From fielding@kiwi.ICS.UCI.EDU Mon Jun 9 13:30:50 1997 Received: from ics.uci.edu by paris.ics.uci.edu id aa03253; 9 Jun 97 13:30 PDT Received: from binky.ics.uci.edu by q2.ics.uci.edu id aa13822; 9 Jun 97 13:30 PDT Received: from ics.uci.edu by binky.ics.uci.edu id aa20377; 9 Jun 97 13:30 PDT Received: from paris.ics.uci.edu by binky.ics.uci.edu id aa20373; 9 Jun 97 13:30 PDT Received: from paris.ics.uci.edu by paris.ics.uci.edu id aa02929; 9 Jun 97 13:26 PDT Received: from kiwi.ics.uci.edu by paris.ics.uci.edu id aa02906; 9 Jun 97 13:26 PDT To: libwww-ada95@ics.uci.edu Subject: About the libwww-ada95@ics.uci.edu mailing list Date: Mon, 09 Jun 1997 13:25:59 -0700 From: "Roy T. Fielding" Message-ID: <9706091326.aa02906@paris.ics.uci.edu> Reply-To: libwww-ada95@ics.uci.edu X-Mailing-List: archive/1997/1 X-Loop: libwww-ada95@ics.uci.edu Precedence: list Errors-To: libwww-ada95-request@ics.uci.edu Sender: libwww-ada95-request@ics.uci.edu Introduction ============ This mailing list has been established for technical discussion about libwww-ada95, including problem reports, interim fixes, suggestions for features, and contributions. The mailing list address is libwww-ada95@ics.uci.edu and administrivia [including (un)subscribe requests] should be sent to libwww-ada95-request@ics.uci.edu There is also a Hypermail Archive of the mailing list at What is libwww-ada95? ===================== libwww-ada95 is a library of Ada95 packages which provides a simple and consistent programming interface (client library) to the World Wide Web. libwww-ada95 is NOT a CGI library, though some parts of it could be used within CGI applications. More information about the library can be found at: libwww-ada95 is still in development and has not been publically released. The purpose of this mailing list ================================ This mailing list provides a forum for all developers to communicate their ideas, problems, suggestions, and contributions with each other in a timely fashion. It also provides a mechanism for recording those discussions so that others can join in development at a later time and learn from the past. What is appropriate discussion for this mailing list? ===================================================== This list is NOT for discussion of general Ada95 or WWW or CGI issues. It is ONLY for discussion about the libwww-ada95 library and the things that are needed to implement the library. The first priority is problem reports! If anyone finds a bug (or what they think is a bug), report it to the list so that we can all (or at least all those affected by the bug) examine and discuss it. Also, once a solution (or many solutions) is found, send that to the list as well so that all the others can stop looking. The second priority is new contributions (i.e. stuff that you think should be added to the "official" release). All contributed source code must be either Public Domain or distributable under the libwww-ada95 LICENSE. Copyright is retained by the original author (as is always true). Where appropriate, attempts are made to identify the original author both in the code and the release documentation. Contributions of documentation are also strongly encouraged -- this would be a great project for a Technical Writing class. ;-) Note that not all contributions will make it into the "official" release. Contributed-but-not-released software is available on the Hypermail Archive (see above). Also, feel free to send useful information which you DO NOT want to be included in the "official" release -- just be sure to mark them as such. If you are working on a project of general usefulness, let everyone else know about it. Sharing and exchanging ideas is the best way to build a quality product, and the time lost talking about it is almost always regained through the insight of others and the additional insight you yourself gain by describing an idea such that others can understand it. Also, if you have a need for some service which is not provided by the library, let us all know -- someone may be looking for a new project and is just lacking the initial impetus. Ah, and finally, never forget to keep your sense of humor. If people say something which sounds a bit odd, you can probably attribute it to the strange customs of our native country and not any intention to offend. I'm half-Yank (USA) and half-Kiwi (NZ) myself, so I'm allowed to say twice as many odd things. Don't forget that