Lighttpd

You have combined multiple requests. Please do not do that.

There have been some issue surrounding the PID when packaging for Gentoo. I wanted to know if we could make adjustments upstream that might help.

Please post in the forums to ask questions before submitting patches for things that might be end-user mistakes instead of issues in lighttpd. What problem are you trying to solve? You have not sufficiently described the problem.

Why are you not using server.pid-file? The pid will then be ... wait for it ... easily readable from the file.

If not using that, any service management, such as systemd or daemontools, can run lighttpd -D (to keep lighttpd in the foreground) and then you can use the service management tools to send signals to the lighttpd pid.

In your patch for printing config, you only print the value if the value is a string. That is a major limitation and unsuitable to even be considered.
(BTW, you have also misspelled "speccific" (sic) and "teh" (sic))

.

Regarding the call to getrandom(), did you test your patch? Does it work? (Odds are "no" and "no" from my reading of the patch.)

Is there any harm in moving the call to `li_rand_reseed()`

General answer: yes there is possible harm. If you do not know what the code does and why, then it is dangerous to move it. The initial call to li_rand_reseed() that you moved in your patch should not do anything. Additionally, the place that you moved it to should not have any perceivable difference with the original, even if the initial call did have an effect.

The code in lighttpd defers seeding the random number generators.

However, that is probably not your problem. If you are using any TLS module, the TLS module initialization is more likely the place where the server is hanging waiting for the system to have sufficient entropy available to continue. This is independent from li_rand_reseed().