A W S - Ada Web Server
1.4 release / SOAP 1.2
Pascal Obry November 13th, 2003
We are happy to announce the availability of the AWS 1.4 release. The API
could change slightly at this stage but should be fairly stable now.
AWS stand for Ada Web Server. It is not a real Web Server like Apache. It is
a small yet powerful HTTP component to embed in any applications. It means
that you can communicate with your application using a standard Web browser
and this without the need for a Web Server. AWS is fully developed in Ada
AWS support SOAP, Server Push, HTTPS/SSL, client HTTP, hotplug modules... We
have worked very hard to make this release as stable as possible. Note that
Hotplug modules are very nice but have a potentially security hole as it is
implemented today. A new secure implementation will be proposed in a future
The SOAP implementation has been validated on http://validator.soapware.org/.
Here are the main changes since AWS 1.3 :
- Add WSDL parser and SOAP stub/skeleton generator. This implementation has
passed many tests (base and B tests) from the SOAP Interoperability Tests
Specification (see http://www.xmethods.net/ilab/ and
- wsdl2aws, new tool. Generate SOAP/AWS stub/skeleton from a WSDL
document. With this tool it is very easy to build or to connect to Web
Services. This tool use the WSDL parser and SOAP generator API. See above.
wsdl2aws has been tested with some services on the Web, one of them was
the Google Search Service WSDL specification which is quite complex.
- Fix memory leak in templates parser when using a non existent template
- Disable keep-alive connection for the last free slots. It make AWS better
for heavy-loaded servers.
- New configuration parameter Free_Slots_Keep_Alive_Limit to control
keep-alive connections on heavy-loaded servers.
- It is now possible to set the message that must be displayed when an
authentication is cancelled.
- Correctly free connection if a SOAP request fails. Fix memory leak.
- LDAP API checks for non initialized Directory object.
- Handle properly arrays of SOAP record. Only basic types were supported.
- Fix space handling bug in SOAP parser. Multiple spaces were removed
from the parsed string.
- Handle properly accents in SOAP messages. Convert properly DOM Utf8
encoded tree back to basic Ada 8 bits string.
- SOAP strings are now properly encoded using Utf8. This is needed for
- It is now possible to retrieve the running server from a callback
procedure. From there it is also possible to get the server
configuration. See AWS.Server.Get_Current routine.
- It is now possible to have access to the server internal status. See
- The Log API is now thread safe.
- Fix regression in client API on standard and proxy authentication.
- Fix two memory leaks in templates parser engine. This version has been
checked with gnatmem in a large project and should be memory leak free
at this point.
- Handle properly file upload for filenames with accents and spaces.
- Add session cleanup callback (called when a session ID has expired).
- Fix memory leak in the sessions container.
- Templates_Parser use lot less stack space than before while parsing a
template file. It is now possible to parse very large template file.
- Important Templates_Parser speed-up (between x10 and x15) for template
files containing large vector or matrix tags.
- Slots activity counters are updated only if an answer has been handled.
- The directory browser's template file can now be configured using the
- Using the session feature without server's session support is now
properly detected (Constraint_Error is raised).
- It is now possible to give back a socket, removed from the server using
Socket_Taken, to the server. This can be used to implement waiting line
for requests taking a long time without loosing the keep-alive status
and without blocking server's resources.
- Support for memory streams and compressed memory streams (deflate /
inflate) based on ZLib library.
- New routines to compress / decompress a Stream_Element_Array.
- New routines to compress / decompress a file using the Gzip encoding.
- Add new routines to retrieve the log filenames and the status of
the log files. See AWS.Server.Log.
- Plus many small fixes, enhancements, API comments, and documentation work.
Non upward compatible changes
Note that the changes listed below can introduce non upward compatibility.
In such a case we try to give proper advice on how to change the code
to work properly. Of course we try to avoid this as much as possible
but we really prefer to have a clean API instead of keeping awkward
- Properly handle SOAPAction in SOAP persistent connections. It was not
possible to change the SOAPAction value for each request. The same
SOAPAction was used for all requests over the same connection.
=> This change is not upward compatible. The SOAPAction value must be
removed from the persistent connection creation (AWS.Client.Create)
and passed to the SOAP.Client.Call.