others on this mailing list may be from anywhere on the Internet. ....Roy T. Fielding Department of ICS, University of California, Irvine USA (fielding@ics.uci.edu) (libwww-ada95-request@ics.uci.edu) From fielding@kiwi.ICS.UCI.EDU Tue Jun 10 16:43:07 1997 Received: from ics.uci.edu by paris.ics.uci.edu id aa00428; 10 Jun 97 16:43 PDT Received: from binky.ics.uci.edu by q2.ics.uci.edu id ai01024; 10 Jun 97 16:43 PDT Received: from ics.uci.edu by binky.ics.uci.edu id aa13845; 10 Jun 97 10:22 PDT Received: from paris.ics.uci.edu by binky.ics.uci.edu id aa13840; 10 Jun 97 10:22 PDT Received: from paris.ics.uci.edu by paris.ics.uci.edu id aa13977; 10 Jun 97 10:20 PDT Received: from kiwi.ics.uci.edu by paris.ics.uci.edu id aa13947; 10 Jun 97 10:20 PDT To: libwww-ada95@kiwi.ICS.UCI.EDU Subject: summary of netnews for Ada95 stream/socket binding Date: Tue, 10 Jun 1997 10:20:45 -0700 From: "Roy T. Fielding" Message-ID: <9706101020.aa13947@paris.ics.uci.edu> Reply-To: libwww-ada95@ics.uci.edu X-Mailing-List: archive/1997/2 X-Loop: libwww-ada95@ics.uci.edu Precedence: list Errors-To: libwww-ada95-request@ics.uci.edu Sender: libwww-ada95-request@ics.uci.edu I asked about network bindings on comp.lang.ada. It turns out that another person asked a similar question the day before I did, so here is a summary of the responses to both. [I've added some editorial comments in brackets]. ================================================================= From: Ken Raeburn Subject: q: pkg for network/UNIX programming; system dependencies Date: 29 May 1997 23:16:44 -0400 Organization: Cygnus Support, Cambridge MA Message-ID: I'm thinking about working on a couple small spare-time UNIX projects in Ada95 to help familiarize myself with the language. Most such projects I'm considering right now make use of networking, both TCP and UDP. I'm looking for a package to help me deal with UNIX system-defined types, socket- and network-related types and functionality, domain name service, etc. I've looked at the "paradise" package, which seems to have some of what I want, but the latest version I could find is about 3 years old, only works on SunOS and HP-UX and with three specific compilers (none of which is GNAT), and is specifically conditionalized in various places for those systems and compilers. If that's the closest thing available, I'll go with it; I just wanted to know if there's anything else. [...] [This is interesting, since Cygnus does Gnu support and thus would be a good candidate to take over support for libwww-ada95...Roy] ================================================================= From: Corey Minyard Newsgroups: comp.lang.ada Subject: Re: q: pkg for network/UNIX programming; system dependencies Date: Fri, 30 May 1997 14:33:00 GMT Organization: Wonderforce Research Message-ID: I have a port of the forest-1.4 library (it compiles without error under GNAT 3.09) with some limited Berkeley socket bindings at http://www.concentric.net/~Minyard. I understand that the standard forest library is not as extensive as paradise, but I have never used paradise. The socket stuff I wrote is pretty limited, and I'm not an Ada95 guru, so if anyone wants to take it and work on it, feel free. [...] [I looked at it and it is indeed limited. I don't like how it is tied to forest (a huge and outdated POSIX library)....Roy] ================================================================= From: "W. Wesley Groleau (Wes)" Subject: Re: q: pkg for network/UNIX programming; system dependencies Message-ID: <338F3177.1A04@pseserv6.fw.hac.com> Organization: Hughes Defense Communications Date: Fri, 30 May 1997 19:58:47 GMT > I have a port of the forest-1.4 library (it compiles without error > under GNAT 3.09) with some limited Berkeley socket bindings at > http://www.concentric.net/~Minyard. ...... FLORIST (Florida State upgrade to FOREST) might also help. [I grabbed FLORIST from but it is a huge POSIX binding, with no socket support....Roy] ================================================================= From: jsa@alexandria (Jon S Anthony) Subject: Re: q: pkg for network/UNIX programming; system dependencies Date: 30 May 1997 22:59:34 GMT Organization: PSI Public Usenet Link Message-ID: Why not use GLADE or ADEPT? Both of these work with GNAT and give you full distributed object style capabilities (per the DSA of Ada95). ADEPT also gives you RMI interaction. ADEPT: http://www.cs.tamu.edu/research/ADEPT/ GLADE: