It's now displayed semi-sorted, because it's initially sorted but then we sum up all versions usages per library. Depending on how many versions and how much the library is used, some will bump up or down a spot here and there.

It should be re-sorted after being summed and filtered. (after this point https://github.com/compiler-explorer/conanproxy/blob/main/html/usage.html#L104)

Example of what is now shown:

https://my.papertrailapp.com/groups/1243314/events?q=program%3Ace-conan&focus=1647886078848028673&selected=1647886078848028673

PayloadTooLargeError: request entity too large
 at readStream (/home/ubuntu/ceconan/conanproxy/node_modules/raw-body/index.js:156:17)
 at getRawBody (/home/ubuntu/ceconan/conanproxy/node_modules/raw-body/index.js:109:12)
 at read (/home/ubuntu/ceconan/conanproxy/node_modules/body-parser/lib/read.js:79:3)
 at jsonParser (/home/ubuntu/ceconan/conanproxy/node_modules/body-parser/lib/types/json.js:135:5)
 at Layer.handle [as handle_request] (/home/ubuntu/ceconan/conanproxy/node_modules/express/lib/router/layer.js:95:5)
 at trim_prefix (/home/ubuntu/ceconan/conanproxy/node_modules/express/lib/router/index.js:328:13)
 at /home/ubuntu/ceconan/conanproxy/node_modules/express/lib/router/index.js:286:9
 at Function.process_params (/home/ubuntu/ceconan/conanproxy/node_modules/express/lib/router/index.js:346:12)
 at next (/home/ubuntu/ceconan/conanproxy/node_modules/express/lib/router/index.js:280:10)
 at cors (/home/ubuntu/ceconan/conanproxy/node_modules/cors/lib/index.js:188:7)

This appears to be happening during the builds for QT 6.4.2

Includes some Clang compilers and the new Gcc Riscv compilers
https://conan.compiler-explorer.com/libraries.html

Perhaps with a # link with some actions, or in worst case perhaps a separate page

The page https://conan.compiler-explorer.com/usage.html retrieves statistics from a CSV and displays them in a D3 chart.
But when there's a lot of libraries listed, it's not very user friendly or readable.

Preferably at least the libraries that have 0 hits should not be in the chart, and even better should be mentioned like "The following libraries are left out because they were not used in the specified time-period: lib1, lib2, lib3, etc"

If there are better ways to graph this out with D3 thats Ok too.

The page can be edited directly from https://github.com/compiler-explorer/conanproxy/blob/main/html/usage.html and should be testable from any browser without having to serve it from a webserver.