[06/21] diff suppression: Fix handling of change kinds.

  When parsing suppression specifications, libabigail attempts to detect
and so ignore useless suppressions. This is an optimisation to avoid
the cost of testing against a suppression that will do nothing.

Unfortunately, with the way the parser currently works, these checks
have to be against an exhaustive list of fields that matter, rather
than by listing the fields that don't (label). They are fragile in the
face of changes that add new fields.

Two of the short-cut checks are in fact buggy, missing out the
change_kind field. One of the checks also risks a null pointer
dereference as it doesn't actually trigger a return from the function.

This patch eliminates (rather than fixing up) this short-cutting on
the grounds that it buys very litte and is a maintenance burden. In a
future version, it could be reinstated with logic that has global
knowledge of which fields are present and which fields have no
suppression semantics and perhaps emitting warning message to the user
if they have supplied a useless specification.

The patch also corrects 4 affected test cases to reflect that
suppression is actually happening (according to change_kind).

	* src/abg-suppression.cc (read_type_suppression): Remove
	short-circuiting of useless suppressions.
	(read_function_suppression): Ditto.
	(read_variable_suppression: Ditto.
	(read_file_suppression): Ditto.
	tests/data/test-diff-suppr/test15-suppr-added-fn-report-5.txt:
	Fix test - something is actually suppressed.
	* tests/data/test-diff-suppr/test16-suppr-removed-fn-report-5.txt:
	Ditto.
	* tests/data/test-diff-suppr/test17-suppr-added-var-report-5.txt:
	Ditto.
	* tests/data/test-diff-suppr/test18-suppr-removed-var-report-5.txt:
	Ditto.

Signed-off-by: Giuliano Procida <gprocida@google.com>
---
 src/abg-suppression.cc                        | 89 +++++--------------
 .../test15-suppr-added-fn-report-5.txt        |  6 +-
 .../test16-suppr-removed-fn-report-5.txt      | 15 +---
 .../test17-suppr-added-var-report-5.txt       | 15 +---
 .../test18-suppr-removed-var-report-5.txt     | 15 +---
 5 files changed, 25 insertions(+), 115 deletions(-)

Message ID	20200423154441.170531-7-gprocida@google.com
State	Superseded
Headers	From: gprocida@google.com (Giuliano Procida) Date: Thu, 23 Apr 2020 16:44:26 +0100 Subject: [PATCH 06/21] diff suppression: Fix handling of change kinds. In-Reply-To: <20200423154441.170531-1-gprocida@google.com> References: <20200423154441.170531-1-gprocida@google.com> Message-ID: <20200423154441.170531-7-gprocida@google.com>
Series	Simplify regex and suppression parsing. \| [00/21] Simplify regex and suppression parsing. [01/21] Move regex definitions to own files. [02/21] Move libxml bits out of abg-sptr-utils.h. [03/21] Simplify generation of symbol whitelist regex. [04/21] Escape names used in symbol whitelisting regex. [05/21] abg-suppression.cc: More uniform variable naming. [06/21] diff suppression: Fix handling of change kinds. [07/21] Add POSIX regex wrapper functions. [08/21] Use regex::compile wrapper instead of regcomp. [09/21] Use regex::match wrapper instead of regexec. [10/21] Refactor read_parameter_spec_from_string logic. [11/21] Compile suppression regexes earlier. [12/21] Reduce direct access to suppression priv_ members. [13/21] Move match methods from priv to suppression_base. [14/21] Remove suppression types' priv class methods. [15/21] abg-suppression.cc: More consistent regex matching. [16/21] abg-tools-utils.cc: Assert generated regexes OK. [17/21] Refactor suppression property string parsing. [18/21] Refactor suppression property regex parsing. [19/21] Warn if user-supplied regexes fail to compile. [20/21] Default construct suppression types. [21/21] Remove unused suppression type priv constructors.

[06/21] diff suppression: Fix handling of change kinds.

Commit Message

Patch