RFR 8059631: Use of '#' to represent MethodHandle kind is confusing

[John Rose]

On Jun 17, 2016, at 4:41 PM, Jonathan Gibbons <jonathan.gibbons at oracle.com> wrote:
> 
> On 06/17/2016 04:33 PM, John Rose wrote:
>> On Jun 17, 2016, at 3:53 PM, Jonathan Gibbons <jonathan.gibbons at oracle.com <mailto:jonathan.gibbons at oracle.com>> wrote:
>>> 
>>> I'll do a build and test run;  I think it should go if it is unused.
>>> 
>>> I wasn't sure at first, since most of the names appear in the 3rd column of JVMS 8, Table 5.4.3.5-A
>>> but the oddball is REF_newInvokeSpecial(8, "newinvokespecial") which isn't a JVMS name.
>> 
>> That's because it corresponds to two instructions, new + invokespecial.
> 
> Sure, I guessed that, but it didn't seem useful to include that in a "name" field, when the enum name
> itself is more indicative of a name defined in JVMS.

Oh, right.  I guess the "name" fields are trying to look like instruction names, but that fails for new + invokespecial.
I agree it's simpler just to use the enum names.

— John
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.openjdk.java.net/pipermail/compiler-dev/attachments/20160617/58902c30/attachment-0001.html>