On 09/26/2012 06:17 PM, Martin Storsjö wrote: > On Wed, 26 Sep 2012, Luca Barbato wrote: > >> On 09/23/2012 12:47 PM, Martin Storsjö wrote: >>> From: Duncan Salerno <[email protected]> >> >>> --- >>> libavformat/http.c | 4 ++-- >>> 1 file changed, 2 insertions(+), 2 deletions(-) >>> >>> diff --git a/libavformat/http.c b/libavformat/http.c >>> index 041a9b0..426e859 100644 >>> --- a/libavformat/http.c >>> +++ b/libavformat/http.c >>> @@ -33,7 +33,7 @@ >>> only a subset of it. */ >>> >>> /* used for protocol handling */ >>> -#define BUFFER_SIZE 1024 >>> +#define BUFFER_SIZE 4096 >>> #define MAX_REDIRECTS 8 >>> >>> typedef struct { >>> @@ -377,7 +377,7 @@ static int http_connect(URLContext *h, const char >>> *path, const char *local_path, >>> { >>> HTTPContext *s = h->priv_data; >>> int post, err; >>> - char headers[1024] = ""; >>> + char headers[4096] = ""; >> >> why not using BUFFER_SIZE while at it? > > It's actually semantically pretty unrelated - BUFFER_SIZE is the size of > the IO buffer, while this is a buffer used for constructing headers. But > as far as I can see, the URLs aren't really ever written into this > buffer, so I don't think changing this is necessary. > >>> char *authstr = NULL, *proxyauthstr = NULL; >>> int64_t off = s->off; >>> int len = 0; >> >> Since uri are unbounded what should we decide to reject? > > Umm, what do you mean? What arbitrary max size we should choose? The > original author of this patch apparently had problems with the limit at > 1024 with some odd HLS streams, while 4096 seemed to be enough.
let's assume 4k could be a valid limit, browsers can go up to 16k safely. _______________________________________________ libav-devel mailing list [email protected] https://lists.libav.org/mailman/listinfo/libav-devel
