selsta: yes, probably tomorrow though

Who should I prod to get the TXT records updated for update.getmonero.org ? Pony or was that handed around ?

i'm not sure, but binaryfate might have access ^

only fluffy has access at the moment

the daemon seems to return multiple www-authenticate headers which causes requests from firefox to fail

the daemon is started with: ./monerod --stagenet --rpc-login superuser:abctesting123 --rpc-access-control-origins localhost:9100

ideally firefox would handle the issue gracefully like the other browsers, but I'm looking to remove the duplicate from the response

@woodser The daemon do the right thing. It's firefox fetch() api that's broken.

I thought the firefox fetch() api was breaking because it can't parse the daemon's www-authenticate response header, which is either duplicated or separated by newlines instead of commas?

I can see the wallet-rpc return 2 lines:

known firefox issue: bugzilla.mozilla.org/1491010

WWW-authenticate: Digest qop="auth",algorithm=MD5,realm="monero-rpc",nonce="k7VW9A39o7Dp7OjSnVCA/w==",stale=false

WWW-authenticate: Digest qop="auth",algorithm=MD5-sess,realm="monero-rpc",nonce="k7VW9A39o7Dp7OjSnVCA/w==",stale=false

I see the same from daemon responses

The RFC nnnn says it's correct. But fetch() discard them as duplicates or something. I'll try to find the issue on the tracker...

while it's technically correct, and that issue should be resolved in firefox, we might be able to support firefox in the interim by cleaning up the response headers

How about not using fetch at all? xmlhttprequest do not suffer from same bug.

worth trying

Or github.com/request/request-promise can handle digest auth.

When using xmlhttprequest one has to handle the construction of the auth response. Not that it's that hard..

selsta see PR 6407

thanks, trying it currently

What's the difference between std::string(host) and std::string{host} ? I tried the former and it did not help, and I thought it was the same.

it is, but maybe applied to the wrong line? the compiler was giving funky line numbers for some reason

I think the compiler was saying the line on get_host was the issue, but it was set_host

I did it for both the set and get IIRC. Even though it only moaned about one.

also C++ compiler error messages are useless

:D

Then I guess I messed up something else -_-

no it gives the correct line number

nah, they are extremely helpful, you just have to be a compiler writer to understand them

lol so I guess not very helpful most of the time

right, I guess you need a lot of experience to to appreciate them

the thing is they are often times giving the equivalent of stack traces for templates

do you know why this compiled on Windows?

so its quite noisy and you have to look at the bottom of the stack first

the tests arent built

or is release-static not building tests?

ok

its a cmake command to request it

Oh. I tried again, it does work, but the error comes up for a second similar case, I did not notice the liner number was further out.

luigi1111: can you merge #6407 to fix compilation? ty

Reminder meeting tomorrow 17 UTC monero-project/meta #449

(take into account hour change tonight)

done