[OpenJDK 2D-Dev] <AWT Dev> RFR: Even more gcc warnings (8035287)
mikael.vidstedt at oracle.com
Wed Feb 19 01:30:01 UTC 2014
On 2014-02-15 09:31, Phil Race wrote:
> Looks OK to me although I just realised there's no bug ID here
> FWIW I develop on WIndows, Mac & Linux and I've noticed widely divergent
> things that the compilers on these platforms warn about. Warning
> free on Linux might not mean warning free on Windows.
> Also, assuming you develop on Linux might want to check if any of
> these make
> the VS compiler less happy about anything.
Acknowledged - not all platforms/compilers complain about the same
thing(s). I tried my best to manually verify that no new warnings are
introduced by building on the usual suspect platforms and grep through
>> * src/solaris/native/sun/awt/awt_Font.c
>> Comparisons with string literals is undefined behavior - keep track
>> of whether the string should be freed explicitly with a boolean instead.
> Gosh .. that code must be from 1996 or thereabouts.
I hope touching it doesn't mean I own it ;)
Anything else I should do/test?
> On 2/15/14 8:37 AM, Mikael Vidstedt wrote:
>> Corrected link - this webrev is based on jdk9/client:
>>> On Feb 14, 2014, at 17:54, Mikael Vidstedt
>>> <mikael.vidstedt at oracle.com> wrote:
>>> A drive-by set of warning fixes:
>>> * src/share/native/com/sun/java/util/jar/pack/bands.cpp
>>> Set the size of the array explicitly to increase likelihood of enum
>>> and struct array being in sync. Arguably this should be changed to
>>> use the (new) [<enumName>] = <value> instead.
>>> Initialize all the fields in the (redundant) terminator struct
>>> Remove unused macro.
>>> * src/share/native/sun/java2d/opengl/OGLContext.c
>>> Get the prototype for jio_snprintf from jvm.h to address an implicit
>>> * src/solaris/native/sun/awt/awt_Font.c
>>> Comparisons with string literals is undefined behavior - keep track
>>> of whether the string should be freed explicitly with a boolean
>>> * src/solaris/native/sun/awt/awt_LoadLibrary.c
>>> The macro is supposed to expand to a void function declaration, but
>>> forgets to actually add the "void".
More information about the 2d-dev