public inbox for libstdc++-cvs@sourceware.org help / color / mirror / Atom feed
From: Jason Merrill <jason@gcc.gnu.org> To: gcc-cvs@gcc.gnu.org, libstdc++-cvs@gcc.gnu.org Subject: [gcc/devel/c++-contracts] c++: comment tweaks Date: Tue, 1 Nov 2022 11:43:05 +0000 (GMT) [thread overview] Message-ID: <20221101114305.F1F463858017@sourceware.org> (raw) https://gcc.gnu.org/g:3a3e900ba634451f95443077a890f70003ec1fc6 commit 3a3e900ba634451f95443077a890f70003ec1fc6 Author: Jason Merrill <jason@redhat.com> Date: Tue Oct 25 15:55:20 2022 -0400 c++: comment tweaks Diff: --- gcc/cp/contracts.cc | 18 +++++++++++------- libstdc++-v3/include/experimental/contract | 2 +- 2 files changed, 12 insertions(+), 8 deletions(-) diff --git a/gcc/cp/contracts.cc b/gcc/cp/contracts.cc index 0a1626d6b68..2fa7023f76b 100644 --- a/gcc/cp/contracts.cc +++ b/gcc/cp/contracts.cc @@ -1,5 +1,5 @@ /* Definitions for C++ contract levels - Copyright (C) 2020 Free Software Foundation, Inc. + Copyright (C) 2020-2022 Free Software Foundation, Inc. Contributed by Jeff Chapman II (jchapman@lock3software.com) This file is part of GCC. @@ -131,20 +131,24 @@ along with GCC; see the file COPYING3. If not see With the idea being that multiple return statements could collapse the function epilogue after inlining the pre/post functions. clang is able to collapse common function epilogues, while gcc needs -O3 -Os combined. - We're already doing this "manually" for cdtors due to the way they're already - implemented, forcing DECL_CDTOR_NEEDS_LABLED_EXIT_P to be true when the - cdtor has active contracts. Directly laying the pre contracts down in the function body doesn't have many issues. The post contracts may need to be repeated multiple times, once - for each return, or a goto epilogue would need generated similarly to cdtors. + for each return, or a goto epilogue would need to be generated. For this initial implementation, generating function calls and letting later optimizations decide whether to inline and duplicate the actual checks or whether to collapse the shared epilogue was chosen. + For cdtors a post contract is implemented using a CLEANUP_STMT. + FIXME the compiler already handles sharing cleanup code on multiple exit - paths properly, this outlining isn't necessary if we represent the - postcondition as a cleanup (like I already did for dtors). */ + paths properly, so this outlining seems unnecessary if we represent the + postcondition as a cleanup for all functions. + + More helpful for optimization might be to make the contracts a wrapper + function (for non-variadic functions), that could be inlined into a + caller while preserving the call to the actual function? Either that or + turn a never-continue post contract into an assume in the caller. */ #include "config.h" #include "system.h" diff --git a/libstdc++-v3/include/experimental/contract b/libstdc++-v3/include/experimental/contract index a52c2a485cf..7a5f743665b 100644 --- a/libstdc++-v3/include/experimental/contract +++ b/libstdc++-v3/include/experimental/contract @@ -1,6 +1,6 @@ // Contracts support header for -*- C++ -*- -// Copyright (C) 1995-2018 Free Software Foundation, Inc. +// Copyright (C) 1995-2022 Free Software Foundation, Inc. // // This file is part of GCC. //
reply other threads:[~2022-11-01 11:43 UTC|newest] Thread overview: [no followups] expand[flat|nested] mbox.gz Atom feed
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=20221101114305.F1F463858017@sourceware.org \ --to=jason@gcc.gnu.org \ --cc=gcc-cvs@gcc.gnu.org \ --cc=libstdc++-cvs@gcc.gnu.org \ /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: linkBe 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).