Alex Rousskov wrote:
> On 04/19/2010 03:26 AM, Amos Jeffries wrote:
>> Kinkie wrote:
>>> On Thu, Apr 15, 2010 at 6:54 AM, Dhaval Varia <dhavalkvaria_at_gmail.com>
>>> wrote:
>>>> Dear Sir,
>>>>
>>>> Please find the attached patch file for the
>>>>
>>>> 1. TCP logging (Newly created) ------- modtcp.patch
>>> The patch is not malformed now. I still don't understand the
>>> file-related error messages, so I leave this to someone with more clue
>>> than I do.
>> Neither. They came in from ModUdp. I suspect they will never happen by
>> have not felt keen to remove them as part of this change.
>>
>>>> 2. UDP logging (Modified) ------- modudpfix.patch
>>> The test reversal seems fine. +1 (I don't understand why adding the
>>> four empty lines in logfile_mod_udp_open(); it clashes with Squid's
>>> coding style guidelines.)
>>>
>> The needed 1-byte change to UDP has been in 3.HEAD for a while now. The
>> UDP patch can be ignored now.
>>
>>>> would you please commit this TCP logging and UDP fix first??
>>>>
>>>> I will see what to do to handle back pressure later.
>>> Ok by me. It's probably implicitly done if the socket is blocking.
>>>
>> It's a regular comm_open TCP socket. So I think its non-blocking by
>> default.
>>
>> Buffering is done by the module. But we neglected to increase the buffer
>> size yet. I'm bumping it up to 64K on commit in a few minutes.
>
> As far as I could see the TCP logging code had no user-visible
> documentation and left most new types/functions undocumented as well,
> including the globally visible function. Please insist on documentation
> when accepting patches.
>
gah. quite right. Sorry.
Amos
-- Please be using Current Stable Squid 2.7.STABLE9 or 3.1.3Received on Sun May 23 2010 - 11:42:22 MDT
This archive was generated by hypermail 2.2.0 : Sun May 23 2010 - 12:00:54 MDT