GitLab

Resolves #26969.

In the implementation of more accurate DNS resolution, the DirContext
used by JNDI lookups was shared in a utility class. It turns out that
InitialDirContext, the implementation we used, is very not thread-safe
and so as a result lookups would often fail to complete.

This commit creates a new DirContext for every lookup. The overhead in
doing so is minimal and ensures that we do not have thread safety
issues.

This commit also increases RDNS_LOOKUP_WORKERS_NUM to 30 (from 5). The
majority of time in these threads is spent waiting for replies and so
there are gains to be had in increasing the number of requests in
flight.

Fixes: #26963

Implements the first part of #26919, which is to find out whether the
"fingerprint" parameter is used at all.

Implements the last part of #23713.

Implements another part of #23713.

Implements part of #23713.

Implements #25448.

Jackson can convert names to snake_case by itself just fine.

Prepares #24296.

The filter is applicable to both bridges and relays. It uses basic
string operations to extract the operating system from the platform line
of descriptors.

Fixes: #6946

A missing descriptor directory has caused an IllegalStateException
when we attempted to write our parse history files. Let's be more
careful about this case and skip this history file if there is no such
directory.

Fixes #26711.

Indexes are created for relays with no known country code or
autonomous system number using the special values "xz" and "AS0"
respectively.

No changes are made to the summary or details documents.

Fixes: #26665

This commit adds two new fields: {un,}verified_host_names.

Whereas previously InetAddress was used to resolve reverse domain
names, this instead changes the lookup mechanism to use JNDI allowing
for a deeper view into the DNS. It also accounts for the fact that
multiple PTR records are not forbidden in the DNS specification and are
often used in shared hosting scenarios.

A host name is considered verified if it has a matching forward record.
If a PTR value is found to have multiple A records, it will be
considered verified if any one of the A records matches the original
address. If no matching record is found, it will be reported as an
unverified host name.

Previously, unverified host names were discarded internally by the
InetAddress lookup mechanism and so this data could not be used.

To maintain "bug compatibility" with the previous implementation of the
"host_name" field, which will now be deprecated, the IP address is
returned when a lookup fails.

The host_name field continues to be used, but now will consider all
verified and unverified host names. If finer grained filtering is
needed, then a seperate ticket could be filed for that, but it is
unclear that it is useful enough to justify the work.

Fixes: #18342

This class uses JNDI to look up resource record values from the Domain
Name System using the system configured resolver.

See: #18342

Update metrics-lib dependency to 2.4.0.

Adapt touched logging statements to our standards.

Implements #25848.

Fixes #25332.

Also adapt tests to check the new functionality.
Implements task-25740.

This makes the git-revision available even when running 'ant clean war'.

 - Fix NullPointerException in cases where a relay/bridge does not
   report to have a (valid) tor version.
 - Set version_status field in bridge details documents, too.

Fixes #25700.

Implements #24256.

Add a utility method for only un-escaping valid utf and supply a test
as well as test data for this issue.

Fixes task-22594.