Arni
Arni
It seems that in `unNameFormat` there's strange problem: 1. Creating a `char* unnamed_str` at heap via std::unique_ptr 2. Using it somehow to calc `char* out` 3. Getting raw `char* ptr`...
Just fucking nice: one place fixed, got one test fixed and other tests leak in some other places. I'm loving it (no)
just... why ... why it's in C's char pointer and not it std::string or any like
Just amazing: solved another leak via converting a lot of dangling fmt::string_view to std::string... And got another one... Creating thread buffer in heap and then saving pointer in std::vector .........
> Memory management in fmtlog is a little bit tricky, which may trigger false alarm by sanitizers, so don't care too much about it. IDK , I solved them somehow,...
> will use (and identify the version of) version 1.8.2 of benchmark on my machine if it's a local package, when CPM_USE_LOCAL_PACKAGES is ON. If CPM_USE_LOCAL_PACKAGES is OFF, it will...
I can confirm it still happens on version `2.3.1` at Gentoo Linux. Also it is kinda strange: it doesn't return failure ( like `int main(){return 1;}` ).
I've used forgejo for my private project. And I've set up to use mydomain.com/git/ for the forgejo instance. I solved everything until I found that actions/checkout in github is broken....