Multisection Location Lists comments

[David B Anderson <davea at quasar dot engr dot sgi dot com>]

A few small comments on the proposal on  Multisection Location Lists



Ron writes, in the propsal on Multisection Location Lists:
||       "Each entry in a range list is either a base address selection
||       entry (see Section 3.1.2) or a range list entry. A range list
|        entry consists of:


Since there are now *three* things here, I suggest

"Each entry in a range list is either a base address selection
 entry (see Section 3.1.2), a location list terminator, 
 or a range list entry. A range list
 entry consists of:

And where a "location list terminator" is what is
described but not named a few lines below in Sec 2.5.4. A 0,0 entry.
Also one without any 'location block'.

That makes it easier to be clear what does and does not have
a location block after the address range pair.




[ ]
|  - Define an attribute DW_AT_base_addresses, which may be used only
|   for the compilation unit DIE, and whose value is a sequence of
    ^^^^^^^
		in a compilation unit DIE, ...

|   addresses where each address is the beginning address of one of
|   the text program sections that hold the machine code.
[ ]



|DW_AT_base_addresses
|
|This attribute is used only on a compilation unit DIE and is used in
|combination with DW_AT_ranges (and instead of DW_AT_low_pc and
|DW_AT_high_pc). The value of this attribute is a sequence of addresses,
|represented using class block, where each address is the beginning address
|of one of the (text) sections that contain some of the machine code for
|the compilation unit.

This means, if I read it correctly, that there is no provision for
'empty' entries in the block.  The block must be exactly right-sized.    
And the size of the block is a multiple of address size.

I can't see how this is a problem, I just wanted to be sure I understood.


Regards,
David B. Anderson davea@sgi.com danderson@acm.org http://reality.sgi.com/davea/