[GR-39230] Backports for 22.2

PullRequest: truffleruby/3431

Author: eregon

CHANGELOG.md

@@ -1,20 +1,3 @@
-# 22.3.0
-
-New features:
-
-
-Bug fixes:
-
-
-Compatibility:
-
-
-Performance:
-
-
-Changes:
-
-
 # 22.2.0
 
 New features:
@@ -40,6 +23,8 @@ Bug fixes:
 * Fix `/#{...}/o` to evaluate only once per context when splitting happens (@eregon).
 * Fix `Kernel#sprintf` formatting of floats to be like CRuby (@aardvark179).
 * Fix `Process.egid=` to accept `String`s (#2615, @ngtban)
+* Fix optional assignment to only evaluate index arguments once (#2658, @aardvark179).
+* Fix `StringIO` to set position correctly after reading multi-byte characters (#2207, @aardvark179).
 
 Compatibility:
 
@@ -61,13 +46,15 @@ Compatibility:
 * Use `$PAGER` for `--help` and `--help*`, similar to CRuby (#2542, @Strech).
 * Ensure all headers are warnings-free (#2662, @eregon).
 * All `IO` instances should have `T_FILE` as their `rb_type()`, not only `File` instances (#2662, @eregon).
+* Make `rb_fd_select` retry on `EINTR` (#1584, @aardvark179).
 
 Performance:
 
 * Reimplement `Float#to_s` for better performance (#1584, @aardvark179).
 * Improve reference processing by making C object free functions and other finalizers more lightweight (@aardvark179).
 * Improve performance of `RSTRING_PTR` for interned strings (@aardvark179).
 * Cache constant argument formats used with `rb_scan_args_kw` (@aardvark179).
+* Report polymorphism inside `Hash#[]` to recover performance (@aardvark179).
 
 Changes:
 

ci.jsonnet

@@ -271,7 +271,7 @@ local part_definitions = {
         bench_machine: ["x52"] + self.normal_machine + ["no_frequency_scaling"],
       },
       docker: {
-        image: "phx.ocir.io/oraclelabs2/c_graal/buildslave:buildslave_ol7",
+        image: "buildslave_ol7",
         mount_modules: true,
       },
     },

common.json

@@ -4,18 +4,18 @@
   "jdks": {
     "openjdk11":          {"name": "openjdk",   "version": "11.0.11+9", "platformspecific": true },
     "oraclejdk11":        {"name": "oraclejdk", "version": "11.0.11+9", "platformspecific": true },
-    "labsjdk-ce-11":      {"name": "labsjdk",   "version": "ce-11.0.16+5-jvmci-22.2-b03", "platformspecific": true },
-    "labsjdk-ce-11-llvm": {"name": "labsjdk",   "version": "ce-11.0.16+5-jvmci-22.2-b03-sulong", "platformspecific": true },
-    "labsjdk-ee-11":      {"name": "labsjdk",   "version": "ee-11.0.16+9-jvmci-22.2-b03", "platformspecific": true },
-    "labsjdk-ee-11-llvm": {"name": "labsjdk",   "version": "ee-11.0.16+9-jvmci-22.2-b03-sulong", "platformspecific": true },
+    "labsjdk-ce-11":      {"name": "labsjdk",   "version": "ce-11.0.16+7-jvmci-22.2-b04", "platformspecific": true },
+    "labsjdk-ce-11-llvm": {"name": "labsjdk",   "version": "ce-11.0.16+7-jvmci-22.2-b04-sulong", "platformspecific": true },
+    "labsjdk-ee-11":      {"name": "labsjdk",   "version": "ee-11.0.16+11-jvmci-22.2-b05", "platformspecific": true },
+    "labsjdk-ee-11-llvm": {"name": "labsjdk",   "version": "ee-11.0.16+11-jvmci-22.2-b05-sulong", "platformspecific": true },
 
