[Zoo-discuss] Strange Apache behavior while handling statusSupported services

[Fenoy Gerald]

Hello Alexander,
nice to hear that you are using ZOO-Projet on windows platform.

I already had similar issue a long time ago, but it should be solved in the past already. Note that I have to concede that I do not use Apache on windows for more than a year now and use Microsoft IIS 7.5 on windows server (mainly because it is installed per default on many Microsoft servers). So the ZOO-Kernel code was adapted for windows platform to be used with IIS server. So first I would like to ask you if you can try to use IIS in replacement of Apache server to check if the behavior is the same or not ? If you really need to use Apache server on windows then I would like to invite you to try closing standard I/O (trying with stdin first) and check again. Once you found the correction then you can share with us then we can add a specific compilation flag specifying if the target Server is IIS or Apache to use the fix or use current code which should work correctly on IIS.

I noticed that you spoke about a 1.2.0 version of ZOO-Project, so I would like to invite you to use the latest 1.3.0 version at least as this issue should probably be solved in the new version. As you modified the code to add debuting 

Note that the ZOO-Kernel should now be able of handling reporting status from ongoing service on each supported platforms (so MacOS X, GNU/Linux and Winows).

I hope to hear back from you,
Best regards,

Gérald Fenoy

ps: I was not there yesterday, so I cannot catch you on IRC, but it was a good communication channel to get direct contact with us.

Le 27 août 2013 à 16:50, Александр Шумилов <shumsan1011 at gmail.com> a écrit :

> Hello,
>    My name is Alexander, I would like to appreciate the great job of developers of ZOO Project and ask community for piece of advice. 
>    I've got Windows Server 2003 x64 ZOO 1.2.0 installation. Couple of days ago I've managed to launch statusSupported applications (I've discovered some issues with shared memory addressing and status XML handling, I guess they were my mistakes, not bugs). So right now I am able to launch LongProcess example, but here is the thing I want to ask about - when I launch request 
> 
> http://***.ru/cgi-bin/zoo_loader.cgi?request=Execute&service=WPS&identifier=longProcess&status=true&storeExecuteResponse=true&version=1&DataInputs=sid=100&DataOutputs=executeResponseLocation
> 
> browser connection stays up until the LongProcess is over. So, when LongProcess is over, already mentioned request returns me STATUS_ACCEPTED XML. When I visit statusLocation URL, it returns me proper XML with service results / percentCompleted document.
> There is a funny thing - if I abort mentioned above request in browser on a half-way through child LongProcess, it returns me STATUS_ACCEPTED XML, just like it has to be done (I've tested same code in Linux environment). If I launch request through console (cmd), it works great, without waiting for child process.
> 
> Has anybody encountered this problem? I've put log messages all the way through the code till "return 0" in zoo_loader.c - log messages are printed in right order, in right time, but connection still stays up, so I'm not able to work normally with it from client-side.
> 
> 
> Also there is a wrapper for executing EXE-files for Windows (with query mapping through standart .zcfg file) - if anybody have done same with Linux executables, we could work together on this issue.
> 
>  
> 
> _______________________________
> С уважением,
> Шумилов Александр
> 
> Тел: 89086606450, 89501017750
> Skype: sashuk-morehod
> Email: shumsan1011 at gmail.com
> Веб-сайт: http://alexshumilov.ru
> _______________________________________________
> Zoo-discuss mailing list
> Zoo-discuss at lists.osgeo.org
> http://lists.osgeo.org/cgi-bin/mailman/listinfo/zoo-discuss