Enforce limit on all formats not just html and make limit = 100

[rufuspollock]

• Why: no risk that someone accidentally gets 500k records and crashes their browser.
• Can allow for getting all (if really wanted) by having _limit=-1 or similar (though not sure this is necessary)
• Limit of 100 is reasonable number (could even have 20 IMO)

[pudo]

• I disagree. Just being able to use /x/y/z.csv as a resource and download it without having to write a complete API client seems like a big advantage. Whats the use case for calling 20 rows of a CSV file?

[rufuspollock]

I don't quite understand why would i need an API client. I'd do /x/y/z.csv?_limit=-1. I agree that isn't quite as pretty as it should be but it would be a price worth paying to avoid people accidentally getting 1GB file. If that is too unpleasant for you perhaps we could just configure this for data types so that e.g. html and json are restricted but csv is not. For me what really matters is json and html because that is what people will grab expecting a sample.

[rufuspollock]

bump! I really think this is important (albeit minor).