     "oraclejdk17":        {"name": "oraclejdk", "version": "17.0.1+12", "platformspecific": true },
-    "labsjdk-ce-17":      {"name": "labsjdk",   "version": "ce-17.0.4+5-jvmci-22.2-b03", "platformspecific": true },
-    "labsjdk-ce-17Debug": {"name": "labsjdk",   "version": "ce-17.0.4+5-jvmci-22.2-b03-debug", "platformspecific": true },
-    "labsjdk-ce-17-llvm": {"name": "labsjdk",   "version": "ce-17.0.4+5-jvmci-22.2-b03-sulong", "platformspecific": true },
-    "labsjdk-ee-17":      {"name": "labsjdk",   "version": "ee-17.0.4+9-jvmci-22.2-b03", "platformspecific": true },
-    "labsjdk-ee-17Debug": {"name": "labsjdk",   "version": "ee-17.0.4+9-jvmci-22.2-b03-debug", "platformspecific": true },
-    "labsjdk-ee-17-llvm": {"name": "labsjdk",   "version": "ee-17.0.4+9-jvmci-22.2-b03-sulong", "platformspecific": true }
+    "labsjdk-ce-17":      {"name": "labsjdk",   "version": "ce-17.0.4+7-jvmci-22.2-b04", "platformspecific": true },
+    "labsjdk-ce-17Debug": {"name": "labsjdk",   "version": "ce-17.0.4+7-jvmci-22.2-b04-debug", "platformspecific": true },
+    "labsjdk-ce-17-llvm": {"name": "labsjdk",   "version": "ce-17.0.4+7-jvmci-22.2-b04-sulong", "platformspecific": true },
+    "labsjdk-ee-17":      {"name": "labsjdk",   "version": "ee-17.0.4+11-jvmci-22.2-b05", "platformspecific": true },
+    "labsjdk-ee-17Debug": {"name": "labsjdk",   "version": "ee-17.0.4+11-jvmci-22.2-b05-debug", "platformspecific": true },
+    "labsjdk-ee-17-llvm": {"name": "labsjdk",   "version": "ee-17.0.4+11-jvmci-22.2-b05-sulong", "platformspecific": true }
   },
 
   "COMMENT.devkits" : "The devkits versions reflect those used to build the JVMCI JDKs (e.g., see devkit_platform_revisions in <jdk>/make/conf/jib-profiles.js)",

lib/cext/ABI_check.txt

@@ -1 +1 @@
-5
+6

lib/truffle/stringio.rb

@@ -211,7 +211,7 @@ def each_byte
     d = @__data__
     string = d.string
 
-    while d.pos < string.length
+    while d.pos < string.bytesize
       check_readable
       byte = string.getbyte d.pos
       d.pos += 1
@@ -468,7 +468,7 @@ def read(length = nil, buffer = nil)
       buffer.replace str if buffer
     end
 
-    d.pos += str.length
+    d.pos += str.bytesize
     str
   end
 

lib/truffle/truffle/openssl-prefix.rb

@@ -8,10 +8,6 @@
 
 # Set OPENSSL_PREFIX in ENV to find the OpenSSL headers
 
-module Truffle
-  OPENSSL_PREFIX_WAS_SET = ENV.key?('OPENSSL_PREFIX')
-end
-
 search_homebrew = -> homebrew {
   if prefix = "#{homebrew}/opt/[email protected]" and Dir.exist?(prefix)
     prefix
@@ -20,10 +16,9 @@ module Truffle
   end
 }
 
-macOS = RUBY_PLATFORM.include?('darwin')
-
-if macOS && !ENV['OPENSSL_PREFIX']
-  if prefix = search_homebrew.call('/usr/local')
+if Truffle::Platform.darwin? && !ENV['OPENSSL_PREFIX']
+  default_homebrew_prefix = Truffle::System.host_cpu == 'aarch64' ? '/opt/homebrew' : '/usr/local'
+  if prefix = search_homebrew.call(default_homebrew_prefix)
     # found
   else
     homebrew = `brew --prefix 2>/dev/null`.strip

mx.truffleruby/suite.py

@@ -7,7 +7,7 @@
             {
                 "name": "regex",
                 "subdir": True,
-                "version": "f33384d85f86bc5f4fd1a92d9d734309c64d381d",
+                "version": "105d87b99fc5879433f9b1c54040200530f9c1aa",
                 "urls": [
                     {"url": "https://github.com/oracle/graal.git", "kind": "git"},
                     {"url": "https://curio.ssw.jku.at/nexus/content/repositories/snapshots", "kind": "binary"},
@@ -16,7 +16,7 @@
             {
                 "name": "sulong",
                 "subdir": True,
-                "version": "f33384d85f86bc5f4fd1a92d9d734309c64d381d",
+                "version": "105d87b99fc5879433f9b1c54040200530f9c1aa",
                 "urls": [
                     {"url": "https://github.com/oracle/graal.git", "kind": "git"},
                     {"url": "https://curio.ssw.jku.at/nexus/content/repositories/snapshots", "kind": "binary"},

spec/ruby/language/optional_assignments_spec.rb

@@ -300,6 +300,44 @@ def []=(k, v)
         (@b[:k] ||= 12).should == 12
       end
 
