569427 - AvmDebugMsg truncates output

Status: RESOLVED FIXED

(Tamarin Graveyard :: Tools, defect)

Description

[Felix S. Klock II [:pnkfelix, :fklock]]

avmplus::AvmDebugMsg is overloaded with two (well, really three) variants:

    void AvmDebugMsg(bool debuggerBreak, const char* format, ...);
    void AvmDebugMsg(const char* msg, bool debuggerBreak);
    void AvmDebugMsg(const wchar* msg, bool debuggerBreak);

However, it seems to me like the first does not properly handle reasonably long strings.

As an illustration, when I instrumented GC::FindMissingWriteBarriers() as follows: 

    void GC::FindMissingWriteBarriers()
    {
        avmplus::AvmDebugMsg(false, "12345678901234567890\n");
        avmplus::AvmDebugMsg(       "abcdefghijklmnopqrst\n", false);

I saw the following on 64-bit builds:

  ...
  12345678abcdefghijklmnopqrst
  12345678abcdefghijklmnopqrst
  ...

and the following on 32-bit builds:

  ...
  1234abcdefghijklmnopqrst
  1234abcdefghijklmnopqrst
  ...

(Its possible that something is wrong with my build configuration, but the above struck me as quite strange.)

Comment 1

[Felix S. Klock II [:pnkfelix, :fklock]]

Doh.  [[ sizeof(buf-2) ]] != [[ sizeof(buf) - 2 ]].

Comment 2

[Felix S. Klock II [:pnkfelix, :fklock]]

Attachment: belt and suspenders belong at waist, not feet. :)

Comment 3

[Lars T Hansen]

Comment on attachment 448613 [details] [diff] [review]
belt and suspenders belong at waist, not feet.  :)

Doh!

Comment 4

[Felix S. Klock II [:pnkfelix, :fklock]]

(In reply to comment #3)
> (From update of attachment 448613 [details] [diff] [review])
> Doh!

push me please.

Comment 5

[Lars T Hansen]

(In reply to comment #4)
> (In reply to comment #3)
> > (From update of attachment 448613 [details] [diff] [review] [details])
> > Doh!
> 
> push me please.

I'll take care of it tomorrow.

Comment 6

[Lars T Hansen]

tamarin-redux changeset:   4754:f1e6cd506cfb