Goodby Xojo.Core/Text/Auto

A long time ago when Xojo.Core/Text/Auto was the new kid on the block I disregarded it as I didn’t want to change all my code base at once. Over time I slowly added more and more of the “new” framework into my projects, eventually my whole code base was saturated.

I have been dreading the switch back to Variant/String/Dictionary, I thought it would be a slow migration back over time but that’s not how it ended up going.

At first I changed a few internal libraries to use regular dictionaries to be compatible with newer projects. Before I knew it had spread to uses of text (string and text keys don’t match). In the end, for one large project it took me an entire week of tracing down each use of the previous framework and converting it to the new thing. I ended up changing over 150 files and 10k lines of code. Each line had to be inspected and tested to make sure I wasn’t introducing a bug.

It was a long and hard process but overall I think it was worth it. I can now get back to productive work without trying to get the frameworks to play nice with each other. I was upset at first about all the name changes and modifications to the framework but I think it was the right direction for Xojo.


The main thing saving me from this pain is that I only started a handful of smaller projects during the reign of Text and Auto… So it’s actually reduced my refactoring load to have just waited a bit :wink:


My resistance to change paid off for me in this case - I never used any of that Xojo.Core stuff.