FAQ
Reviewers: golang-dev1,

Message:
Hello golang-dev@googlegroups.com (cc: golang-dev@googlegroups.com),

I'd like you to review this change to
https://go.googlecode.com/hg/


Description:
cmd/gc: compute initialization order for top-level blank vars too.

Fixes issue 5244.

Please review this at https://codereview.appspot.com/8601044/

Affected files:
    M src/cmd/gc/sinit.c
    A test/fixedbugs/issue5244.go


Index: src/cmd/gc/sinit.c
===================================================================
--- a/src/cmd/gc/sinit.c
+++ b/src/cmd/gc/sinit.c
@@ -53,9 +53,7 @@
     if(isblank(n) && n->curfn == N && n->defn != N && n->defn->initorder ==
InitNotStarted) {
      // blank names initialization is part of init() but not
      // when they are inside a function.
- n->defn->initorder = InitDone;
- if(debug['%']) dump("nonstatic", n->defn);
- *out = list(*out, n->defn);
+ break;
     }
     return;
    }
@@ -130,7 +128,7 @@
       init2(n->defn->right, out);
       if(debug['j'])
        print("%S\n", n->sym);
- if(!staticinit(n, out)) {
+ if(isblank(n) || !staticinit(n, out)) {
        if(debug['%']) dump("nonstatic", n->defn);
        *out = list(*out, n->defn);
       }
Index: test/fixedbugs/issue5244.go
===================================================================
new file mode 100644
--- /dev/null
+++ b/test/fixedbugs/issue5244.go
@@ -0,0 +1,18 @@
+// run
+
+// Copyright 2013 The Go Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+// Issue 5244: the init order computation uses the wrong
+// order for top-level blank identifier assignments.
+// The example used to panic because it tries calling a
+// nil function instead of assigning to f before.
+
+package main
+
+var f = func() int { return 1 }
+var _ = f() + g()
+var g = func() int { return 2 }
+
+func main() {}


--

---
You received this message because you are subscribed to the Google Groups "golang-dev" group.
To unsubscribe from this group and stop receiving emails from it, send an email to golang-dev+unsubscribe@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.

Search Discussions

  • Rsc at Jun 10, 2013 at 10:25 pm
    LGTM

    Ian, what do you think about this for Go 1.1.1?


    https://codereview.appspot.com/8601044/

    --

    ---
    You received this message because you are subscribed to the Google Groups "golang-dev" group.
    To unsubscribe from this group and stop receiving emails from it, send an email to golang-dev+unsubscribe@googlegroups.com.
    For more options, visit https://groups.google.com/groups/opt_out.
  • Ian Lance Taylor at Jun 10, 2013 at 10:35 pm

    On Mon, Jun 10, 2013 at 3:25 PM, wrote:
    LGTM

    Ian, what do you think about this for Go 1.1.1?
    I don't know gc well enough to judge the safety, but in general I'm in
    favor. It doesn't quite meet the criteria, in that there is a
    workaround, but the failure mode doesn't really lead to that
    workaround.

    Ian

    --

    ---
    You received this message because you are subscribed to the Google Groups "golang-dev" group.
    To unsubscribe from this group and stop receiving emails from it, send an email to golang-dev+unsubscribe@googlegroups.com.
    For more options, visit https://groups.google.com/groups/opt_out.
  • R at Jun 10, 2013 at 10:57 pm
    I think 1.1.1 is warranted

    https://codereview.appspot.com/8601044/

    --

    ---
    You received this message because you are subscribed to the Google Groups "golang-dev" group.
    To unsubscribe from this group and stop receiving emails from it, send an email to golang-dev+unsubscribe@googlegroups.com.
    For more options, visit https://groups.google.com/groups/opt_out.

Related Discussions

Discussion Navigation
viewthread | post
Discussion Overview
groupgolang-dev @
categoriesgo
postedJun 10, '13 at 9:38p
activeJun 10, '13 at 10:57p
posts4
users4
websitegolang.org

People

Translate

site design / logo © 2022 Grokbase