- Change AWS.Client.SOAP_Post and SOAP.Client.Call spec for the persistent
=> This is not upward compatible but easier to use. The calls were
passing the persistent connection using an access mode. Just remove
the 'Access attribute to pass the connection object.
- Change the way Size of resources are computed. The size is now part of
the objects (File, Embedded or Stream resources). This is a better design.
AWS.Response.Stream and AWS.Response.Set.Stream does not have the stream
size as parameter. Furthermore the AWS.Response.Set.Content_Length
routine has been removed. The stream size is now given by overloading the
Size stream's method (see routine AWS.Resources.Streams.Size).
=> This is not upward compatible. Remove calls to
AWS.Response.Set.Content_Length and implement the
AWS.Resources.Streams.Size method for the stream object.
- AWS.Response.Set.Message_Body (with an access to a Stream_Element_Array)
has been removed. There was not clean way to integrate this with the ZLib
memory stream support.
=> This is not upward compatible. Add ".all" to the parameter to use the
version with a Stream_Element_Array formal parameter.
In each new version we try to be upward compatible with previous
version. This is really important, but in some cases it seems that a
"redesign" of the API would be good in the long term. All obsolescent
features will be listed in this section. Those features will be
removed in the next version. Note that you can check usage of those
features in your application by using the GNAT's -gnatwj option as we
have tagged all obsolescent features with a pragma.
Note that pragma Obsolescent and -gnatwj option is only supported
since GNAT 3.16.
use AWS.Server.Log.Start instead.
use AWS.Server.Log.Stop instead.
use AWS.Server.Start_Error instead.
use AWS.Server.Stop_Error instead.
You can have a look at docs/TODO file to see what are the topics that we will
probably implement in future releases.
NOTE: Since we have switched to the .PNG file format we have found that
Netscape Navigator is not able to display the PNG transparent layer properly!
The OpenSSL libraries (optional) distributed are for Windows only. On UNIX
you'll have to build the libraries from sources, it is quite easy to do
so. This has been tested on GNU/Linux without trouble.
The LDAP binding will use the LDAP dynamic library on Windows. On UNIX you
need to build and install OpenLDAP.
See documentation for build information.
AWS 1.4 has been compiled and has passed all tests on:
Windows XP, GNAT 3.16a1 and 5.01a
GNU/Linux x86, GNAT 3.16a1 and 5.01a
SPARC Solaris 8, GNAT 3.16a1
Others platforms / compiler version combinations have not been tested, it
does not mean that it's not working.
Previous version of AWS have been build on FreeBSD 4.1 and MacOSX.
AWS User's Mailing List:
AWS Home Page (sources and printable documentations in Postscript and PDF):
Templates_Parser module (sources and documentation) is provided with AWS
distribution. Version 6.2 is distributed with AWS 1.4.
Latest version of this module and the documentation can be found at:
Templates_Parser is a very useful add-on for AWS. You should have a look at
it if you plan to develop a Web application. Templates_Parser permits to
completely separate the HTML design from the Ada code.
Some other Templates engine are WebMacro, FreeMarker, PHP, ASP, JSP and
Velocity. All of them are based on explicit iterators (#foreach with a
variable) where Templates_Parser is based on implicit ones (you use a more
intuitive table iterator). Be sure to check the documentation. Only
the Velocity project has the goal to support complete separation of HTML
design and code.
GNU/Ada - GNAT
You need at least version 3.15 to build and use AWS 1.4.
Socket binding (Optional) :
Since AWS 1.2 you need at least version 1.0 of the Socket binding. Note
that by default AWS uses GNAT.Sockets.
You need this library only if you want to use AWS SOAP feature. You need
at least XMLada 1.0.
POSIX Binding (optional) :
OpenSSL library (optional) :
Sources for UNIX or Win32:
binaries for Win32:
Included with the main AWS distribution (win32 directory).
Note that we have used and we distribute (for Win32 platform) OpenSSL
version 0.9.7c with this AWS release. OpenSSL has been built with GNAT 5.01a
C subsystem (based on GCC 3.2.3) with -O3 optimization level.
