From: Giuliano Procida <gprocida@google.com>
To: Dodji Seketeli <dodji@seketeli.org>
Cc: libabigail@sourceware.org, kernel-team@android.com,
"Matthias Männich" <maennich@google.com>
Subject: Re: [PATCH 3/7] Better suppression section parsing delegation.
Date: Fri, 2 Oct 2020 08:20:16 +0100 [thread overview]
Message-ID: <CAGvU0HmcU2aVA12eoWhkwv+WQxUztk7vGwj_apzi1XBos_ya3Q@mail.gmail.com> (raw)
In-Reply-To: <87o8lmos73.fsf@seketeli.org>
Hi Dodji.
This is the start of a sequence of much longer changes to the suppression
code that make the parser into a real entity in its own right.
https://github.com/myxoid/libabigail/commits/suppression-parsing-error-handling
gets to the point of having a table-driven parser where error status is
threaded to the top-level (and is followed by further series).
The redundancy of error checking you suggest could be kept but would be
somewhat out of place in the table-driven parser.
If you can see a way that these changes might be useful, perhaps with small
or large changes, let me know. Otherwise it may be time to abandon the
effort.
Regards,
Giuliano.
On Thu, 1 Oct 2020 at 16:56, Dodji Seketeli <dodji@seketeli.org> wrote:
> Giuliano Procida <gprocida@google.com> a écrit:
>
> > The function read_suppressions hands the same ini config section to
> > each of four parsing functions, until one succeeds. Each of those in
> > turn has an early exit if the name of the section doesn't correspond.
>
> Another way of seeing it is that the section is read either as type,
> function, variable or file suppression.
>
> > This can be simplified.
>
> It's not obvious to me that the result you are proposing below is
> "simpler" per se. It seems to me that this just boils down to a matter
> of taste.
>
> >
> > This patch moves the name checking into read_suppressions so that
> > exactly one parsing function is called per section.
> >
> > * src/abg-suppression.cc (read_suppressions): Call appropriate
> > read_foo_suppression function based on section name.
> > (read_type_suppression): Remove section name check.
> > (read_function_suppression): Ditto.
> > (read_variable_suppression): Ditto.
> > (read_file_suppression): Ditto.
> >
> > Signed-off-by: Giuliano Procida <gprocida@google.com>
> > ---
> > src/abg-suppression.cc | 34 +++++++++++++++-------------------
> > 1 file changed, 15 insertions(+), 19 deletions(-)
> >
> > diff --git a/src/abg-suppression.cc b/src/abg-suppression.cc
> > index ae7cc95ce..682bb8742 100644
> > --- a/src/abg-suppression.cc
> > +++ b/src/abg-suppression.cc
> > @@ -372,17 +372,25 @@ static void
> > read_suppressions(const ini::config& config,
> > suppressions_type& suppressions)
> > {
> > - suppression_sptr s;
> > for (ini::config::sections_type::const_iterator i =
> > config.get_sections().begin();
> > i != config.get_sections().end();
> > ++i)
> > - if ((s = read_type_suppression(**i))
> > - || (s = read_function_suppression(**i))
> > - || (s = read_variable_suppression(**i))
> > - || (s = read_file_suppression(**i)))
> > - suppressions.push_back(s);
> > -
>
> As I was saying earlier, the above can be understood as "the section is
> read either as a type, function, variable or file suppression."
>
> > + {
> > + const ini::config::section_sptr& section = *i;
> > + const std::string& name = section->get_name();
> > + suppression_sptr s;
> > + if (name == "suppress_type")
> > + s = read_type_suppression(*section);
> > + else if (name == "suppress_function")
> > + s = read_function_suppression(*section);
> > + else if (name == "suppress_variable")
> > + s = read_variable_suppression(*section);
> > + else if (name == "suppress_file")
> > + s = read_file_suppression(*section);
> > + if (s)
> > + suppressions.push_back(s);
> > + }
> > }
>
> I don't see how this is simpler than the initial code, quite frankly.
>
> >
> > /// Read suppressions specifications from an input stream.
> > @@ -1564,9 +1572,6 @@ read_type_suppression(const ini::config::section&
> section)
> > {
> > type_suppression_sptr result;
> >
> > - if (section.get_name() != "suppress_type")
> > - return result;
> > -
>
> Actually, I'd prefer making this function return early when wrongly
> called on a section that is not a type suppression. So I think removing
> this check actually makes the function less robust.
>
> I have a similar comment for the changes to the other parsing
> functions below this one.
>
> [...]
>
> Cheers,
>
> --
> Dodji
>
next prev parent reply other threads:[~2020-10-02 7:20 UTC|newest]
Thread overview: 14+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-08-17 9:38 [PATCH 0/7] Suppression parsing - preparatory work Giuliano Procida
2020-08-17 9:38 ` [PATCH 1/7] Add missing newlines to end of test files Giuliano Procida
2020-10-01 14:36 ` Dodji Seketeli
2020-08-17 9:38 ` [PATCH 2/7] Fix two wrongs in test suppression regex Giuliano Procida
2020-10-01 14:42 ` Dodji Seketeli
2020-08-17 9:38 ` [PATCH 3/7] Better suppression section parsing delegation Giuliano Procida
2020-10-01 15:56 ` Dodji Seketeli
2020-10-02 7:20 ` Giuliano Procida [this message]
2020-08-17 9:38 ` [PATCH 4/7] Add read_*_suppression success/failure plumbing Giuliano Procida
2020-08-17 9:38 ` [PATCH 5/7] Add error handling to read_suppressions Giuliano Procida
2020-08-17 9:38 ` [PATCH 6/7] Default construct suppression types Giuliano Procida
2020-08-17 9:38 ` [PATCH 7/7] Refresh getter/setter comments Giuliano Procida
2020-10-02 7:25 ` [PATCH 0/7] Suppression parsing - preparatory work Dodji Seketeli
2020-10-06 20:19 ` Giuliano Procida
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=CAGvU0HmcU2aVA12eoWhkwv+WQxUztk7vGwj_apzi1XBos_ya3Q@mail.gmail.com \
--to=gprocida@google.com \
--cc=dodji@seketeli.org \
--cc=kernel-team@android.com \
--cc=libabigail@sourceware.org \
--cc=maennich@google.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).