html_url,issue_url,id,node_id,user,user_label,created_at,updated_at,author_association,body,reactions,issue,issue_label,performed_via_github_app
https://github.com/simonw/datasette/issues/272#issuecomment-503351966,https://api.github.com/repos/simonw/datasette/issues/272,503351966,MDEyOklzc3VlQ29tbWVudDUwMzM1MTk2Ng==,9599,simonw,2019-06-18T23:45:17Z,2019-06-18T23:45:17Z,OWNER,Uvicorn 0.8.1 is our and supports `raw_path`!,"{""total_count"": 0, ""+1"": 0, ""-1"": 0, ""laugh"": 0, ""hooray"": 0, ""confused"": 0, ""heart"": 0, ""rocket"": 0, ""eyes"": 0}",324188953,Port Datasette to ASGI,
https://github.com/simonw/datasette/issues/513#issuecomment-503249999,https://api.github.com/repos/simonw/datasette/issues/513,503249999,MDEyOklzc3VlQ29tbWVudDUwMzI0OTk5OQ==,7936571,chrismp,2019-06-18T18:11:36Z,2019-06-18T18:11:36Z,NONE,"Ah, so basically put the SQLite databases on Linode, for example, and run `datasette serve` on there? I'm comfortable with that. ","{""total_count"": 0, ""+1"": 0, ""-1"": 0, ""laugh"": 0, ""hooray"": 0, ""confused"": 0, ""heart"": 0, ""rocket"": 0, ""eyes"": 0}",457201907,Is it possible to publish to Heroku despite slug size being  too large?,
https://github.com/simonw/datasette/issues/502#issuecomment-503237884,https://api.github.com/repos/simonw/datasette/issues/502,503237884,MDEyOklzc3VlQ29tbWVudDUwMzIzNzg4NA==,7936571,chrismp,2019-06-18T17:39:18Z,2019-06-18T17:46:08Z,NONE,It appears that I cannot reopen this issue but the proposed solution did not solve it. The link is not there. I have full text search enabled for a bunch of tables in my database and even clicking the link to reveal hidden tables did not show the download DB link.,"{""total_count"": 0, ""+1"": 0, ""-1"": 0, ""laugh"": 0, ""hooray"": 0, ""confused"": 0, ""heart"": 0, ""rocket"": 0, ""eyes"": 0}",453131917,Exporting sqlite database(s)?,
https://github.com/simonw/datasette/issues/512#issuecomment-503236800,https://api.github.com/repos/simonw/datasette/issues/512,503236800,MDEyOklzc3VlQ29tbWVudDUwMzIzNjgwMA==,7936571,chrismp,2019-06-18T17:36:37Z,2019-06-18T17:36:37Z,NONE,Oh I didn't know the `description` field could be used for a database's metadata. ,"{""total_count"": 0, ""+1"": 0, ""-1"": 0, ""laugh"": 0, ""hooray"": 0, ""confused"": 0, ""heart"": 0, ""rocket"": 0, ""eyes"": 0}",457147936,"""about"" parameter in metadata does not appear when alone",
https://github.com/simonw/datasette/issues/512#issuecomment-503200024,https://api.github.com/repos/simonw/datasette/issues/512,503200024,MDEyOklzc3VlQ29tbWVudDUwMzIwMDAyNA==,9599,simonw,2019-06-18T15:58:18Z,2019-06-18T15:58:18Z,OWNER,"The `about` and `license` and `source` keys are all intended to be used for links - so if you provide a `about_url` it will be displayed as a URL, and you can then use the `about` key to customize the link label.

There are `description` and `title` fields which can be used to display text without linking to anything.

I'm definitely open to reconsidering how these work - I don't think they quite serve people's needs as they are right now, so suggestions for improving them would be very welcome.","{""total_count"": 0, ""+1"": 0, ""-1"": 0, ""laugh"": 0, ""hooray"": 0, ""confused"": 0, ""heart"": 0, ""rocket"": 0, ""eyes"": 0}",457147936,"""about"" parameter in metadata does not appear when alone",
https://github.com/simonw/datasette/issues/513#issuecomment-503199253,https://api.github.com/repos/simonw/datasette/issues/513,503199253,MDEyOklzc3VlQ29tbWVudDUwMzE5OTI1Mw==,9599,simonw,2019-06-18T15:56:29Z,2019-06-18T15:56:29Z,OWNER,"Unfortunately not - I really wish this was possible. I have not yet found a great serverless solution for publishing 1GB+ databases - they're too big for Heroku, Cloud Run OR Zeit Now. Once databases get that big the only option I've found is to run a VPS (or an EC2 instance) with a mounted hard drive volume and execute `datasette serve` on that instance, with an nginx running on port 80 that proxies traffic back to Datasette.

I'd love to figure out a way to make hosting larger databases as easy as it currently is to host small ones.","{""total_count"": 0, ""+1"": 0, ""-1"": 0, ""laugh"": 0, ""hooray"": 0, ""confused"": 0, ""heart"": 0, ""rocket"": 0, ""eyes"": 0}",457201907,Is it possible to publish to Heroku despite slug size being  too large?,
https://github.com/simonw/datasette/issues/272#issuecomment-503195217,https://api.github.com/repos/simonw/datasette/issues/272,503195217,MDEyOklzc3VlQ29tbWVudDUwMzE5NTIxNw==,9599,simonw,2019-06-18T15:46:31Z,2019-06-18T15:54:18Z,OWNER,"How should file serving work?

Starlette and Sanic both use `aiofiles` - https://github.com/Tinche/aiofiles  - which is a small wrapper around file operations which runs them all in an executor thread. It doesn't have any C dependencies so it looks like a good option. [Quart uses it too](https://gitlab.com/pgjones/quart/blob/317562ea660edb7159efc20fa57b95223d408ea0/quart/wrappers/response.py#L122-169).

`aiohttp` does things differently: it has [an implementation based on sendfile](https://github.com/aio-libs/aiohttp/blob/7a324fd46ff7dc9bb0bb1bc5afb326e04cf7cef0/aiohttp/web_fileresponse.py#L46-L122) with [an alternative fallback](https://github.com/aio-libs/aiohttp/blob/7a324fd46ff7dc9bb0bb1bc5afb326e04cf7cef0/aiohttp/web_fileresponse.py#L175-L200) which reads chunks from a file object and yields them one chunk at a time, 
","{""total_count"": 0, ""+1"": 0, ""-1"": 0, ""laugh"": 0, ""hooray"": 0, ""confused"": 0, ""heart"": 0, ""rocket"": 0, ""eyes"": 0}",324188953,Port Datasette to ASGI,