ssl

Checklist

• Dependencies installed
• No typos
• Searched existing issues and docs

Issue Description

When using the RateLimiter Middleware with a rate between 0 and 1 all events will be rejected instead of applying the specified rate. E.g.: e.Use(middleware.RateLimiter(middleware.NewRateLimiterMemoryStore(0.5)))

I am not saying that it is a common use case to have

As noted in #14464, the randomisation used does not need to be cryptographically secure. Any form of randomisation is sufficient.

The project could implement a low quality but fast RNG for the purposes generating this kind of blinding material. There is an existing low quality RNG in impl_cache_flush_cache() in crypto/property/property.c which could be used more widely. There is another

Context and Description

The READMEs and any example code in all projects should be updated to reflect the move from the IBM-Swift organization to the Kitura organization.

If anyone wants to take on all or part of this, please comment here so other's know what you're working on and submit PR's. :-)

Thanks!

There's little information about what keys and values are in the output, what it means and how they are related to the screen output. In general that needs to be added. (special topics see #1675, #1674)

Problem:

A common pattern is:

GUARD(s2n_stuffer_skip_write(stuffer, bytes_to_write));
uint8_t* ptr = suffer->blob.data + stuffer->write_cursor - bytes_to_write;

which could be simplified.

Solution:

*ptr could be an *out parameter to s2n_stuffer_skip_write

• Does this change what S2N sends over the wire? No.
• Does this change any public APIs? No.

Running the oauth2-proxy with the "--display-htpasswd-form" and "--htpasswd-file" set to true, using a set of incorrect credentials doesn't show anything to the end user.
The end-user needs to know that his/her attempt failed (due to incorrect username/password) otherwise they don't even know that the attempt was done. For them

There are advantages to grouping commonly-used fields in structures together. On Cortex-M0, an access to the first 128 elements of a structure (p->x when offsetof(t, x) / sizeof(x) < 128 where sizeof(x) is 1, 2 or 4) uses less code than an access beyond this boundary. On platforms with a cache, putting commonly-used fields in the same cache line optimizes cache use.

Anecdotal evidence sug