Re: [PATCH] block/curl: rewrite http header parsing function

[Kevin Wolf]

Am 29.06.2024 um 16:25 hat Michael Tokarev geschrieben:
> Existing code was long, unclear and twisty.
> 
> Signed-off-by: Michael Tokarev <mjt@tls.msk.ru>
> ---
>  block/curl.c | 44 ++++++++++++++++++--------------------------
>  1 file changed, 18 insertions(+), 26 deletions(-)
> 
> diff --git a/block/curl.c b/block/curl.c
> index 419f7c89ef..9802d0319d 100644
> --- a/block/curl.c
> +++ b/block/curl.c
> @@ -210,37 +210,29 @@ static size_t curl_header_cb(void *ptr, size_t size, 
> size_t nmemb, void *opaque)
>  {
>      BDRVCURLState *s = opaque;
>      size_t realsize = size * nmemb;
> -    const char *header = (char *)ptr;
> -    const char *end = header + realsize;
> -    const char *accept_ranges = "accept-ranges:";
> -    const char *bytes = "bytes";
> +    const char *p = ptr;
> +    const char *end = p + realsize;
> +    const char *t = "accept-ranges : bytes "; /* A lowercase template */

I don't think spaces between the field name and the colon are allowed
in the spec (and in the old code), only before and after the value.

> -    if (realsize >= strlen(accept_ranges)
> -        && g_ascii_strncasecmp(header, accept_ranges,
> -                               strlen(accept_ranges)) == 0) {
> -
> -        char *p = strchr(header, ':') + 1;
> -
> -        /* Skip whitespace between the header name and value. */
> -        while (p < end && *p && g_ascii_isspace(*p)) {
> -            p++;
> -        }
> -
> -        if (end - p >= strlen(bytes)
> -            && strncmp(p, bytes, strlen(bytes)) == 0) {
> -
> -            /* Check that there is nothing but whitespace after the value. */
> -            p += strlen(bytes);
> -            while (p < end && *p && g_ascii_isspace(*p)) {
> -                p++;
> -            }
> -
> -            if (p == end || !*p) {
> -                s->accept_range = true;
> +    /* check if header matches the "t" template */
> +    for (;;) {
> +        if (*t == ' ') { /* space in t matches any amount of isspace in p */
> +            if (p < end && g_ascii_isspace(*p)) {
> +                ++p;
> +            } else {
> +                ++t;
>              }
> +        } else if (*t && p < end && *t == g_ascii_tolower(*p)) {
> +            ++p, ++t;
> +        } else {
> +            break;
>          }
>      }
>  
> +    if (!*t && p == end) { /* if we managed to reach ends of both strings */
> +        s->accept_range = true;
> +    }

Maybe make the generic comparison with a template a separate function
(maybe even in cutils.c?) so that curl_header_cb() essentially only has
something like this any more:

if (!qemu_memcasecmp_space(ptr, end, "accept-ranges: bytes ")) {
    s->accept_range = true;
}

(A better name for the function would be preferable, of course. Maybe
also a bool return value, but if it has a name related to memcmp() or
strcmp(), then 0 must mean it matches.)

Then this would really highlight the curl specific logic rather than the
string parser in curl_header_cb().

Kevin