The config file is written poorly: first off, not all available options
are written (they have to be written manually, since there's no struct
enumeration); then, default values are not commented out.
Pool state machine is incomplete, and enabled_pools is not increased if
pool state is not explicitly set to "enabled" in config. However,
there is no reason why the user should not be allowed to disable
all pools.
enabled_pools is a bookkeeping variable, and bookkeeping is cumbersome.
Rather than trying to fix this regression, get rid of another place
where enabled_pools is used.
Closes https://github.com/veox/sgminer/issues/75
Previous defaults (before 10/30) were optimised for Bitcoin that has
a much higher average block time. Optimise for Litecoin (currently
largest scrypt network.)
Not pushing it further down so as not to scare people with dropping
hashrates.
Thanks to MissedOutOnBTC @ freenode for testing.
Closes https://github.com/veox/sgminer/issues/22
Closes https://github.com/veox/sgminer/issues/51
Using j_a_i allows us to get rid of accounting vars like total_poolnames,
and pool names don't get mixed up if they're not specified for some pools.
It would still be desirable to show the URL as a pool name if poolname is not
set.
Use `#defined(...)` instead of `#defined ...`.
Remove comments about issues - anyone removing those lines can check blame
to see why they were added.
Add tonobitc to AUTHORS.md.
Introduced by commit 87e5f35745
The cast/ref/deref is performed wrong and breaks share/best share display.
It is not the value of target that should be increased, but its address.
If all else fails, add an `#ifdef _MSVC_VER`; we should avoid it if possible.
In either case, constructs like `&(*(((uint64_t *)target) + 24))` must be
explained (in a comment).
moved enable_pool(), disable_pool() and other functions related to
state. These could probably be factored out altogether.
Pool state default is now "enabled" - it was previosly "disabled",
but there was an unconditional function call to enable all pools
in main() previously. It was factored out by joe's earlier commits,
so not visible in this one.