Where do I find the doumentation for the xojo_project file structure?

[James_Dooley]

@Tim_Parnell

I copied directly from the EULA tab from the About window of 2020r2.1

I’m trying to suggest the respectful route of contacting Xojo directly to resolve matters. They are the ones who can decide if OP is in violation of the EULA.

Nothing in your quote prevents the creation of code generators and courts decide on breach of contracts.

[Michael_Dettmer]

Wow. I am very impressed about this forum again

Let me explain the background of my question again. I am developing a generator tool based on Apache Velocity and Xojo. It uses a(ny) database as source for the generators. It can generate any language code - even PHP and Java.

And it can and will generate Xojo code. This is nice and useful - like ARgen does in a similar way. I believe this will be good for Xojo and their customers also.

I am using a structured approach for the generated apps. This means I do not have an interface designer necessarily. But I use the database and an input structure with parameters.

I just asked a question and I got an answer. Though not satisfying I will try to live with that. I can understand Xojo because if they would deliver this information the next step would be getting errors reading the project files. And this could geopardize the integrity of the IDE.

So if there are any comments regarding the question itself they are very welcome.

[Arnaud_N]

@Douglas_Handy

If they had to publish and not break backward compatibility, I suspect their own development time would be at negatively impacted.

They could publish it only as a reference. No need to be responsible of what others do with that.

[Arnaud_N]

@Tim_Hare

There is no guarantee your code won’t break if they change the format (and they have - and my code broke, but I couldn’t complain, because it wasn’t documented).

Why would documenting it (for the sake of just information) directly means one can complain?

[Arnaud_N]

Great, Michael. You’re welcome.

[Thomas_Tempelmann]

My tool Arbed reads and writes project files in all 3 format, and from my experience, the binary format is the most stable and easiest to get right, xml is a bit more of a problem because the IDE’s xml parser appears to self-written and not very forgiving about white space (even though if it’s valid xml). The text format is the worst to get right. Sometimes even an extra empty lines matters.

The binary format is structurally identical to the xml format, without the parsing issued. But you need a translation table between the codes used in the binary vs xml formats.

[Michael_Dettmer]

This is the reason why I asked for some documentation:

The text format is the worst to get right. Sometimes even an extra empty lines matters.

Problem for me is that I am generating the code completely from scratch. There is no project when I start. I am using templates yet - and they contain stuff I do not know anything about. So you say I should try to generate XML projects?

[Thomas_Tempelmann]

The safest approach would be to create a dummy project in the Xojo IDE that contains each of the elements you want to replicate, then copy & modify those elements in the file somehow.

[Tim_Hare]

I use XML format for code generation. I generate modules though, not entire projects. I did write a program that parses binary format projects to do a compare of them and show what changed.