance with RFC 4627 the smallest JSON is an empty array or object.

```
[]
```
```
{}
```

I’ll see if I can make this a pull request later on.

Add a blank w32 `.dll`

Is it useful to have a minimal .dll?
e.g. https://github.com/katahiromz/smalldll

Fonts - TTF

Maybe https://github.com/matthewlein/inline-block-no-space-font/tree/master/fonts

[Requesting comment] Smallest possible XML file

By strict reading of the XML 1.0 standard am I right to assume the smallest possible XML file is the following?

<?xml version='1.0'?><_/>

This does validate as ‘well-formed XML’ on the W3C Markup Validator. The only odd part is that the root element is seen as _/ rather than _. Looks like a bug, because it does treat the slash as a closing slash.

Not sure if I should pull, requesting comment.

A XML document is:

document      ::=   prolog element Misc*

Where:

prolog        ::=   XMLDecl? Misc* (doctypedecl Misc*)?
XMLDecl       ::=   '<?xml' VersionInfo EncodingDecl? SDDecl? S? '?>'
VersionInfo   ::=   S 'version' Eq ("'" VersionNum "'" | '"' VersionNum '"')
Eq            ::=   S? '=' S?
VersionNum    ::=   '1.' [0-9]+

And:

element       ::=   EmptyElemTag
                    | STag content ETag

With:

EmptyElemTag  ::=   '<' Name (S Attribute)* S? '/>'

While:

Name          ::=   NameStartChar (NameChar)*

In short we get this when all optional (marked by ? or *) parts are left out:

document      ::= '<?xml' S 'version' '=' ("'" '1.' [0-9]+ "'" | '"' '1.' [0-9]+ '"') '?>' '<' NameStartChar '/>'

C code is implementation dependent

According to the C99 standard:

The function called at program startup is named main. The implementation declares no
prototype for this function. It shall be defined with a return type of int and with no
parameters:
int main(void) { /\* ... \*/ }
or with two parameters (referred to here as argc and argv, though any names may be
used, as they are local to the function in which they are declared):
int main(int argc, char \*argv[]) { /\* ... \*/ }
or equivalent;10) or in some other implementation-defined manner.

Also, it requires a new line at the end:

A nonempty source file does not end in a new-line character which is not immediately
preceded by a backslash character or ends in a partial preprocessing token or
comment (5.1.1.2).

(Section J.2 Undefined behaviour)

So if you want something that should work in all implementations that conform to the standard, the file c.c has to be (return 0; is implied at the end, as per 5.1.2.2.3).

int main(void){}

(With a newline at the end, 17 bytes in total)

jpeg is not usable

Firefox, IE, and Windows Photo Viewer all claim that the 125-byte JPEG image is corrupt.

Add PL/I

a:proc options(main);end;