ftp.gnat.com:pub/gnat [The Distributed Systems Annex (DSA) is for running parallelized programs across multiple systems. Both of these DSA libraries have a binding to an RPC mechanism, but it would take some work to extract just a simple binding to sockets from them....Roy] ================================================================= From: Rolf Ebert Subject: Re: looking for network streams binding Date: Sun, 1 Jun 1997 15:52:46 +0200 Message-ID: RTF> I am looking for a freeware library or set of packages for binding RTF> Ada95 to TCP/IP network streams (or even plain sockets). I have RTF> checked the usual archives, to no avail. The purpose is to provide RTF> a native interface for Ada95 clients using HTTP requests, with RTF> message retrieval, parsing, and processing using concurrent RTF> threads. I'd like to see something like the above myself. Same experience, I could not find anything. There are minimal bindings to sockets in forest-1.4 (kind of POSIX binding), but nothing more. RTF> It needs to be something I can redistribute (eventually) for free. RTF> OTOH, I'd be willing to help coordinate a collaborative effort, RTF> similar to what I have already done for libwww-perl and Apache. I already thought about that, too. I really like the way Perl packages are organized and maintained. Something similar should be established for Ada as well. If you are going to write an Ada version of libwww-perl, please keep me informed. In my limited spare time, I prefer to maintain the Emacs Ada mode. Perhaps you can change my mind. Rolf ================================================================= From: karl@grebyn.com Subject: Re: looking for network streams binding Date: Mon, 2 Jun 97 18:13:29 EDT Message-Id: <9706022213.AA27657@numbers.grebyn.com> Organization: Grebyn Corporation There's Paradise and a couple others. You might even look into the ACE C++ stuff from wustl as a possibility (either as paradigm or binding). I also thought there was some Ada CGI stuff around someplace (part of WebAda) - that might be a worthwhile starting place. Let me know if you want any of that stuff and can't find it elsewhere. [Paradise is ancient. ACE was done by Doug Schmidt (former UCI dude), but it is much too complicated for our purposes. The Ada CGI stuff does exist, but it has no relation to a WWW client library (CGI communication is through stdin/stdout)....Roy] ================================================================= In summary, there is no suitable binding available -- we'll have to create our own. However, it looks like there are a few people that might be willing to help build and test the library. Once we get the initial design done and some notion of the scope, I'll see if they'd like to join the mailing list. In the mean time, it looks like W3C libwww now has a clean objective-C'ish binding to streams, so we may be able to use that as a guide. Docs are available at . ....Roy From fielding@kiwi.ICS.UCI.EDU Fri Jun 20 20:31:57 1997 Received: from ics.uci.edu by paris.ics.uci.edu id aa11728; 20 Jun 97 20:31 PDT Received: from binky.ics.uci.edu by q2.ics.uci.edu id aa11384; 20 Jun 97 20:31 PDT Received: from ics.uci.edu by binky.ics.uci.edu id aa27818; 20 Jun 97 20:31 PDT Received: from paris.ics.uci.edu by binky.ics.uci.edu id aa27814; 20 Jun 97 20:31 PDT Received: from paris.ics.uci.edu by paris.ics.uci.edu id aa11712; 20 Jun 97 20:31 PDT Received: from kiwi.ics.uci.edu by paris.ics.uci.edu id aa11692; 20 Jun 97 20:31 PDT To: libwww-ada95@ics.uci.edu Subject: Re: GNAT and streams interfaces In-reply-to: Your message of "Fri, 20 Jun 1997 13:37:39 PDT." <6746.866839059@elysees.ics.uci.edu> Date: Fri, 20 Jun 1997 20:28:29 -0700 From: "Roy T. Fielding" Message-ID: <9706202031.aa11692@paris.ics.uci.edu> Reply-To: libwww-ada95@ics.uci.edu X-Mailing-List: archive/1997/3 X-Loop: libwww-ada95@ics.uci.edu Precedence: list Errors-To: libwww-ada95-request@ics.uci.edu Sender: libwww-ada95-request@ics.uci.edu Unfortunately, this seems to be another case where the same word (streams) is used in too many different ways. I've been looking through this stuff and it appears their are three different uses: Ada streams: (a-stream.ads) A mechanism for reading and writing Ada objects and data structures to/from persistent storage. This is essentially providing methods for linearizing a data structure into a byte stream such that the typing information is not lost. Subclassing Ada.streams allows you to