<rbrunner7> @tobtoht:monero.social: I have a new problem with the Polyseed code that must have crept in quite recently. Symptom is a problem with the Polyseed dependencies. By setting a breakpoint in the Dependency constructor, I can confirm that it runs and initializes static dependency deps; in polyseed.cpp. But if a bit later a P [... too long, see mrelay.p2pool.observer/e/5IHvmaELRWcydzZK ]

<rbrunner7> One suspicion that I have but that I could not yet verify so far is that after the recent Polyseed PR (see tevador/polyseed #20/changes) we somehow have that static deps variable twice, and only one of the two gets initialized correctly.

<rbrunner7> (twice in memory, in 2 different parts of the resulting binary)

<rbrunner7> If that does not look like the probable culprint, any idea where and how to continue the search for the root of the problem? This goes quite a bit over my low-level C/C++ knowledge :)

bas-kirill opened issue #11126: Suggest `hw_reconnect` when hardware wallet HID communication fails du...

> monero-project/monero #11126

<tobtoht> @rbrunner7:monero.social: Is the issue still present in your rebase? I can look at it now.

<rbrunner7> Yes, I rebased now to test again, just to make sure that I don't chase any gosts. The issue persists.

.merge+ 11081 11080 11079 11078 11077 11075 11073 11062

Added

<sneedlewoods_xmr:matrix.org> I tested it with polyseed_static changed back to polyseed in src/mnemonics/polyseed/CMakeLists.txt and it works. > <@rbrunner7> One suspicion that I have but that I could not yet verify so far is that after the recent Polyseed PR (see tevador/polyseed #20/changes) we somehow have that static deps variable twice, and only one of the two gets initialized correctly.

<sneedlewoods_xmr:matrix.org> With ldd monero-wallet-cli I see that libpolyseed.so.2 is missing when using polyseed_static in target_link_libraries(). The binary size is 23MB for non-static and 24MB for static.

<tobtoht> Reverting to dynamic linking is not a solution. It must be linked statically for release builds.

<syntheticbird> dynamic linking in general is a failure

<sneedlewoods_xmr:matrix.org> AFAICT polyseed_static works for release builds but not for debug builds, which is what confuses me

<rbrunner7> Now, that's interesting. Can confirm: A release build works, a debug build doesn't.

<rbrunner7> Well, could we take out the hammer and use polyseed_static for release builds, and polyseed for debug builds? Surely that CMake thingy is smart enough for something like that? :)

.merges

10865 10872 10886 10887 10964 11046 11066 11077 11078

<syntheticbird> .merged

<syntheticbird> damn i can't even joke without a typo

Ap4sh opened pull request #11127: epee: parse HTTP methods case-sensitively

> monero-project/monero #11127

<syntheticbird> LETS FUCKING GO

<syntheticbird> sorry i overreacted

<basses:matrix.org> only syn issues get fixed in less than a day

<syntheticbird> my issues are small tbf