I made an app to show the books you have lend at a public library
And to make sure it can be used with all existing libraries web sites and all opacs, you can enter any url and an xpath expressions, and then it runs the xpath expressions each day and shows the result as list of books.
And since I wrote it 15 years ago, there were no tech libraries for headless browsers available. So I wrote half of my own browser, and an XPath interpreter. (Modern XPath is actually Turing-complete, and with the EXPath file module, it can read and write to any file)
So, if the store does not allow custom browsers or Python interpreters, those are two reasons it would not allow my app.
Rendering engine is not the issue. It's the JIT - firefox could publish a browser, it would just need to have interpreted javascript and it'd be useless.
Well, then it is good, I did not make a JIT (only XPath AST, not even bytecode)
Although I have been thinking about making a JIT. Building on for x86 is easy enough, but then it is useless for ARM. I would not want to build two JITs. Or four with 32/64-bits.
This is just wrong. Apple disallows apps that duplicate existing functionality, i.e. browsers. You can write whatever wack headless browser you want, and there are tons of frameworks for this exact purpose, too: All manner of methods to do network things without invoking Safari.
This is a jawbreaker of hyperbole around a chewy gum center of truth.
And to make sure it can be used with all existing libraries web sites and all opacs, you can enter any url and an xpath expressions, and then it runs the xpath expressions each day and shows the result as list of books.
And since I wrote it 15 years ago, there were no tech libraries for headless browsers available. So I wrote half of my own browser, and an XPath interpreter. (Modern XPath is actually Turing-complete, and with the EXPath file module, it can read and write to any file)
So, if the store does not allow custom browsers or Python interpreters, those are two reasons it would not allow my app.