<div dir="ltr">@amos or anyone else from dev team<div><br></div><div>Can you confirm this is intentional behavior or bug ?</div></div><div class="gmail_extra"><br><div class="gmail_quote">On Mon, Jan 2, 2017 at 9:18 AM, Alex Rousskov <span dir="ltr"><<a href="mailto:rousskov@measurement-factory.com" target="_blank">rousskov@measurement-factory.com</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">On 12/27/2016 04:50 AM, Hardik Dangar wrote:<br>
<br>
> If i remove !serverIsws somehow websockets will not work.<br>
<br>
Then there is a bug somewhere AFAICT. It is your call whether to find<br>
out what that bug is [while continuing to use a potentially dangerous<br>
workaround].<br>
<br>
Alex.<br>
<br>
<br>
> On Tue, Dec 20, 2016 at 10:27 PM, Alex Rousskov wrote:<br>
><br>
> On 12/20/2016 02:42 AM, Hardik Dangar wrote:<br>
> > Following changes in config works and whatsapp starts working,<br>
> ><br>
> > acl serverIsws ssl::server_name_regex ^w[0-9]+\.web\.whatsapp\.com$<br>
> ><br>
> > acl step1 at_step SslBump1<br>
> > ssl_bump peek step1<br>
> > ssl_bump splice serverIsws<br>
> > ssl_bump bump !serverIsws all<br>
><br>
> You do not need the "!serverIsws" part because if serverIsws matches,<br>
> then the splice rule wins, and Squid does not reach the bump rule. This<br>
> configuration is sufficient:<br>
><br>
> ssl_bump peek step1<br>
> ssl_bump splice serverIsws<br>
> ssl_bump bump all<br>
><br>
> In theory, adding "!serverIsws" does not hurt. However, negating complex<br>
> ACLs is tricky/dangerous and should be avoided when possible.<br>
><br>
> Alex.<br>
><br>
><br>
<br>
</blockquote></div><br></div>