[Mapserver-dev] FastCGI and MapServer 4.4

[Frank Warmerdam]

Daniel Morissette wrote:
> To be honest, I'm a bit worried about doing that kind of changes (mostly 
> connection pooling) at the last minute before a release, but if you are 
> confident that they can be compiled out by default and not cause any 
> side-effects and would not delay the release, then let's try to fit them 
> in 4.4.

Daniel,

I am pretty confident that the fastcgi support would essentially be compiled
out in normal builds.

For the purposes of the 4.4 release I would be happy enough to only apply
"new style" connection pooling in mapsde.c with it potentially controlled
by an #ifdef to fallback to the old style if that is an issue.  However,
ideally I would like the new connection pooling to be implemented for any
of the providers where the maintainer is willing to apply it and test it.

I have no strong personal opinion on whether it should be exposed in
MapScript.

> But once again it's not just my call: I am just trying to facilitate the 
> release process between all the developers. So far only you and Assefa 
> have replied to my call for a 4.4 release plan... what do the other 
> developers think of my proposed plan?

Well, in fact, I think we should appoint a formal release manager to make
decisions on what goes in, whether things are ready and so forth.

I hearby nominate Daniel as MapServer 4.4 release manager and give him my vote.

Of course, hopefully, even once given omnipotent power as release manager, he
would still consider the views of the developers as much as is practical.

To that end, I hearby propose that:

  1) FastCGI support go into 4.4, but only be enabled if explicitly requested
     (not just if the fastcgi libraries happen to be sitting around).

  2) Connection pooling go into 4.4 but only for layer types where someone is
     willing to integrate and test it reasonably thoroughly.  (SDE and Oracle
     would be the types that would get a huge benefit).

  3) The "raw" connection pooling API be exposed to MapScript, but documented
     as "experimental".

  4) Frank Warmerdam must have the implementation complete by October 17th.

This proposal is of course the "give me everything I want" proposal, but I am
calling for votes for and against from serious MapServer contributors.  Any
votes against (or even for) are encouraged to include reasoning and suggested
variances.

Best regards,
-- 
---------------------------------------+--------------------------------------
I set the clouds in motion - turn up   | Frank Warmerdam, warmerdam at pobox.com
light and sound - activate the windows | http://pobox.com/~warmerdam
and watch the world go round - Rush    | Geospatial Programmer for Rent