Using a standard ending here will let other tools that expect
markdown understand our output here.
This commit was automatically generated with:
for fn in $(find src -name '*.dox'); do \
git mv "$fn" "${fn%.dox}.md"; \
done
This includes app, core, feature, lib, and tools, but excludes
ext, test, and trunnel.
This was generated by the following shell script:
cd src
for dname in $(find lib core feature app tools -type d |grep -v \\.deps$); do
keyword="$(echo "$dname" |sed -e "s/\//_/" )"
target="${dname}/${keyword}.dox"
echo "$target"
cat <<EOF >"$target"
/**
@dir ${dname}
@brief ${dname}
**/
EOF
git add "$target"
done
The code here parses the fields from the microdescriptor, including
possible annotations, and stores them into a microdesc_t object.
This commit is almost pure code movement; I recommend using
--color-moved to review it.
This code is logically independent of the rest of the function, and
goes better in its own function.
This is almost purely code movement; I suggest reviewing with
--color-moved.
We have code in microdescs_parse_from_string() to record the digests
of microdescriptors that we could not parse. But right now, that
code looks at the md->digest field, which is a bit inelegant, and
will stand in the way of sensible refactoring.
Instead, use a local variable to hold the digest.
Previously we had decoded the asn.1 to get a public key, and then
discarded the asn.1 so that we had to re-encode the key to store it
in the onion_pkey field of a microdesc_t or routerinfo_t.
Now we can just do a tor_memdup() instead, which should be loads
faster.
Previously, we would decode the PEM wrapper for keys twice: once in
get_next_token, and once later in PEM decode. Now we just do all of
the wrapper and base64 stuff in get_next_token, and store the
base64-decoded part in the token object for keys and non-keys alike.
This change should speed up parsing slightly by letting us skip a
bunch of stuff in crypto_pk_read_*from_string(), including the tag
detection parts of pem_decode(), and an extra key allocation and
deallocation pair.
Retaining the base64-decoded part in the token object will allow us
to speed up our microdesc parsing, since it is the asn1 portion that
we actually want to retain.
The strcmp_len() function was somewhat misconceived, since we're
only using it to test whether a length+extent string is equal to a
NUL-terminated string or not. By simplifying it and making it
inlined, we should be able to make it a little faster.
(It *does* show up in profiles.)
Closes ticket 28856.
I believe we originally added this for "just in case" safety, but it
isn't actually needed -- we never copy uninitialized stack here.
What's more, this one memset is showing up on our startup profiles,
so we ought to remove it.
Closes ticket 28852.
