RES: [UMN_MAPSERVER-USERS] RES: [UMN_MAPSERVER-USERS] php/mapscript bug - URGENT

[Ed McNierney]

Murilo -

Yes, I've just started using FastCGI with MapServer, and I'm quite happy with the results.

     - Ed

-----Original Message-----
From: UMN MapServer Users List [mailto:MAPSERVER-USERS at LISTS.UMN.EDU] On Behalf Of Murilo Lacerda Yoshida
Sent: Monday, February 14, 2005 11:56 PM
To: MAPSERVER-USERS at LISTS.UMN.EDU
Subject: [UMN_MAPSERVER-USERS] RES: [UMN_MAPSERVER-USERS] RES: [UMN_MAPSERVER-USERS] php/mapscript bug - URGENT

  Ed,
  I'm comparing the same maps with the same operations (zoom and pan).
With cgi it takes a lot more time to draw the same map. But now with
fastcgi mapscript is drawing nice.
  Just a note, I am in a local network with the server, so maybe this
affects the results.
  

  Murilo

-----Mensagem original-----
De: Ed McNierney [mailto:ed at topozone.com] 
Enviada em: terça-feira, 15 de fevereiro de 2005 00:48
Para: Murilo Lacerda Yoshida; MAPSERVER-USERS at LISTS.UMN.EDU
Assunto: RE: [UMN_MAPSERVER-USERS] RES: [UMN_MAPSERVER-USERS]
php/mapscript bug - URGENT

Murilo -

Any software can be made to run as quickly as you like, IF you do not
require it to give you the right answer!

I am a little surprised that you see that much difference between two
situations when you're looking at one image being drawn one time.  Are
you sure you are comparing similar situations?

     - Ed

Ed McNierney
TopoZone.com

-----Original Message-----
From: UMN MapServer Users List [mailto:MAPSERVER-USERS at LISTS.UMN.EDU] On
Behalf Of Murilo Lacerda Yoshida
Sent: Monday, February 14, 2005 11:40 PM
To: MAPSERVER-USERS at LISTS.UMN.EDU
Subject: [UMN_MAPSERVER-USERS] RES: [UMN_MAPSERVER-USERS] php/mapscript
bug - URGENT

  Steve,
  I installed the second php on my Server, as a cgi. It worked, the bug
isn't there anymore.
  But is amazing how the mapscript performance working with php as cgi
is so much lower than that with php as a dso... If the image I want to
see has never been drawed in that session, it takes around 4 seconds to
draw it (I have direct access to the server trough the lan), while with
dso it takes less than a second in the worst case...
  Do you know if the proj and shape functions of mapscript are also
affected by this bug? If I understood it right the bug is related to
thread problems, right? So there is a chance that this functions (shape
and proj) are affected as they are called by a thread?

  Do you know if mapserver works with apache 2.0 and php5 with cgid?
There is a rumor that is faster than regular cgi.....
  Thanks,
    Murilo

-----Mensagem original-----
De: UMN MapServer Users List [mailto:MAPSERVER-USERS at LISTS.UMN.EDU] Em
nome de Stephen Woodbridge
Enviada em: terça-feira, 15 de fevereiro de 2005 00:08
Para: MAPSERVER-USERS at LISTS.UMN.EDU
Assunto: Re: [UMN_MAPSERVER-USERS] php/mapscript bug - URGENT

Murilo,

I'm not sure anyone can say it works definitively, so the safe answer is
do not use the DSO version, especially if you are have problems with it.
Mapserver working in a DSO is very new and we as a group have very
little experience with it.

If you can afford to help us out testing and reporting results using the
DSO version that would be grand, but if you need to get you site up and
working reliably, we all know that the cgi stuff works and works well.
You can run all your other non mapscript code using the DSO, just
configure apache to use the CGI version for the directories that contain
mapscript.

-Steve W.

Murilo Lacerda Yoshida wrote:
>   Hi all,
>   This is about that saveWebImage bug, that shows random images to the
> user.
>   I am going to install a second php on my server, as a cgi, and run
> mapscript on it. The problem is that I have several pages that do not
> show images on screen, these pages does some stuff like creating
> shapefiles, making a projection conversion, this kind of stuff.
>
>   What I wanted to know is if this image bug is restricted to the
> images... Can I still use mapscript with php-dso when I don't how
> images?
>   I think that I can, but I'm not sure.
>   Thanks,
>     Murilo
>