501614 - jsinterp.cpp:1705: warning: ‘isFunction’ may be used uninitialized in this function

Status: RESOLVED WORKSFORME

(Core :: JavaScript Engine, defect)

Description

[Julian Seward [:jseward]]

User-Agent:       Mozilla/5.0 (X11; U; Linux i686 (x86_64); en-US; rv:1.9.1) Gecko/20090624 Firefox/3.5
Build Identifier: 

Branch: tracemonkey.  With a debug build of js configured thusly

CC="gcc -m32" CXX="g++ -m32" AR=ar ../configure     --enable-debug  \
   --enable-optimize="-g -O" --target=i686-pc-linux-gnu

gcc-4.3.1 complains that
jsinterp.cpp:1705: warning: ‘isFunction’ may be used uninitialized in this function

and at least from a naive interpretation of the control flow, it is
possible to get from the entrance to a use point, which I suspect is

    type = (attrs == JSPROP_INITIALIZER)
           ? "property"
           : (oldAttrs & attrs & JSPROP_GETTER)
           ? js_getter_str
           : (oldAttrs & attrs & JSPROP_SETTER)
           ? js_setter_str
           : (oldAttrs & JSPROP_READONLY)
           ? js_const_str
           : isFunction              <-----------------------
           ? js_function_str
           : js_var_str;

without going through a definition.

Merely changing the declaration of isFunction thusly shuts it up and
does not appear to break anything:

    JSBool isFunction = false; /* keeps gcc happy */

It may be that I'm seeing this because my debug build is still done with
gcc -O (hence dataflow analysis), which is probably unusual.


Reproducible: Always

Comment 1

[Steven Johnson]

strictly speaking this isn't a nanojit bug, as jsinterp.cpp doesn't live there.

Comment 2

[Atul Aggarwal]

I could not see any warning in the jsinterp.cpp file. I also verified this with the debug build and with nightly build of linux and macos.

Hence, closing this bug.