See OpenSSL license (docs/openssl.license).
OpenLDAP library (optional) :
Sources for UNIX or Win32:
binaries for Win32:
Included with the main AWS distribution (win32 directory). The import
library will bind to the Microsoft LDAP dynamic library.
Windows Services API (optional):
To build the runme demo as a Windows NT/2000 services you must download
the services API made by Ted Dennison for his SETI@Home project.
You can report bugs to:
Dmitriy Anisimkov [log in to unmask]
Pascal Obry [log in to unmask]
It would be nice if you could also sent us a note if you are using AWS just
to know if it is used at all or not :) And if you are ok, we'll add an entry
for your project in the next section.
AWS User's Mailing List:
A good way to keep informed of AWS news and to share experiences with other
AWS users is to register to the AWS dedicated mailing list. See:
Thanks to the contributors and peoples who send feedbacks, ideas
about AWS. In the early stage of the project this is very valuable.
So thanks goes to Ted Dennison, Wiljan Derks, Sune Falck, David C. Hoos,
Thierry Lelegard, Nicolas Lesbats, Jean-François Rameau, Maxim Reznik,
Jean-Pierre Rosen, Jerôme Roussel, Ariane Sinibardy.
- SETI@Home from Ted Dennison. AWS is used as a "plugable" GUI to retrieve
different program status.
- DOCWEBSERVER from Wiljan Derks
In our department we keep our documents in a directory tree. These documents
are all project related and have a certain naming convention to be able to
find the right document. In the past I already wrote a program that searches
though this directory and then converts the found documents into fixed html
pages. With AWS I was able to get a much nicer setup. I have now a server
that can do the following:
- browse through the projects in explorer style. The html contains info
about the document like date and title.
- one can check in documents through the web interface
- it shows our download page as I have send you in the example
- we have now all our documentation in small pieces of html as is needed
to build .chm (w2k compiled help) files. For these we use a content
file, that is also stored in the document archive.
The docwebserver gives by reading all this stuff the direct view on this
documentation. On the other hand I can run some tool and automatically
generate the .chm files.
- OESM Server (OESM=Overall Equipment Status Monitoring) from Wiljan Derks
I am working on a project now for our factories. ITEC mainly delivers
equipment for discrete semiconductor assembly. Almost all of that equipment
is now controlled by a similar Ada 95 based code with having a lot of code in
common. One of the common things, is the way we log errors and state changes
of our equipment.
The OESM Server is an application which copies all this information
continuously to its local pc by opening the proper files on the remote
equipment. That data copied is also stored in local files. The web server
component of the application can then, making use of that data, give reports
that show things like the amount of products produced in a certain period,
error paretos of equipment, mtbf, %time in production and of course many
The cool thing of course is that this information can easily be charted (I
am use kavachart) and it allows simple navigation through different groups
of equipments and different views on the equipment.
- WORM from Pascal Obry
A Web server to share bookmarks, this server was using a standard CGI
design. To keep session information we were using a GLADE partition. With
AWS the design has been really simplified, there is no need for a session
partition, there is no need to build all CGI as partitions too. GLADE is now
used only to handle distributed objects. Indeed WORM is a multi-server
system (using RACW) with a register/unregister mechanism.
Also the server seems to be fastest, there is no more CGI to spawn.
- Internet Currency Trading System at http://www.actforex.com by
This is a server is used to keep historical data about currency trading to
build charts of currency prices. The charts viewer part is written in Java
and loaded through AWS. This server can be reach on the Internet.
Ongoing work is done to based this development on AWS framework only and
to remove all the Java layers. It is also interesting to note that this is
an heavy loaded server, it handle something like 40 to 50 requests per
seconds on a Windows 2000 Server.
- http://www.ada-ru.org (Ada in Russian)
This Web Site is powered by AWS.
Thanks to all who have reported bugs and have sent us patches.
Dmitriy & Pascal.
--| Pascal Obry Team-Ada Member
--| 45, rue Gabriel Peri - 78114 Magny Les Hameaux FRANCE
--| "The best way to travel is by means of imagination"
--| gpg --keyserver wwwkeys.pgp.net --recv-key C1082595