moneromooo that seems to have done it (both)

moneromooo: gitlab security release

Updated, thanks.

i am looking for a js monero address validation, something better then a regex and for all current formats [ account / integrated / sub - addresses ]

NyMonero probably has that.

MyMonero?

Software which does monero things in javascript. Which seems relevant to your question.

spoke0: you can look at the source of xmr.llcoins.net/addresstests.html too

<woodser> two issues to report with emscripten

<woodser> (1) adding the wallet2.cpp source file in CMakeLists gives a compile error ‘openssl/evp.h’ file not found: paste.debian.net/1119462. I’m not building OpenSSL but it shouldn’t be necessary when a custom implementation of the abstract http client is used

<woodser> (2) HAVE_MLOCK and CRITICAL_REGION_LOCAL in contrib/epee/src/mlocker.cpp cause the built binary to hang. the issue is resolved by these changes but maybe there is a better way: woodser/monero dc5388e

It's fair to define out the mlock code there, it's likely all going to splat everywhere in memory anyway.

I'm also fine with the http client parameter.

Is evp.h actually needed ?

<woodser> not for webassembly

<woodser> how can I get around the #include <openssl/evp.h> import not being found when the http client parameter is used?

Comment it out, see what breaks, if anything.

<woodser> that leads to this dependency which is again not needed if using a custom http client: paste.debian.net/1119464

OK. I think I'd be OK with a -DENABLE_SSL=OFF switch in cmake, if it defaults to ON.

Anyone thinks it's a dumb idea ?

Can openssl not be built with enscripten, or did you choose not to do it ?

<woodser> I had difficulties with it. endogenic claims success with trial so it’s possible

<woodser> I did not not see that it was necessary if using a custom http client. Although ideally monero’s http client would be webassembly compatible but I’m not sure support is there yet

Then it's better to not add this -DENABLE_SSL=OFF. emscripten build files can be added to contrib/emscripten.

endogenic: are you willing to contribute this (enscripten build/config for openssl to work with monero) ?

<woodser> by “It’s fair to define out the mlock code there”, what exactly are you recommending? the workaround comments a lot out

I'll make a patch and post it so you can confirm it works for you.

<woodser> ok

yeah i mean our clientside work is open source and i've been hoping it'll converge with core -- have at it.. pleased it's a help

pls take a look at our monero-core-custom repo

we make core embeddable

ndorf fyi

it's 100% not the cleanest

Great. If there are contrib build scripts, it's less likely it'll slowly break again.

Or at least I'll get notified when it does :)

most def

spoke0: check github.com/mymonero/mymonero-core-js

it's up to date

current tradeoff (without work to allow configuring these things) is you have to embed the whole module rather than just addr validation

for a time there was a pure js impl of addr validation in there - it's in history - supports subaddrs too

but if you use that, though it's probably ok bc it's mostly a trivial function, it will indeed diverge from core

does any issue at the moment require a CLI point release?

if not we will tag GUI v0.15.0.2

Yes. We might not have a patch yet though.

That db corruption thing that a few people got.

Like, er, you.

It can be later though if you want to release.

I have a feeling it may take a while to confirm a fix...

yes I don’t think we will fix that soon

I’ve been running hyc’s patch the last days but no crash yet

crash / corruption

binaryFate: look at PM pls

we’re having issues with our public node repeatedly dying

2019-12-02 07:05:22.562 [RPC1] ERROR net contrib/epee/include/net/abstract_tcp_server2.inl:759 Setting timer on a shut down object

Dying with an error, or a crash ?

s/error/purposeful exit/

not a crash, it just stops responding

Are you implying it's still running ?

It does still run, but none of our RPC requests get responses.

gdb into it, and "thread apply all bt"

gdb /path/to/binary `pidof monerod`

OK - waiting for Mykola to do it, will get back to you

what to do after those commands moneromooo?

Paste the output to fpaste.org or similar.

There'll be several pages, so make sure you go through to the end of the pager.

OK

seems we are missing some libraries but can’t figure out what

There's not even the first screenful of thread apply all bt here.

You can ignore the "no such file or directory". Kernel source missing, I don't need it.

OK

No apparent reason for a hang. run "set_log 1,net*:TRACE,*perf*:INFO" in monerod, then issue a get_info call. Then paste the resulting monerod logs.

Be quick to do this in case the node is talking to the p2p network, this is going to be very noisy.

The perf* logs are only written to file. Can you paste that version please ?