Bug 5993 – Strange thread exceptions with DMD 2.053 beta 13th May

Strange exceptions from within threads for code compiled with dmd 2.053 beta (13th May and previous) not seen using 2.052 or 2.051. Backtrace: Thread died unexpectedly: %score.thread.ThreadException@src/core/thread.d(2338): Unable to suspend thread ---------------- ./.bob/bob(extern (C) void core.thread.thread_suspendAll().void suspend(core.thread.Thread)+0x86) [0x80f5d7a] ./.bob/bob(thread_suspendAll+0x7f) [0x80f5cdf] ./.bob/bob(uint gc.gcx.Gcx.fullcollect(void*)+0x14) [0x80ed6c0] ./.bob/bob(uint gc.gcx.Gcx.fullcollectshell()+0x26) [0x80ed69e] ./.bob/bob(void* gc.gcx.Gcx.bigAlloc(uint, gc.gcx.Pool**, uint*)+0x8a) [0x80ed226] ./.bob/bob(void* gc.gcx.GC.mallocNoSync(uint, uint, uint*)+0x11d) [0x80eb9ed] ./.bob/bob(void* gc.gcx.GC.malloc(uint, uint, uint*)+0x4b) [0x80eb8a3] ./.bob/bob(gc_malloc+0x1e) [0x80cb4ae] ./.bob/bob(void* core.memory.GC.malloc(uint, uint)+0xf) [0x80c8173] ./.bob/bob(void[] std.file.read(const(char[]), const(uint))+0xc2) [0x80dc972] ./.bob/bob(immutable(char)[] std.file.readText!(immutable(char)[]).readText(const(char[]))+0x16) [0x80bdfd6] ./.bob/bob(void bob.read_options()+0x73) [0x80b46f3] ... < SNIP > Thread died unexpectedly: %sobject.Error: Got signal 11 SIGSEGV ---------------- ./.bob/bob(signalHandler+0x4f) [0x80c5abb] [0x33f400] ./.bob/bob(gc.gcx.BlkInfo gc.gcx.GC.queryNoSync(void*)+0x17) [0x80ec493] ./.bob/bob(gc.gcx.BlkInfo gc.gcx.GC.query(void*)+0x80) [0x80ec448] ./.bob/bob(gc_query+0x26) [0x80cb6f6] ./.bob/bob(_d_arrayappendcTX+0x5e) [0x80ce19e] ./.bob/bob(_d_arrayappendT+0x2a) [0x80ce09a] ./.bob/bob(bob.Obj bob.Obj.__ctor(bob.Unit, bob.File, immutable(char)[], bob.DynamicLib[], bob.StaticLib[])+0x177) [0x80b7393] ./.bob/bob(bob.Obj bob.Obj.__ctor(bob.Unit, immutable(char)[], immutable(char)[], bob.DynamicLib[], bob.StaticLib[])+0xc7) [0x80b75 ... < SNIP >

Changed summary to indicate it's a new bug, not a regression.

Created attachment 979 Backtrace I encountered the same GC issue when updating ldc2 to the new frontend and druntime.

Is this still an issue in 2.069? Is there a test case?

(In reply to Infiltrator from comment #3) > Is this still an issue in 2.069? IIRC this was fixed and the user code that broke resumed working. I think this bug is a relic now. Apologies for not following it up.