+      it 'correctly handles a splatted argument for the index' do
+        (@b[*[:k]] ||= 12).should == 12
+      end
+
+      it "evaluates the index precisely once" do
+        ary = [:x, :y]
+        @a[:x] = 15
+        @a[ary.pop] ||= 25
+        ary.should == [:x]
+        @a.should == { x: 15, y: 25 }
+      end
+
+      it "evaluates the index arguments in the correct order" do
+        ary = Class.new(Array) do
+          def [](x, y)
+            super(x + 3 * y)
+          end
+
+          def []=(x, y, value)
+            super(x + 3 * y, value)
+          end
+        end.new
+        ary[0, 0] = 1
+        ary[1, 0] = 1
+        ary[2, 0] = nil
+        ary[3, 0] = 1
+        ary[4, 0] = 1
+        ary[5, 0] = 1
+        ary[6, 0] = nil
+
+        foo = [0, 2]
+
+        ary[foo.pop, foo.pop] ||= 2
+
+        ary[2, 0].should == 2
+        ary[6, 0].should == nil
+      end
+
       it 'returns the assigned value, not the result of the []= method with +=' do
         @b[:k] = 17
         (@b[:k] += 12).should == 29

spec/ruby/library/stringio/shared/read.rb

@@ -89,6 +89,12 @@
     @io.send(@method)
     @io.pos.should eql(7)
   end
+
+  it "correctly update the current position in bytes when multi-byte characters are used" do
+    @io.print("example\u03A3") # Overwrite the original string with 8 characters containing 9 bytes.
+    @io.send(@method)
+    @io.pos.should eql(9)
+  end
 end
 
 describe :stringio_read_nil, shared: true do

src/main/c/cext/fd.c

@@ -9,7 +9,9 @@
  */
 #include <truffleruby-impl.h>
 #include <fcntl.h>
+#include <errno.h>
 #include <ruby/thread.h>
+#include <time.h>
 
 // For howmany()
 #ifdef HAVE_SYS_PARAM_H
@@ -105,6 +107,65 @@ void rb_fd_dup(rb_fdset_t *dst, const rb_fdset_t *src) {
   memcpy(dst->fdset, src->fdset, size);
 }
 
+void rb_fd_init_copy(rb_fdset_t *dst, rb_fdset_t *src) {
+  size_t size = howmany(rb_fd_max(src), NFDBITS) * sizeof(fd_mask);
+
+  if (size < sizeof(fd_set)) {
+    size = sizeof(fd_set);
+  }
+  dst->maxfd = src->maxfd;
+  dst->fdset = xmalloc(size);
+  memcpy(dst->fdset, src->fdset, size);
+}
+
+static bool timespec_subtract(struct timespec *result, struct timespec x, struct timespec y) {
+  /* Perform the carry for the later subtraction by updating y. */
+  if (x.tv_nsec < y.tv_nsec) {
+    long nsec = (y.tv_nsec - x.tv_nsec) / 1000000000 + 1;
+    y.tv_nsec -= 1000000000 * nsec;
+    y.tv_sec += nsec;
+  }
+  if (x.tv_nsec - y.tv_nsec > 1000000000) {
+    long nsec = (x.tv_nsec - y.tv_nsec) / 1000000000;
+    y.tv_nsec += 1000000000 * nsec;
+    y.tv_sec -= nsec;
+  }
+
+  /* Compute the time remaining to wait.
+     tv_nsec is certainly positive. */
+  result->tv_sec = x.tv_sec - y.tv_sec;
+  result->tv_nsec = x.tv_nsec - y.tv_nsec;
+
+  /* Return 1 if result is negative. */
+  return x.tv_sec < y.tv_sec;
+}
+
+static bool timeval_subtract(struct timeval *result, struct timeval x, struct timeval y) {
+  /* Perform the carry for the later subtraction by updating y. */
+  if (x.tv_usec < y.tv_usec) {
+    long usec = (y.tv_usec - x.tv_usec) / 1000000 + 1;
+    y.tv_usec -= 1000000 * usec;
+    y.tv_sec += usec;
+  }
+  if (x.tv_usec - y.tv_usec > 1000000) {
+    long usec = (x.tv_usec - y.tv_usec) / 1000000;
+    y.tv_usec += 1000000 * usec;
+    y.tv_sec -= usec;
+  }
+
+  /* Compute the time remaining to wait.
+     tv_usec is certainly positive. */
+  result->tv_sec = x.tv_sec - y.tv_sec;
+  result->tv_usec = x.tv_usec - y.tv_usec;
+
+  /* Return 1 if result is negative. */
+  return x.tv_sec < y.tv_sec;
+}
+
+static int should_retry(int result) {
+  return (result < 0) && (errno == EINTR);
+}
+
 int rb_fd_select(int n, rb_fdset_t *readfds, rb_fdset_t *writefds, rb_fdset_t *exceptfds, struct timeval *timeout) {
   fd_set *r = NULL, *w = NULL, *e = NULL;
   if (readfds) {
@@ -122,21 +183,90 @@ int rb_fd_select(int n, rb_fdset_t *readfds, rb_fdset_t *writefds, rb_fdset_t *e
   return select(n, r, w, e, timeout);
 }
 
