<html>
<head>
<meta http-equiv="content-type" content="text/html; charset=utf-8">
</head>
<body bgcolor="#FFFFFF" text="#000000">
Followed your suggestions in [t4] patch.<br>
<br>
Eduard.<br>
<br>
<br>
<span style="color: rgb(34, 34, 34); font-family: arial, sans-serif;
font-size: small; font-style: normal; font-variant: normal;
font-weight: normal; letter-spacing: normal; line-height: normal;
orphans: auto; text-align: start; text-indent: 0px;
text-transform: none; white-space: normal; widows: 1;
word-spacing: 0px; -webkit-text-stroke-width: 0px; display: inline
!important; float: none; background-color: rgb(255, 255, 255);"></span>2016-03-07
17:20 GMT+03:00 Amos Jeffries <a class="moz-txt-link-rfc2396E" href="mailto:squid3@treenet.co.nz"><squid3@treenet.co.nz></a>:<br>
> * To fit with the rest of the directives this should be called<br>
> server_pconn_for_nonretriable.<br>
> <br>
> * This is not a RFC violation, so you dont need to wrap any of
its code<br>
> in USE_HTTP_VIOLATIONS.<br>
> The current Squid behaviour was just a hack Alex added years
back to<br>
> prevent the old behaviour this is trying to selectively
re-enable now.<br>
> Since it was causing so many client-visible error messages when
we first<br>
> moved to HTTP/1.1 defaults.<br>
> <br>
> * What the risk is should probaly be noted in the
documentation; That<br>
> the requests it enables to re-use pconn can result in 500
errors<br>
> reaching clients/users eyes if there are any network delivery
issues, or<br>
> the server closes its connection while the request is still
bufferd by TCP.<br>
> <br>
> * only indent cf.data.pre text with 1 tab, not 2.<br>
<br>
</body>
</html>