Determine supported image file types by checking the file header #36

Damien-Chen · 2024-08-22T12:09:53Z

Based on the PNG header format and JPG/JPEG header format。Parsing the header of image to determine if the supported format is PNG or JPEG.

src/image.c

jserv

Use git rebase -i to squash the commits and improve the git commit messages.

src/image.c

jserv · 2024-08-22T17:31:27Z

Consider the following:

#if __BYTE_ORDER == __BIG_ENDIAN
static const uint8_t header_png[] = {0x89, 0x50, 0x4E, 0x47, 0x0D, 0x0A, 0x1A, 0x0A};
#else
static const uint8_t header_png[] = {0x0A, 0x1A, 0x0A, 0x0D, 0x47, 0x4E, 0x50, 0x89};
#endif
static const uint8_t header_jpeg[] = {0xFF, 0xD8, 0xFF};

static twin_image_format_t image_type_detect(const char *path)
{   
    twin_image_format_t type = IMAGE_TYPE_unknown;
    FILE *file = fopen(path, "rb");
    ...
    
    if (bytes_read < 8) /* incomplete image file */
        return IMAGE_TYPE_unknown;
#if LOADER_HAS(PNG)
    else if (!memcmp(header, header_png, sizeof(header_png)))
        type = IMAGE_TYPE_png;
#endif
#if LOADER_HAS(JPEG)
    else if (!memcmp(header, header_jpeg, sizeof(header_jpeg)))
        type = IMAGE_TYPE_jpeg;
#endif

    /* otherwise, unsupported format */                                                                                                                       
    return type;
}

src/image.c

jserv

For git commit messages, elaborate on the shortcomings of relying solely on file extensions for file type identification, and discuss the advantages of verifying image files by examining their headers. Include why reading the actual file data provides more reliable file type validation compared to trusting the extension alone.

Always squash git commits for changes serving a single purpose.

src/image.c

Damien-Chen · 2024-08-22T18:48:34Z

For my knowlege
The format of a file is determined by the bit layout in the file. The extension is a convenient human-readable piece of metadata that has nothing whatsoever to do with the contents of a file. One can easily changing file extension, which means a .jpg might not actually be a JPEG image but could be another file type entirely, such as a harmful executable.

Changing a file’s extension by renaming it does not change the internal header format that identifies the file format. If determine file format by only its extension, it may cause incorrect result.

File headers contain unique, format-specific signatures, which offers a more stable way to identify the true nature of a file.

jserv · 2024-08-22T18:52:53Z

For my knowlege ...

Refine the git commit messages accordingly.

jserv

Read https://cbea.ms/git-commit/ carefully and enforce the rules.
Remove the leading "Fix:", which looks a bit irrelevant to the proposed changes.

jserv

Read https://cbea.ms/git-commit/ carefully and enforce the rules.

Previously, the image format was determined by checking the file extension. The format of a file is determined by its internal bit layout rather than its extension. File extensions are merely human-readable metadata and do not reflect the actual contents of a file. For instance, renaming a .jpg file to another extension does not alter its internal format and could result in a file of a completely d ifferent type, such as a harmful executable. Relying solely on file extensions to determine file format is unreliable, as it may lead to incorrect results. Instead, file headers contain unique, format-specific signatures that provide a more accurate and stable method for identifying a file's true nature. Therefore, I updates the implementation to determine the image format by checking the file header and remove previous checking for file extension.

jserv · 2024-08-23T15:27:02Z

Thank @Damien-Chen for contributing! I amended the commit messages.