write/read the object to/from a C stream. C streams: (i-cstrea.ads) Basically, what most people call standard IO. Operations for buffered reads/writes to Unix file descriptors. We could use these for writing to a socket, but only after the socket is bound to a file descriptor. Unix streams: A set of kernel mechanisms that support the development of network services and data communication drivers. It defines interface standards for character input/output within the kernel and between the kernel and user level processes. The STREAMS mechanism is composed of utility routines, kernel facilities and a set of data structures. A stream is a full-duplex data path within the kernel between a user process and driver routines. The primary components are a stream head, a driver and zero or more modules between the stream head and driver. A stream is analogous to a shell pipeline except that data flow and pro- cessing are bidirectional. In a stream, the stream head is the end of the stream that provides the interface between the stream and a user pro- cess. The principle functions of the stream head are pro- cessing STREAMS-related system calls, and passing data and information between a user process and the stream. I am still looking for better docs on Unix streams (the above is from `man -s 2 intro`). See also the man pages for getmsg() and putmsg(). What we want is a library interface for handling Unix streams. I have downloaded the Ada DSA for GNAT called GLADE which seems to include an RPC interface to TCP/IP, but from reading the documentation it would seem to be designed as Unix streams (or at least that is the impression I get from their description of "channels", which is almost identical to the W3C library's use of channels). I'll have to look further into the code before I can tell what it actually is using. If we are extremely lucky (unlikely), we might be able to use a subset of the Garlic library rather than invent our own. Anyway, all this stuff (currently GNAT 3.09) is installed under the "extra" directory in our group account. .....Roy From fielding@kiwi.ICS.UCI.EDU Mon Jun 23 11:28:12 1997 Received: from ics.uci.edu by paris.ics.uci.edu id aa27869; 23 Jun 97 11:28 PDT Received: from binky.ics.uci.edu by q2.ics.uci.edu id aa09481; 23 Jun 97 11:28 PDT Received: from ics.uci.edu by binky.ics.uci.edu id aa10009; 23 Jun 97 11:28 PDT Received: from paris.ics.uci.edu by binky.ics.uci.edu id aa10005; 23 Jun 97 11:27 PDT Received: from paris.ics.uci.edu by paris.ics.uci.edu id aa27693; 23 Jun 97 11:25 PDT Received: from kiwi.ics.uci.edu by paris.ics.uci.edu id aa27673; 23 Jun 97 11:24 PDT To: libwww-ada95@kiwi.ICS.UCI.EDU Subject: TCP/socket binding in Garlic Date: Mon, 23 Jun 1997 11:21:09 -0700 From: "Roy T. Fielding" Message-ID: <9706231124.aa27673@paris.ics.uci.edu> Reply-To: libwww-ada95@ics.uci.edu X-Mailing-List: archive/1997/4 X-Loop: libwww-ada95@ics.uci.edu Precedence: list Errors-To: libwww-ada95-request@ics.uci.edu Sender: libwww-ada95-request@ics.uci.edu GLADE's protocol library (Garlic) does indeed have a complete thin socket binding (s-garthi.ads) and thicker binding for the functions that socket applications need for client and server operation (s-gartcp.adb). It will probably take some work to put together a subset which does not include the distributed services annex, but at least we have a working example. The Garlic source is in /extra/fielding0/src/glade/1.02/Garlic .....Roy From fielding@kiwi.ICS.UCI.EDU Tue Jun 24 12:51:32 1997 Received: from ics.uci.edu by paris.ics.uci.edu id aa17343; 24 Jun 97 12:51 PDT Received: from binky.ics.uci.edu by q2.ics.uci.edu id aa14818; 24 Jun 97 12:51 PDT Received: from ics.uci.edu by binky.ics.uci.edu id aa14393; 24 Jun 97 12:51 PDT Received: from paris.ics.uci.edu by binky.ics.uci.edu id aa14389; 24 Jun 97 12:51 PDT Received: from paris.ics.uci.edu by paris.ics.uci.edu id aa17120; 24 Jun 97 12:48 PDT Received: from kiwi.ics.uci.edu by paris.ics.uci.edu id aa17098; 24 Jun 97 12:48 PDT To: libwww-ada95@kiwi.ICS.UCI.EDU Subject: HTTP Tutorial Date: Tue, 24 Jun 1997 12:43:59 -0700 From: "Roy T. Fielding" Message-ID: <9706241248.aa17098@paris.ics.uci.edu> Reply-To: libwww-ada95@ics.uci.edu X-Mailing-List: archive/1997/5 X-Loop: libwww-ada95@ics.uci.edu Precedence: list Errors-To: libwww-ada95-request@ics.uci.edu Sender: libwww-ada95-request@ics.uci.edu This is a gentler intro to HTTP: http://www.jmarshall.com/easy/http/ ....Roy