Re: [PATCH v4] gdb/manual: Introduce location specs

[Pedro Alves <pedro@palves.net>]

On 2022-05-27 15:16, Eli Zaretskii wrote:
>> From: Pedro Alves <pedro@palves.net>
>> Date: Thu, 26 May 2022 20:42:50 +0100
>>
>>  gdb/doc/gdb.texinfo | 419 +++++++++++++++++++++++++-------------------
>>  gdb/doc/guile.texi  |   2 +-
>>  gdb/doc/python.texi |   5 +-
>>  3 files changed, 246 insertions(+), 180 deletions(-)
> 
> Thanks.  The changes in descriptions of the commands that accept
> location specs seem more-or-less okay, although I didn't yet read all
> of them  in detail (so please don't push just yet).  The "Location
> Specifications" section, OTOH, needs more work.  I didn't yet make up
> my mind regarding what exactly should it look like, and one reason why
> I'm not there yet is that your text doesn't seem to describe the "code
> location" in full:
> 
>   > +A concrete code location in your program is uniquely identifiable by a
>   > +set of logical attributes.  Typically, a line number, the source file
>   > +the line belongs to, the fully-qualified and prototyped function it is
>   > +defined in, and an instruction address.  Because each inferior has its
>   > +own address space, also an inferior number.
> 
> The "typically" part and the overall style seem to say that this is
> not the exhaustive list of all the attributes of a code location, just
> a general idea.  Can you please present a full exhaustive list of the
> attributes of a code location?

I meant to remove the "typically", and forgot it, sorry.  It is not
supposed to be there.

> 
> And another question: does the process of resolving a location spec to
> obtain the corresponding code locations involve only filling in of the
> attributes that were omitted from the spec, or does it also produce
> attributes that can _never_ be part of the location spec?  IOW, can
> the user type a location spec which will yield a code location that is
> 100% identical to the input spec?

Currently there's no way to explicitly specify the inferior with
any format of location specifications.  So if you do "b func", and you
have multiple inferiors, GDB will find code locations for "func" in all
the inferiors.  All the other attributes you can explicitly specify.
Not sure whether that answers your question.  I am not sure what you
mean by "100% identical".  A spec can never the identical to the actual
thing, the same way a cake recipe can never be identical to a cake, for
they are things of different nature.  It can only be that the actual
thing complies with or follows the spec.

> 
> IMO, the best way to describe "location specifications", "code
> locations", and how the former leads to the latter depends on the
> answers to those two questions.
> 
> Thanks.
>