+
 // NOTE: MRI's version has more fields
 struct select_set {
   int max;
   rb_fdset_t *rset;
   rb_fdset_t *wset;
   rb_fdset_t *eset;
+  rb_fdset_t *orig_rset;
+  rb_fdset_t *orig_wset;
+  rb_fdset_t *orig_eset;
   struct timeval *timeout;
+  struct timeval *orig_timeout;
 };
 
+static inline void restore_fds(rb_fdset_t *dst, rb_fdset_t *src) {
+  if (dst) {
+    rb_fd_dup(dst, src);
+  }
+}
+
+static bool update_timeout(struct timeval *timeout, struct timeval *orig_timeout, struct timespec *starttime) {
+  struct timespec currenttime;
+  struct timespec difftime;
+  struct timeval difftimeout;
+  bool timeleft = true;
+  if (timeout) {
+    clock_gettime(CLOCK_MONOTONIC, &currenttime);
+    timespec_subtract(&difftime, currenttime, *starttime);
+    difftimeout.tv_sec = difftime.tv_sec;
+    difftimeout.tv_usec = difftime.tv_nsec / 1000;
+    timeleft = timeval_subtract(timeout, *orig_timeout, difftimeout);
+  }
+
+  return timeleft;
+}
+
 static void* rb_thread_fd_select_blocking(void *data) {
   struct select_set *set = (struct select_set*)data;
-  int result = rb_fd_select(set->max, set->rset, set->wset, set->eset, set->timeout);
+  struct timespec starttime;
+
+  if (set->timeout) {
+    clock_gettime(CLOCK_MONOTONIC, &starttime);
+  }
+
+  int result = 0;
+  bool timeleft = true;
+  do {
+    restore_fds(set->rset, set->orig_rset);
+    restore_fds(set->wset, set->orig_wset);
+    restore_fds(set->eset, set->orig_eset);
+    timeleft = update_timeout(set->timeout, set->orig_timeout, &starttime);
+    if (!timeleft) {
+      break;
+    }
+    result = rb_fd_select(set->max, set->rset, set->wset, set->eset, set->timeout);
+  } while (should_retry(result));
   return (void*)(long)result;
 }
 
+static void* rb_thread_fd_select_internal(void *sets) {
+  return rb_thread_call_without_gvl(rb_thread_fd_select_blocking, sets, RUBY_UBF_IO, 0);
+}
+
+static void rb_thread_fd_select_set_free(struct select_set *sets) {
+  if (sets->orig_rset) {
+    rb_fd_term(sets->orig_rset);
+  }
+  if (sets->orig_wset) {
+    rb_fd_term(sets->orig_wset);
+  }
+  if (sets->orig_eset) {
+    rb_fd_term(sets->orig_eset);
+  }
+}
+
+static void fd_init_copy(rb_fdset_t *dst, int max, rb_fdset_t *src) {
+  if (src) {
+    rb_fd_resize(max - 1, src);
+    if (dst != src) {
+      rb_fd_init_copy(dst, src);
+    }
+  }
+}
+
 int rb_thread_fd_select(int max, rb_fdset_t *read, rb_fdset_t *write, rb_fdset_t *except, struct timeval *timeout) {
   // NOTE: MRI has more logic in here
   struct select_set set;
@@ -145,7 +275,12 @@ int rb_thread_fd_select(int max, rb_fdset_t *read, rb_fdset_t *write, rb_fdset_t
   set.wset = write;
   set.eset = except;
   set.timeout = timeout;
+  fd_init_copy(set.orig_rset, set.max, set.rset);
+  fd_init_copy(set.orig_wset, set.max, set.wset);
+  fd_init_copy(set.orig_eset, set.max, set.eset);
+  struct timeval orig_timeval = *timeout;
+  set.orig_timeout = &orig_timeval;
 
-  void* result = rb_thread_call_without_gvl(rb_thread_fd_select_blocking, (void*)(&set), RUBY_UBF_IO, 0);
+  void* result = rb_ensure(rb_thread_fd_select_internal, (VALUE)&set, rb_thread_fd_select_set_free, (VALUE)&set);
   return (int)(long)result;
 }