Skip to content

Commit

Permalink
bin/base64: decode: ignore newlines
Browse files Browse the repository at this point in the history
By popular demand, ignore newlines in the encoded input. This achieves
bug compatibility with GNU base64.

The algorithm is quite naive and slow because it checks each byte
independently. There is definitely room for improvement.
  • Loading branch information
aklomp committed Jan 10, 2024
1 parent 6756832 commit 335c27b
Showing 1 changed file with 56 additions and 10 deletions.
66 changes: 56 additions & 10 deletions bin/base64.c
Original file line number Diff line number Diff line change
Expand Up @@ -340,29 +340,75 @@ encode (const struct config *config, struct buffer *buf)
return true;
}

static int
static inline size_t
find_newline (const char *p, const size_t avail)
{
// This is very naive and can probably be improved by vectorization.
for (size_t len = 0; len < avail; len++) {
if (p[len] == '\n') {
return len;
}
}

return avail;
}

static bool
decode (const struct config *config, struct buffer *buf)
{
size_t nread, nout;
size_t avail;
struct base64_state state;

// Initialize the decoder's state structure.
base64_stream_decode_init(&state, 0);

// Read encoded data into the buffer. Use the smallest buffer size to
// be on the safe side: the decoded output will fit the raw buffer.
while ((nread = fread(buf->enc, 1, BUFFER_RAW_SIZE, config->fp)) > 0) {
while ((avail = fread(buf->enc, 1, BUFFER_RAW_SIZE, config->fp)) > 0) {
char *start = buf->enc;
char *outbuf = buf->raw;
size_t ototal = 0;

// By popular demand, this utility tries to be bug-compatible
// with GNU `base64'. That includes silently ignoring newlines
// in the input. Tokenize the input on newline characters.
while (avail > 0) {

// Find the offset of the next newline character, which
// is also the length of the next chunk.
size_t outlen, len = find_newline(start, avail);

// Ignore empty chunks.
if (len == 0) {
start++;
avail--;
continue;
}

// Decode the input into the raw buffer.
if (base64_stream_decode(&state, buf->enc, nread,
buf->raw, &nout) == 0) {
fprintf(stderr, "%s: %s: decoding error\n",
config->name, config->file);
return false;
// Decode the chunk into the raw buffer.
if (base64_stream_decode(&state, start, len,
outbuf, &outlen) == 0) {
fprintf(stderr, "%s: %s: decoding error\n",
config->name, config->file);
return false;
}

// Update the output buffer pointer and total size.
outbuf += outlen;
ototal += outlen;

// Bail out if the whole string has been consumed.
if (len == avail) {
break;
}

// Move the start pointer past the newline.
start += len + 1;
avail -= len + 1;
}

// Append the raw data to the output stream.
if (write_stdout(config, buf->raw, nout) == false) {
if (write_stdout(config, buf->raw, ototal) == false) {
return false;
}
}
Expand Down

0 comments on commit 335c27b

Please sign in to comment.