Eric Hanson
Eric Hanson
I agree. Ok.
Done.
This would not support schema changes. If two versions of a row have different schemas, boom.
Would love to see http/2 support in uWSGI!
Getting close. Running the [extension-to-bundle.sql](https://github.com/aquametalabs/aquameta/blob/f54572b089130da03f20595f2847ddd78b65078a/experimental/extension-to-bundle.sql) script calls [extension_to_bundle()](https://github.com/aquametalabs/aquameta/blob/f54572b089130da03f20595f2847ddd78b65078a/experimental/extension-to-bundle.sql#L51) on each extension except meta and bundle, and outputs below. After the conversion to bundles, I deleted all the bundled schemas and...
Timing on bundle views, with everything installed: ``` dev@aquameta:~/dev/aquameta/experimental$ cat view-timing.sql | psql -e select count(*) from _bundle_blob; count ------- 5620 (1 row) Time: 243.264 ms select count(*) from head_commit_field;...
Or maybe use permissions? Once we get the requirement of being superuser out of `checkout()`, this could at least be possible.
This dovetails into the whole question of "What even is a trackable row." Knowns: - `untracked_rows` is an awesome feature. There's something very magic though about knowing that *everything* in...
There may be an analog here to the infamous [Oops, You Wrote a Database](https://dx.tips/oops-database) w.r.t. grant, policies, roles, and row-level security. All of this can probably be declared and enforced...
Ok with me. Sent from my iPhone > On Jan 10, 2014, at 9:55 PM, Martin Chase [email protected] wrote: > > I fixed it on pypi. PR in to fix...