8.5

Warning
It is highly recommended to make all processes using the HTTP Server Input task self-replicating and
to reduce their polling interval in the Process Properties.
Input
The HTTPServer Input task does not, by itself, capture any files. Neither does it directly wait for requests to
be received. Actually, it is the HTTPservice that receives the requests and places them in a specific location
on the drive. When a request is received, the HTTPServer Input polls that location and finds the requests
and all attachments. It will always pick up the "oldest"request received.
The request can contain one or more files, one being an XMLfile containing the request information as well
as any GETor POSTvariables that were received within this request. Other files are POSTattachments.
Technical
By default, the request XMLalso contains a CDATAsection which contains the raw input data,
effectively doubling the size of the incoming file. Due to technical restrictions, the incoming XMLfile
cannot be more than 400MB, which because of CDATA is reduced to around 200MB. To help in this
situation, you may elect to omit CDATA from the attachment, which can be changed in HTTP Server
Input User Options. Please note that this limitation also affects incoming binary files (sent through file
upload in a form), regardless of CDATA settings.
Processing
Depending on the options chosen in the HTTPServer Input task properties, the task may choose to ignore
some of the files. For example, using the "Do not include XMLenvelope"means that only the
POSTattachments will be used in the process, the XMLfile will be discarded. Attachments are always
saved on disk in a specific location, which is accessible either directly in the XML or directly as a data file
through the "Loop each attachment as data file"option.
Output
First, the output inside the process itself is, depending on the selected options, an XMLrequest files,
POSTAttachments files, either one or both.