Re: opcodes-2.15.vi clarification

[Ian Lance Taylor]

Clytie Siddall <address@hidden> writes:

> > IC:%s [%s] has no terminals or sub-classes\n
> 
> "terminals" in this case, I am assuming, don't mean VDU installations
> or command-line windows. Would I be correct in translating this to
> mean "terminal variables"? I am unsure. Please dispel a little of the
> fog in which I am struggling. :)

This means terminals in the technical sense.  It means something like
"an item which has been full expanded".  The word is often used in the
context of parsing; perhaps you can find other examples if you can
look at a Vietnamese compiler text.

> > no insns mapped directly to terminal IC %s [%s]
> 
> Does "terminal" here have the same meaning as in #1? Is it an
> adjective applying to "IC", or a noun in its own right? I need to
> know, because adjectives follow the noun in Vietnamese: "terminal IC"
> (if the terminal is describing the IC) would be "IC terminal". I don't
> want to split up the end of the string, however, if it is all part of
> a complex variable...

Terminal can be either an adjective or a noun.  Here it is an
adjective.

(You do know that these are just debugging messages, right?  No normal
user will ever see them.)

> > Warning: rsrc %s (%s) has no chks%s\n
> 
> I am assuming "resource" for "rsrc", and think "chks" is probably
> "checks": is that correct?

"resource" is correct.  "checks" is the English word here; it is used
in the sense of "prevents."

> > rsrc %s (%s) has no regs\n
> 
> "regs": registers? ?

Yes, although it really means a resource which is used by this
instruction.

> > opcode %s has no class (ops %d %d %d)\n
> 
> "ops": operands? operators? operations?

Operands.

> > parse_addr16: invalid opindex.
> 
> "opindex": an index of (again) operands? operators? operations?

An index into the list of operand types, in this case.

> > <internal error in opcode table: %s %s>\n
> 
> "opcode": op? as above (I hope they're not all different, but just in
> case, I'm enquiring for each case of "op" here. If there is only one
> meaning for that abbreviation throughout, please just say so. :)

Here it is opcode, not operand.

> > # <dis error: %08x>
> 
> disassembly/disassembler?

Yes.  This is printed for an instruction which could not be
disassembled.

> > Bad case %d (%s) in %s:%d\n
> 
> "case" as in "upper/lower case", or "case" as in "situation, example"?

The latter.

> > index register in load range
> 
> "index" as in "root, base" or "index" as in "table"?

The latter, I suppose.  An index register is a register used in a
memory address with an optional scaling factor.  It gets the name from
"array index", because that is a common use.  That said, I suspect
that this particular error message should use "base register" instead.

> > unknown operand shift: %x\n
> 
> "shift" as in "change"?

Shift as in left shift or right shift of a number.

> > unknown pop reg: %d\n
> 
> "pop reg" as in "population register"?

No, this is a register which appears in a stack pop instruction.

> > No relocation for small immediate
> 
> "immediate number" or "immediate expression"?

An immediate number in an instruction.


Most, perhaps all, of these messages are debugging messages or error
messages for cases which are supposed to be impossible.  They are not
messages which will ever be seen in any normal circumstances.  It's
not clear to me that these strings should be marked for translation.

Thanks for doing this, though.

> In future, in order to obviate this process with later files, or at
> least make it easier :)  I'd encourage you to include a contact
> address in the Report-Msgid-Bugs-to field, and to use the Developer
> Comments field for each string to explain meanings, to make sure what
> you mean is actually transmitted in the localized version.

Thanks for the suggestions.  I agree that there ought to be something
in the Report-Msgid-Bugs-to field, although I don't myself know how to
set it.  I just see it in opcodes.pot, which is a generated file.
Wishing for better Developer Comments for these cases is, I think,
hopeless.  The messages aren't really comprehensible in English,
either, except to a developer.

Ian