refilling often is good, but refilling often has unclear side effects
on a) cpu load, and b) making sure every cell, ever, is sent out one at
a time

fixes bug 4124, as noticed in bug 4115

removes another avenue for enumerating bridges.

Otherwise, we can wind up munging our reference counts if we set it in
the middle of loading the nodes.  This happens because
nodelist_set_consensus() and microdesc_reload_cache() are both in the
business of adjusting microdescriptors' references.

The fix is to turn held_by_node into a reference count.

Fixes bug 4118; bugfix on 0.2.3.1-alpha.

we don't need to check whether we don't have enough guards right after
concluding that we do have enough.

slight efficiency fix suggested by an anonymous fellow on irc.

We were doing "divide bandwidth by 1000, then multiply by msec", but
that would lose accuracy: instead of getting your full bandwidth,
you'd lose up to 999 bytes per sec.  (Not a big deal, but every byte
helps.)

Instead, do the multiply first, then the division.  This can easily
overflow a 32-bit value, so make sure to do it as a 64-bit operation.

Florian Tschorsch authored 13 years ago and Nick Mathewson committed 13 years ago

Implements bug3630.

Mansour Moufid authored 13 years ago and Nick Mathewson committed 13 years ago

GCC 4.2 and maybe other compilers optimize away unsigned integer
overflow checks of the form (foo + bar < foo), for all bar.

Fix one such check in `src/common/OpenBSD_malloc_Linux.c'.

This may help diagnose bug 4027.