Re: Historic Curiosity Question: Websites in QSYS.LIB

Hi Patrik

Just a clarification or extension - DLO is Document Library Object - we see it today as QDLS - a file system whose naming rules are PC DOS - all capitals, maximum length of name components is 8.3. This was used to map drive in PC Support. There are commands like WRKFLR and WRKDOC (I think) that we can still use. Your user name has to be enrolled in the system distribution directory in order to use QDLS - still do, I think. It probably is actually in files in a library, hence some of the ineffeciencies of using QDLS - IBM recommend strongly never to use QDLS these days, for reasons of poor performance, limited naming convention, and the extra authority requirement. E.g., I just tried WRKFLR on a system where I have *SECOFR user class, I could not, because I'm not enrolled.

BTW, have you heard of the CMD menus? Type GO CMDxxx where xxx is a command action like DSP or a command object like PGM - or DLO - so GO CMDDLO gives you a menu with all the commands for DLO on your system. GO CMD* can give you a list of all of these menus. A wonderful thing, the naming convention of commands on this system.

Cheers
Vern

On 8/17/2023 2:35 PM, Patrik Schindler wrote:

Hello,

I have been told in a private mail exchange that a long time ago — before the IFS has come to the AS/400 — "websites" have been hosted within what we now recognize as QSYS.LIB file system. I saw on my P02 that the installable LPPs list in V2R3 features TCP/IP, while at the same time there is no IFS. Just the DLO "file system", which is — well known — a crude emulation of a hierarchical file system with objects being placed in a normal library (QDOC?). Back then, it could only be accessed through DLO commands.

I found evidence that hosting web content in QSYS.LIB is indeed possible from the commented template configuration of the V4 IBM http server: https://try-as400.pocnet.net/wiki/IBM_HTTP-Server_for_AS/400_Configuration

Scroll down to the "Pass - Accept matching requests" section. There are examples for how to code different file system paths:

# QDLS: Pass /doc/* /QDLS/WebTEST/*.HTM
# Root: Pass /root/* /WebSAMP/*.html
# QOpenSys: Pass /x/* /QOpenSys/WebSAMP/*.html
# QSYS: Pass /lib/* /QSYS.LIB/WebSAMP.LIB/HTMLDOC.FILE/*.MBR

While I see no problem with loading content from individual PF members in principle, what makes me really wonder is: How to cope with stream files, aka graphics? Does anyone have some memory about that?


Part two.

I know that newer releases of the (i)ntegrated database support BLOB fields defined though SQL. Those are a way to store and retrieve graphics. But I fail to understand how to access graphics saved in this way from within a web page, with an <img src> tag. Maybe this isn't possible at all? This is good example where IBM documentation largely fails for me. There is documentation about DB2 and BLOB fields. There is documentation praising IBM i's web serving abilities and how to configure those — in the end it's more or less Apache, no big deal. But some "in between" documentation gaps stay. Thus part two of my question to get some hints about where to probably take a closer look and read. :-)

Thanks!

:wq! PoC