This partially reverts commit "Windows: Improve readablity of
getenv() encoding" 14453a256d.
The `getenv` macro defined here is to also substitute the function
in dln_find.c, not only in this file.
getenv() did use the expected codepage as an implicit parameter of the macro.
This is mis-leading since include/ruby/win32.h has a different definition.
Using the "cp" variable explicit (like the other function calls) makes it
more readable and consistent.
The compilation for JIT compaction is very heavy. Triggering a second
compaction to include one more new method is probably not worth it. So
this triggers JIT compaction for ten more new methods after each
compaction.
Because d80226e7bd often reduces the number of unloaded units, it
increases the number of unload_units calls, which are heavy.
To mitigate that, this throttles unload_units per `max_cache_size / 10`.
Also hoping to fix
36552382/job/kjmjgw9cjyf2ksd7
I found that fractional seconds were doubled when using the timezone
feature of Time in Sequel's named_timezones extension (which uses
TZInfo for the timezone object), and traced the problem to this code.
There is no subsecx being modified in the utc_to_local call below
this, and I'm not sure why you would want to add in the fractional
seconds unless you assumed the timezone conversion would drop the
existing fractional seconds (TZInfo doesn't drop fractional seconds).
to avoid "Too many JIT code, but skipped unloading units for JIT compaction".
Now we can forget the `in_compact` locking.
Moving some functions from mjit.c to mjit_worker.c because mjit_worker.c
should have functions executed in the JIT worker.
The check output a warning on M1 Mac mini
20201127T074507Z.log.html.gz
```
checking x86intrin.h usability... no
checking x86intrin.h presence... yes
configure: WARNING: x86intrin.h: present but cannot be compiled
configure: WARNING: x86intrin.h: check for missing prerequisite headers?
configure: WARNING: x86intrin.h: see the Autoconf documentation
configure: WARNING: x86intrin.h: section "Present But Cannot Be Compiled"
configure: WARNING: x86intrin.h: proceeding with the compiler's result
checking for x86intrin.h... no
```
Random generators are not Ractor-safe, so we need to prepare
per-ractor default random genearators. This patch set
`Random::DEFAULT = Randm` (not a Random instance, but the Random
class) and singleton methods like `Random.rand()` use a per-ractor
random generator.
[Feature #17322]
Performance is probably improved?
$ benchmark-driver -v --rbenv 'before --jit;after --jit' --repeat-count=12 --alternate --output=all benchmark.yml
before --jit: ruby 3.0.0dev (2020-11-27T04:37:47Z master 69e77e81dc) +JIT [x86_64-linux]
after --jit: ruby 3.0.0dev (2020-11-27T05:28:19Z master df6b05c6dd) +JIT [x86_64-linux]
last_commit=Set VM_FRAME_FLAG_FINISH at once
Calculating -------------------------------------
before --jit after --jit
Optcarrot Lan_Master.nes 80.89292998533379 82.19497327502751 fps
80.93130641142331 85.13943315260148
81.06214830270119 87.43757879797808
82.29172808453910 87.89942441487113
84.61206450455929 87.91309779491075
85.44545883567997 87.98026086648694
86.02923132404449 88.03081060383973
86.07411817365879 88.14650206137341
86.34348799602836 88.32791633649961
87.90257338977324 88.57599644892220
88.58006509876580 88.67426384743277
89.26611118140011 88.81669430874207
This should have no bad impact on VM because this function is ALWAYS_INLINE.
Both explicit compaction routines (gc_compact and the verify references form)
need to clear the heap before executing compaction. Otherwise some
objects may not be alive, and we'll need the read barrier. The heap
must only contain *live* objects if we want to disable the read barrier
during explicit compaction.
The previous commit was missing the "clear the heap" phase from the
"verify references" explicit compaction function.
Fixes [Bug #17306]
transient_heap_evacuate() evacuates all living objects from theap
to malloc mangaed memory. This evacuation can run in any ractors
in parallel with other ractor's execution and it is danger.
To stop such danger parallel execution, use rb_vm_barrier() to
stop all other ractors.
Ractor's error will be printed if Thread#report_on_exception
is true (default), and error message is used. Without this patch,
the exception object is sent by Ractor.yield and it can be shared
with another ractor.
http://ci.rvm.jp/results/trunk-random3@phosphorus-docker/3269368
To prevent such sharing, show errors befor Ractor.yield().
