V10 Test Problems

Downloaded V10 for testing purposes. I thought by now that the worst of the bugs would be sorted… silly me.

Even something as simple as creating a blank project for testing purposes just fails. The Create button does nothing If I click cancel the annoying ‘Loading’ rotating gif appears and never goes away.

Hardly a good advert for the product (See fix below).

Installation
The first big time waster was the install. Everything looked fine, but when you ran the IDE, it just hung up. 2 hours later (after reinstalls, manually updating the version, yada, yada, I eventually fixed it by clearing all browser casche. Something in the install (on a different machine than the v9) corrupts the cache. I eventually ran it in incognito mode and it worked. When I cleared cache in the browser it worked fine.

Using the AI Capability
Go figure… you need to enter a licence key to access the AI setup, which of course you cant do with a trial version. SO the main capability of V10 cannot be tested. I sent a message to Netmake to ask how I do it… they let me have a temporary key.

Once in the AI module you neeed to setup a link to your preferred AI… Which makes sense if you think aout it.
Scriptcase does not host the LLM (Large Language Model) or pay for your computation. By forcing you to use your own external API key (from OpenAI, Gemini, or Claude), they pass the variable infrastructure and token costs entirely to you. This prevents Scriptcase from having to charge heavy, recurring monthly AI subscription fees on top of their software license.

I don’t necessarily think this is a problem as I’m not convinced you will need heavy AI token usage in normal dev work, but only time will tell.
I signed up for a free key using gemini, but just do a coppilot search or go onto your google account and it will give you options. Not sure if I did it right, but I had to setup a £20 credit on the google account and I set a £5per month spending cap to allow me to test.

Loading a Project
I exported and loaded a v9 - (350+ app) project fairly straightforwardly and it seemed to work. I haven’t fully tested it yet as I want to work through it in a structured way. Here are my first comments.

Dashboards
Kept coming up with a header warning error on a (previously working) Dashboard. (My home screen so easy to spot!) The error indicated that some output was being sent to the header after it was locked. I disabled all header stuff in settings and it still persisted. The only way I found to fix it was to delete all widgets and add them back. (What an absolute waste of effort).

Thats It for now, I’ll add any other snippets as I delve deeper into the software.

Dashboard error reported here:

With some indifferent support from Scriptcase AI to find it. It was only marginally better than doing it separately using copilot. It made the similar instructions to do things you cant actually do in scriptcase and wanted me to debug a standard dashboard created in v10 that exhibited the same problem!

I’ve temporarily suppressed the warnings causing the problem using the follwing ‘bandaid’ code: in the NEW onApplicationInit event on a dashboard.
error_reporting(0);
ini_set(‘display_errors’, 0);

More Builds

Generating Data Dictionary
Dont know if it worked or not. Got an unhelpful error message saying ‘An error occurred… It may hve not been saved’. Since I was doing the whole database, it just left me feeling it doesn’t work properly. Something as important as this cant be left to chance. I can only suggest you do it a piece at a time.

New Gantt Chart Capability
Overall pretty nice. There are a few niggles, like you cant set a time period to start with plus other time options. If you tick all the timeline period options, the grid uses all of them. If you don’t tick them then the grid is OK, but you cant select them. Also, you cant set the height of the bars so you get some UGLY results depending on the height of the window and the number of tasks in the Gantt. It would be much more sensible to allow a min/max setting for the bar heights. One really REALLY iritating problem though is that for some reason, when the task is wide enough to need a scrollbar, the bar is at the bottom and you need to scroll right down to get to it. Even woese is that for some reason, the task line doesnt always show the full left hand characters even if you haven’t scrolled it, so the text just looks missing or wrong. The scrollbar should be more accessible (top?) and more importantly the field should always start showing the left hand characters.

The final nail in this functionality is that you can’t print the Gantt. Ctrl P is your only option and that is just plain prehistoric. (Read unprofessional)

Performance in the IDE
Generally it seemed faster and fewer (if any) glitches in the refreshes. As part of the migration I also took the opportunity to only bring forward recent versions and so clean the scriptcase db, so that might have helped.