The distance between the core published API and the actual layer at which
system services requests are made, is very far. The is a set of .java routines
that just dont get changed, and are common between win&solaris machines. Then
there is a set of '.java' routines, that are particular to windoz, and a set
for solaris ports. These particular java routines take care of the necessary
transslations between the OS and the core java routines that just dont get
changed. System pecularities should never have to go beyond the 'system'
interface routines, or bubble up to, or beyond these service routines.
But eventually someone is gonna want to try to build a >4gig file ( >32bit
seeks ) on the alpha/linux box. Maybe someone is gonna try to read from a
16gig file. These are all valid uses. And Should I limit these people to 32bit
file sizes, or allow them the full use of the OS capabilities? And what of the
reverse - If I create a 14gig file on alpha/linux using the Alpha JDK and
attempt to shuve it onto an intel box, you may not be able to move that file to
that intel box., because the file system may not be able to handle that file
size. Would u blame the OS, the Intel JDK, or the Alpha JDK.
>From the front everything appears to be homogenious, (even on my alpha/linux
port ) but in the back there are some real life issues. And they tend to get
resolved without the java user ever noticing.
Tell me, did Microsoft alter the published core API to suit their own
sensibilities? I dont know, as I didnt pay much attention. Or did they just
manipulate the back-end services.
Cees de Groot wrote:
> Uncle George:
>
> >Actually u do have to change the '.java' files ( simply for practicable
> >convience ). Having extensive experience in porting sun's JDK, I find
> >it rather difficult to store 64bits into a 32bit java int. Maybe u can
> >review the java-linux-porting diffs of their 1.2 port. I'd like to know,
> >for example, how the solaris printer 'queue' switches was changed to be
> >compatable with linux printer 'queue' commands.
>
> Where do 32/64bit differences on the platform bubble up to the java.*
> level that you would need to change this? If there are such locations,
> these should be reported as bugs because on the java level you shouldn't
> be able to determine the native wordsize of the underlying platform.
>
> Changing the wordsize of standard API level interfaces during a port
> would completely defeat the purpose of the virtual machine, and I for
> one are very glad that Sun wouldn't allow this. If your Alpha port
> has changed ints and longs in API .java files it's just as much Java
> as Microsoft's stuff is...
>
>
----------------------------------------------------------------------
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
