[hunchentoot-devel] 'max-threads' behavior for Hunchentoot

[Faré]

Dear Edi,

is there anything I can do to improve the patch that would help it
make it to the official upstream hunchentoot?

[ François-René ÐVB Rideau | Reflection&Cybernethics | http://fare.tunes.org ]
Any time you're asking the user to make a choice they don't care about,
you have failed the user — Jeff Atwood




On 24 August 2010 22:16, Faré <fahree at gmail.com> wrote:
> Dear Edi,
>
> Here's a revised patch against the current svn. It's xcvb and version
> changes reverted, otherwise what I'm running at ITA, and it seems to
> be passing the QRes precheckin. It looks like it passes the elementary
> test under Lispworks, but I don't personally have a working Lispworks
> setup, so I can't say (will it work on LW Personal?).
>
> Can you review and apply if satisfactory?
>
>> Yes, the LispWorks version of Hunchentoot (which is the original
>> Hunchentoot which only ran on LW) doesn't use any of the compatibility
>> libs like BT, usocket, and cl+ssl.  I want it to stay that way - the
>> less dependencies, the better.
>>
> Understood. I also saw that you defined trivial wrappers rather than
> imported LW symbols, and kept things that way.
>
>> Can't you just comment out all the new stuff with #-:lispworks?  As I
>> said, I'll take care of the LW version.
>>
> Scott did some work so it compiles under Lispworks, and
> though it's wholly untested, I fixed it rather than scrapped it.
> Please take it with a pinch of salt.
>
>
> Suggested commit message:
>
>   Extend Hunchentoot's 'one-thread-per-connection-taskmaster'
>     to support 'max-threads' semantics, i.e., don't create
>     a new thread if we've max out.
>
>   Add a 'pooled-thread-per-connection-taskmaster' that will
>     eventually use a thread pool, if profiling indicates.
>   Fix the 'handle-incoming-connection' to implement the
>     new behavior.
>   Add a commented-out implementation of 'accept-connections'
>     that might give better performance.  This needs to be
>     discussed with the Hunchentoot maintainers.
>
>   Address the review comments and discussions between Scott McKay
>   and Hans Huebner.
>   Also correctly issue HTTP 503 when the server runs out of threads.
>
> (Work by Scott McKay, merge by François-René Rideau)
>
> [ François-René ÐVB Rideau | Reflection&Cybernethics | http://fare.tunes.org ]
> Moving parts in rubbing contact require lubrication to avoid excessive wear.
> Honorifics and formal politeness provide lubrication where people rub together.
> Often the very young, the untraveled, the naïve, the unsophisticated deplore
> these formalities as "empty", "meaningless", or "dishonest", and scorn to use
> them. No matter how "pure" their motives, they thereby throw sand into
> machinery that does not work too well at best.
>        — Robert Heinlein, "Time Enough For Love"
>