This pull request has been shunk down to size with other changes now to be in separate pull requests.
The change request here simply splits moves some of "net" debugging into "net2", "block" and "tx".
This is a cut-down version of #8728 which also seeks to make debug message conform to a style.
Sorry for all the pull requests as a result of breaking this up, but I felt there was too much to it in it's previous format, and therefore was unlikely any of it would be merged.
Concept ACK
Wouldn't need init.cpp (debugCategories) an update as well?
Wouldn't need init.cpp (debugCategories) an update as well?
Yes, needs to be documented in option help.
ok, will add a commit to update option help... shall I list each debug string e.g. "tx, tx2, tx3"? or is it better to show the level of granularity another way, e.g. "tx{,2,3}" or "tx[1:3]", "tx[1-3]"...? not sure if there's a standard for denoting this sort of thing.
412 | @@ -413,7 +413,7 @@ std::string HelpMessage(HelpMessageMode mode) 413 | strUsage += HelpMessageOpt("-limitdescendantsize=<n>", strprintf("Do not accept transactions if any ancestor would have more than <n> kilobytes of in-mempool descendants (default: %u).", DEFAULT_DESCENDANT_SIZE_LIMIT)); 414 | strUsage += HelpMessageOpt("-bip9params=deployment:start:end", "Use given start/end times for specified BIP9 deployment (regtest-only)"); 415 | } 416 | - string debugCategories = "addrman, alert, bench, coindb, db, http, libevent, lock, mempool, mempoolrej, net, proxy, prune, rand, reindex, rpc, selectcoins, tor, zmq"; // Don't translate these and qt below 417 | + string debugCategories = "addrman, alert, bench, block, coindb, db, estimatefee{,2}, http, libevent, lock, mempool(,2}, mempoolrej, net{,2}, proxy, prune, rand, reindex, rpc, selectcoins, tx{,2}, tor, zmq"; // Don't translate these and qt below
Can you also add cmpctblock?
μNit: I was wondering if it could help to add a "rule of thumb" somewhere to explain when to use category vs. category2.
@MarcoFalke Done. I propose that for future use, "cmpctblock" be restricted to the actual debug's relating to creation/etc of compact blocks, but for the transmission and download that block{,2} etc be used (since "block", "tx" are subsets of "net"). Sound ok?
Regarding rule of thumb, happy to do this (perhaps in a future pull) - best put into init.cpp or into a readme file somewhere?
On a slightly OT note - can someone tell me how one gets notifications when someone comments on a pull request? My currently practice is to scan through my pull requests to see if comments have appeared, but I suspect there may be a better way!!
still testing. will re-open when more done.
Can you please stop opening and closing this PR?
@rebroad Just leave this open. You can do as much testing as you want locally, but be sure to compile and run the tests before you push.
4784 | @@ -4775,12 +4785,15 @@ void static ProcessGetData(CNode* pfrom, const Consensus::Params& consensusParam 4785 | CBlock block; 4786 | if (!ReadBlockFromDisk(block, (*mi).second, consensusParams)) 4787 | assert(!"cannot load block from disk"); 4788 | - if (inv.type == MSG_BLOCK) 4789 | + if (inv.type == MSG_BLOCK) { 4790 | + LogPrint(fRecent ? "block" : "block2", "sending regular %s (%d) to peer=%d\n", inv.ToString(), nHeight, pfrom->id);
Any reason not to add the size of the block too? Useful to analyse bandwidth usage by block height. Although that is logged in beginmessage/endmessage already so a nasty bit of awking should be able to join them together.
do you mean log the size when sending it or when receiving it, or both?
Click on Details... https://travis-ci.org/bitcoin/bitcoin/jobs/155868520#L1230