Nikita Popov

I think in principle the feature makes sense. This is blocked on #305 though, which will allow us to know the size of the cache entry without allocating SHM memory...

@mikehaertl top output is notoriously hard to interpret. RES is not a good measure of per-process memory usage, because is also includes (resident) shared memory. Using RES-SHR (with a fixed...

I've implemented two changes to address this: * https://github.com/krakjoe/apcu/commit/ab88120bcd1d6b920b3e79b396ac0224f57a5d6e makes apc_inc/dec respect the per-entry TTL. If the TTL is expired, the entry is treated as if it doesn't exist and...

@fabicz If the record did not expire, the original TTL is preserved. The TTL is since the time of creation, not since the last update.

Could you please try if updating to APCu 5.1.11 fixes this issue? A bunch of crashes and a deadlock have been fixed since 5.1.8, it's quite likely that one of...

@markus-hsk Thanks for testing this so quickly! Unfortunately I'm not familiar with the debugging procedure on Windows. There are some instructions on [bugs.php.net](https://bugs.php.net/bugs-generating-backtrace-win32.php), but they don't look terribly helpful for...

Assertion failure: ``` rustc: /home/nikic/rust/src/llvm-project/llvm/lib/IR/Function.cpp:117: llvm::Type* llvm::Argument::getParamByValType() const: Assertion `getType()->isPointerTy() && "Only pointers have byval types"' failed. ``` Trace excerpt: ``` #0 __GI_raise (sig=sig@entry=6) at ../sysdeps/unix/sysv/linux/raise.c:51 #1 0x00007ffff4d0e801 in __GI_abort...

Minimal IR reproducer: ``` define x86_intrcc void @test(i128 %x) { ret void } ``` I'm not familiar with this calling convention, but would suspect that having an i128 argument just...

> Honestly, the question I have is why it isn't e.g. returned in a single xmm register instead. It doesn't make sense to me that LLVM would drop a value...

@bors try @rust-timer queue