You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Since we're really only storing one thing, it's not really required that we put it in a database.
We could for example move to storing the zip files on disk a bit like this:
archives
└── <tla>
└── <sha25sum>.zip
We'd still need to serve the files ourselves in order to enforce our auth requirements, but this might be simpler to work with.
A downsides here is that adding metadata gets tricky. For example it's now a bit more effort to work out which was the most recent upload - we either rely on filesystem timestamps (fine, but may not have the nicest API) or munge it into the file names too.
The text was updated successfully, but these errors were encountered:
Even if we don't change to using filesystem storage, it may still be useful to move the pks of the database rows to being a content id rather than an incrementing number.
Since we're really only storing one thing, it's not really required that we put it in a database.
We could for example move to storing the zip files on disk a bit like this:
We'd still need to serve the files ourselves in order to enforce our auth requirements, but this might be simpler to work with.
A downsides here is that adding metadata gets tricky. For example it's now a bit more effort to work out which was the most recent upload - we either rely on filesystem timestamps (fine, but may not have the nicest API) or munge it into the file names too.
The text was updated successfully, but these errors were encountered: