Hi Michael,

This might be a long shot but whenever i get an odd error 11's (like they should not be happening) then it seems like my file gets corrupted.
    for example writing without a dom and getting an error 11
    or getting they key(file) then trying to read with same key and getting an error 11.
if i do a keyed load on file it usually fixes it.

Jerry

James, I did that. I am taking an error branch (seterr in effect) due to an error 11.  As I say, it is a write with no DOM or ERR on an mkeyed file.

If you ESCAPE before the WRITE, you can step through it. The embedded IO routing will be steppable and you can find the offending situation that way.

In an embedded IO routine (post_write) I am getting an error 11 on a write to a file with no DOM specified.  If  I write it in console mode and I have no issue. This is an mkeyed file like this:  keyed "test",[1:1:2]+[2:1:10]+[3:1:6],0,-300.  Is some strange IO option in play during embedded IO?  I'm not seeing any such setting in the io program itself.

Hi

Thank you. that worked.
it must have been that 'XI' parameter which I dont think i can shut off across board but maybe right prior to sql extracts.

Jerry

As far as I can tell, that error is either coming from *winqry or *winlist.  You should be able to debug *winqry (*winlist is owner code protected) to see what's happening on line 12100.

PxPlus will do that for you, but you can change the SQL syntax it uses by changing option EXTROPT (see https://manual.pvxplus.com/?command_tags/mysql.htm).  I just tried this:

Session #1:
1>print pth(1)," % ",opt(1)
[MySQL]MySQL;bitfile;KEY=KEYFIELD % ;user=root;pswd=pswd;prepare=N;strip
1>extract (1,key="05")
1>print prm('XI')
 0

Session #2:
1>print pth(1)," # ",opt(1)
[MySQL]MySQL;bitfile;KEY=KEYFIELD # ;user=root;pswd=pswd;prepare=N;strip
1>extract(1,key="05")
Error #0: Record/file busy
Last IO to [MySQL]MySQL;bitfile;KEY=KEYFIELD, channel 1
MYSQL:1205:Lock wait timeout exceeded; try restarting transaction

One thin client user doing a query is suddenly getting:

Query Error 41: Invalid integer encountered (range error or non-integer)
(12100)
 
on any query object in our query.en file.  Not only happens in any of our programs using question mark query, but also same error directly in Nomads using test function on any query object. Just this one user, which just started happening.  Everyone else is ok.  Thin client user is remote with win 11.  Everyone using Providex v5.0

Re-installed our code on the thin client did not resolve.  Any ideas?

Quick update from our side:

AI-driven PxPlus development with Git and automated deployment

We're working on a PxPlus-based ERP/logistics system and have set up a workflow where AI (Claude) doesn't just write the PxPlus source code, but also generates database definitions and NOMADS panels.

Everything runs through GitHub with automated deployment.

What the AI writes:

• PxPlus source code — Programs, API endpoints, calculations. The AI has full knowledge of the PxPlus language and writes code that compiles directly.
  
• Data Dictionary create scripts — Instead of manually creating table definitions, the AI generates create scripts for the data dictionary, describing the full database structure.
  
• NOMADS panels — The AI also generates screens. We export NOMADS panels to ".pxpnl" text files (the native PxPlus TO_CVS format). This format is human-readable and diffable.
  

Git workflow:

• Branches: feature branches → dev → test → main (live)
  
• Source only in Git: Compiled ".pxp" files are in ".gitignore"; only "src/" directories are committed.
  
• NOMADS watcher: A PowerShell FileSystemWatcher monitors ".en" library files. On every change, panels are automatically exported to ".pxpnl" text files, so the diff is visible in Git.
  

Automated deployment:
A deploy script ("deploy.sh") handles the full chain on the server:

• 1. git pull - fetch source code
• 2. Compilation - compile all source files with "pxplus -cpl"
  
• 3. Panel import - import ".pxpnl" text files back into the NOMADS ".en" libraries
  
• 4. Database migration - run new or changed create scripts to update the data dictionary
  
• 5. Environment routing - deploy across three environments with their own paths and branches (dev, test, live)
  

What this gives us:

• Version control on everything: Code, screens, and database structure, all as readable text in Git.
  
• Reproducible environments: Setting up a new environment is just a "git clone" + "deploy.sh".
  
• AI as a development tool: The AI has full knowledge of PxPlus syntax and conventions, the NOMADS format, and the data dictionary structure and generates working code including screens and database definitions.
  

The hung sessions all have this USER_AGENT in common:  Mozilla/5.0 (Windows NT 6.1; WOW64) SkypeUriPreview Preview/0.5 skype-url-preview@microsoft.com

After digging, I see that this might be a web crawler used by Skype to generate link previews.
Or, a real user is referencing a web page within MS Teams or Skype and a thumbnail preview is being invoked. 

Either way, the web crawler or preview is interfering with the HTTP_CONNECTION 'close' request.

I'm tending toward a user and their Skype setting being the issue, rather than a web crawler. 

Cheers.