||Use HEX not DEC in tables
If you're working with the values in section 7 ("Data
Representation"), the difference between hex values and decimal
values seems a little inconsistent. Most of the tables, e.g.
Figure 16 "Tag Encodings" [7.5.4 "Attribute Encodings], pp
125-127 draft8, show their values in hex. Other tables had
only a few entries, with values between 0 and 9, so they used
decimal as a shortcut, e.g. [7.9 "Accessibility Codes"] Figure
24, "Accessibility encodings", p. 141.
However, I think several of these decimal tables have grown in
DWARF3 so they're using values greater than 9, but still in
decimal. Specifically [7.21 "Line Number Information"] Figures
33 & 34 ("Line Number Standard Opcode Encodings" & "Line Number
Extended Opcode Encodings") pp 148-149 and [7.22 "Macro
Information"] Figure 35 "Macinfo Type Encodings" p. 150.
Also there are two tables where the hex values are shown with
only one digit except for a couple of values where they require
two -- [7.15 "Calling Convention Encodings"] Figure 29 p. 144
and [7.8 "Base Type Encodings"] Figure 23 p. 141. Given that
all tables except the Language Encodings uses 8-bit hex values,
this table stands out as peculiar.
I propose that all of these values should be 2-digit hex
values, with the exception of the 4-digit Language Encoding
values and the "High 2 bits" column on the CFI table (Figure
36), which should remain 1 digit.
Even for the tables where they have values [0, 1, 2], for
consistency it would be nice to have them formatted
It's hardly earth-shatteringly important or anything, but it
would be helpful if an implementor is transcribing these from
the standard document. I don't know if these sorts of
modifications are still being accepted to the standard at this
point, of course..