I usually compile the client myself since I often have some small patches which I like to use.
I've always been shy to submit any poll request for such, mostly because I'm not really sure how welcome they to be, and nobody likes being rejected
Let ma give you some example.
From time to time I get requests from people I know who have problems with their wallets and so they want me to fix it.
Usually it's because they used to work with several wallets at the same PC, and so they ended up with transactions that have only a question sign, and zero confirmations...
Normally I just tell them to use the "-rescan" option, but sometimes it doesn't help, because the transaction they have in the wallet was somehow a double-spend, and then the only way to fix their wallet is to remove all the transactions and then do the rescan.
So I have this simple patch in init.c, which goes like this:
bool AppInit2()
{
// ...
if (GetBoolArg("-rescan"))
{
pindexRescan = pindexGenesisBlock;
// And here goes my code:
if (GetBoolArg("-purgetransactions"))
{
for (map<uint256, CWalletTx>::iterator it = pwalletMain->mapWallet.begin(); it != pwalletMain->mapWallet.end(); ++it)
{
CWalletTx* wtx = &((*it).second);
pwalletMain->EraseFromWallet(wtx->GetHash());
}
}
}
// ...
}
As you can figure, it basically removes all the local txs from the wallet, if you use both "-rescan" and "-purgetransactions" at the same time.
So: would it be useful enough to bother with a pull request?
And if it would - do I need to add some test cases for it, or what?
Another example, though a bit more complex, would be a possibility to use
* as the
<fromaccount> for
sendmany command.
It's just weird that you cannot specify "any account" there, and I have an application that really sucks because of it.
If such a simple patches are be welcome, I'm more than happy to create some pull requests for them, but if not - no problem, I will just keep patching them in myself.
So please let me know what you think.
