monero-wallet-cli regularly saves the wallet to disk, but monero-wallet-rpc never, unless "store" is explicitly called or when it's sent kill signal and left enough time to do so?

I remember this biting me in the past, still the case right?

No Sir. I had rpc running for the last 24hrs. I just closed the wallet using Ctrl-c, opened the same wallet file using cli and am fully synced

ok thanks, somehow I must not be graceful enough

The kill signal will not save (can't be hooked) but the int signal can will save, or ^C will too.

On Linux (and probably mac) anyway. Who knows what windows does with signals.

thanks will check my signals

Sorry, I meant term, not int. But int should also save as it's ^C.

Right, see src/common/util.h, look for SIGINT

you can't really send signals on windows. but the console driver will recognize break and ctrl-c

I wrote signal support for MSYS1 many years ago twitter.com/hyc_symas/status/1154448029791002624

is there an effort to reduce the size of the wallet cache? thought I remember seeing that talked about somewhere

woodser[m]: that was me, for the seraphis wallet work

Maybe could be adapted to wallet2

The idea is to only store block id checkpoints instead of all block ids.

ok. makes sense

is there any way to merge all data from two wallet cache blobs from the same wallet?

if I'm not mistaken, key images and outputs could be merged using existing export/import functions, but this would not transfer destination addresses and other local data

or is it possible to import the secret tx keys from another wallet cache?

i don't recall seeing any update to this over the years: monero.stackexchange.com/questions/…llets-that-control-the-same-address

hm, yeah

tx keys can be transferred using set_tx_key

I am trying to identify all RingCT outputs and rings using the get_transactions RPC call. If tx version == 2, are all outputs in the tx RingCT outputs? If input amount == 0, are all rings RingCT type? And are all coinbase outputs since the genesis block eligible to be spent in a RingCT input?

I think so, yes, no, resp.

For the second question, re-reading it, it doesn't mean what I thought you meant, and probably what you meant to say.

The answer is yes if s/are all rings/is that ring/

I don't think you can mix rct and non rct rings, but I am not 100% sure.

In fact, I'm pretty sure you can't. Just not 100% sure...

In any case, an out is rct iff amount == 0 or coinbase out after the rct fork.

(the latter is deemed rct really, though isn't, but it's treated as one)

Yes, sorry I embedded the question of "must all inputs in a given tx be the same type of ring?" in there.

I can just check individually for each input.

To be absolutely sure about question (1), instead of just using tx version == 2, can I just check that the "amount" of each output is zero?

Amount being zero implies rct.

Thanks!

hey, is the matrix a mirror of the irc? I wanted to ask if there is any history being kept but if the former is true no answer would be needed

And sold to google ;)