Re: [PATCH] pipeline-shutup.patch

I prefer disabling the parallell fetching completely. Have been bitten
too many times by errors related to the parallell processing of
pipelined requests.

Is why there nowdays is a directive to turn off these parallell fetches:
pipeline_prefetch

Debug levels now adjusted to 2 which is more apropriate.

Regards
Henrik

Miquel van Smoorenburg wrote:
>
> This patch is part of the patchset I'm using in the Debian squid-2.4.3
> package. Though the patches here are against 2.4.3, they should apply
> to 2.5 with minor difficulty.
>
> pipeline-shutup.patch Lower priority of some debug msgs of pipelining code
>
> The Debian 'apt' utility actually uses pipelining, and then the squid
> logs fill up with debug code. Probably because not many people use
> the pipelining code ;)
>
> BTW, I had to turn off pipelining in 2.2 an 2.3 because it didn't
> work very well, but I've had no complaints with 2.4 yet.
>
> --- squid-2.4.1/src/client_side.c.orig Thu Mar 22 21:05:08 2001
> +++ squid-2.4.1/src/client_side.c Tue Mar 27 17:14:05 2001
> @@ -1699,7 +1699,7 @@
> fd, storeUrl(entry), (int) http->out.offset);
> if (conn->chr != http) {
> /* there is another object in progress, defer this one */
> - debug(33, 1) ("clientSendMoreData: Deferring %s\n", storeUrl(entry));
> + debug(33, 3) ("clientSendMoreData: Deferring %s\n", storeUrl(entry));
> memFree(buf, MEM_CLIENT_SOCK_BUF);
> return;
> } else if (entry && EBIT_TEST(entry->flags, ENTRY_ABORTED)) {
> @@ -1873,7 +1873,7 @@
> * execution will resume after the operation completes.
> */
> } else {
> - debug(33, 1) ("clientKeepaliveNextRequest: FD %d Sending next\n",
> + debug(33, 3) ("clientKeepaliveNextRequest: FD %d Sending next\n",
> conn->fd);
> assert(entry);
> if (0 == storeClientCopyPending(http->sc, entry, http)) {
Received on Wed Dec 12 2001 - 16:53:24 MST