issues

I can make this change only after releasing 0.28 - if I make the change earlier than that publish heroku etc will break because they will install the latest release of Datasette which will not understand the -i option.

This is a one-line fix: replace this:

https://github.com/simonw/datasette/blob/2ad9d15cd6901654e6801e2faa29e6fc08bae5fa/datasette/utils.py#L489

With this: (need to do it for other publishers too though)

quoted_files = " ".join( ["-i {}".format(shlex.quote(file_name)) for file_name in file_names] )

The README is quite out of date now. It includes out-dated copies of help files, promotes the old Zeit Now integration and duplicates a lot of material from the docs.

So much new stuff!

https://github.com/simonw/datasette/compare/0.27...master

If the file changes during download it will probably result in a corrupt download. Safer not to allow downloads at all of mutable databases.

Added in https://github.com/simonw/datasette/commit/bf6b0f918de4aeee7c1036ac975ce2fb23237da7 without docs.

It doesn't work any more, and it's not a particularly useful feature - I've hardly used it since I added it.

See #419: if Datasette is going to work against mutable SQLite files it can no longer assume that the .inspect() method will have cached the correct schema for all tables in all attached databases.

So everywhere in the code at the moment that relies on .inspect() data needs to be modified to use live introspection of the schema instead.

From a comment later on: here are the uses I need to fix as a checklist:

• [x] table_exists()
• [x] info["file"] in .execute()
• [x] resolve_db_name()
• [x] .database_url(database)
• [x] DatabaseDownload file path
• [x] sortable_columns_for_table() uses it to find the columns in a table
• [x] expandable_columns() uses it to find foreign keys
• [x] expand_foreign_keys() uses it to find foreign keys
• [x] display_columns_and_rows() uses it to find primary keys and foreign keys... but also has access to a cursor.description which it uses to list the columns
• [x] TableView.data uses it to lookup columns and primary keys and the table_rows_count (used if the thing isn't a view) and probably a few more things, this method is huge!
• [x] RowView.data uses it for primary keys
• [x] foreign_key_tables() uses it for foreign keys
• [x] DatabaseView list of tables
• [x] IndexView

Split from #419. Mainly useful to see what is connected as mutable v.s. immutable.

Also helps fill the gap left by /-/inspect until #465

This will help clarify that you need a nowv1 account use it.

Follow on from #418 - the Datasette documentation needs an entire section (probably a new page) describing exactly how the hash-in-URL caching mechanism works.