The IDE include checker got to the point where it is actually useful and
this removes a lot of unneeded includes.
Naturally, especially for headers like util.h, this may mean we need to
re-add includes in consuming cpp files that bloats the diff a bit.
We now removed the need for Boost:chrono in all the libs, to avoid
accidentally linking to it again this change makes the apps link to
the actual specific libs instead of just all.
The idea of binding to interfaces now will take into account the
interfaces maybe becoming available only after the server started.
If your server starts at machine boot, it is a 50/50 chance that the
network interfaces are already fully configured and have received
addresses. In case of dhcp, more often than not this means that your
server will not be listening at the main interface because it wasn't up
yet.
This new api allows the server to give a function to register a new
interface and we have some linux specific code that will notice changes
in the interfaces and we'll allow the app to bind to it a moment or two
after that.
As we moved most of the creation of a BufferPool to be via the
Streaming::pool() method, which uses a thread-local, it makes sense
to start cleaning up the design and make it more modern C++.
The above mentioned method would return a reference and you'd see
loads of places use `auto &pool =` which is less than ideal.
As the number of places where we actually instantiate a BufferPool
goes down, the usage of some sort of smart pointer makes more sense.
This now makes all APIs use BufferPool be wrapped in a shared_ptr.
This follows the coding style guideline that the file that contains a
class should follow the exact name of this (main) class.
pubkey.{h|cpp} -> PublicKey.{h|cpp}
pubkey_utils.{h|cpp} -> PublicKeyUtils.{h|cpp}
A simple port caused a regression in the unit tests that took me a long
time to track down, as a result the code has been commented, prettyfied
and more.
The final issue was that the return time of QList::size() turned from an
int to a ulong and I stored that in a QDataStream with the streaming
operator.
On extracing the now twice as long number I just put it into an int,
causing 4 bytes to not be read and a subsequent off-by-one in reading
further data from the file (a list of ints).
Memmapping can use any filesize, but the pages are allocated in a power
of two amount.
Assuming 4K pages we now aim to have 65536 pages to map a big file.
This gets around the sillyness where you need an extra dir for a config.
~/.config/flowee/rest-service/rest-service.conf
now you can just use
~/.config/flowee/rest-service.conf
When there is no config we now bind on all available (at startup)
interfaces to listen for service requests.
This solves the problem where a fresh server would not listen to
requests from the outside world but only on localhost.
* avoid using warning/fatal log levels as they go to stderr
* correctly exit at end in more cases
* Limit size of transaction we dump on screen unasked.
* remove dead code.
Instead of asking for a block and reacting when we get a failure, this
just uses the current height instead.
This also makes 'flush' happen in each indexers' own thread, which should
be good for performance.
Ask the hub for the chain-length and when we processed the last block,
immediately flush the data to the SQL DB.
This makes use represent the whole world faster.
tomFlowee
John Galt