- Use a simple struct instead of two slices to store the data and the checksums
of resources
- Remove a superfluous call to Sprintf
- Factorise presence check and data retrieval in some maps
- Size the maps when possible
There is no need to perform a heavy-weight SQL query gathering all the
information available on a feed when we're only interested in its last check
timestamp.
- Replace a call to fmt.Sprintf with a concatenation
- Explicit declaration of return values in FetchJobs
- Initialize the size of FetchJobs return value to b.limit: when b.limit is
used, which is most of the time, this avoid resizing the slice, and when it
isn't, the size of the map is set to 0, which is equivalent to the previous
situation anyway.
- Move a call to `request.UserID(r)` to a lower scope.
- Move the seeking inside of DetectFeedFormat instead of having it everywhere
in ParseFeed
- Provide a hint for the compiler to eliminate a useless bound check in
DetectJSONFormat, otherwise it'll check that buffer[i] is valid on every
iteration of the loop. This shouldn't make a big difference, but oh well.
Apparently, postgresql isn't smart enough to realize that once a true value
value is found as part of a `SELECT true`, there is no need to scan the rest of
the table, so we have to make this explicit. We could also have used the
`SELECT EXISTS(…)` construct, but it's more verbose and I don't like it.
- The JS bundle has its own isolated scope
- There is no need to use IIFEs anymore (Immediately Invoked Function Expressions)
- Modules are executed after the HTML document is fully parsed, similar to `defer` attribute
- There is no need to use `DOMContentLoaded` anymore
- Module scripts inherently run in strict mode (no need to define `use strict` anymore)
Since Go doesn't support unions, and because the translation format is a bit
wacky with the same field having multiple types, we must resort to
introspection to switch between single-item translation (for singular), and
multi-items ones (for plurals).
Previously, introspection was done at runtime, which is not only slow, but will
also only catch typing errors while trying to use the translations. The current
approach is to use a struct with a different field per possible type, and
implement a custom unmarshaller to dispatch the translations to the right one.
This should marginally reduce the memory consumption since interface-boxing
doesn't happen anymore, speed up the translations matching, and enforce proper
typing earlier. This also allows us to remove a bunch of now-useless tests.
As stated in the documentation:
> Read calls io.ReadFull on Reader and crashes the program irrecoverably if an
error is returned. The default Reader uses operating system APIs that are
documented to never return an error on all but legacy Linux systems.
Maintaining a separate ChangeLog file is redundant and error-prone,
as it largely duplicated the Git commit history without adding meaningful context.
Release notes are still available on GitHub Releases and the Miniflux website.
