From mboxrd@z Thu Jan 1 00:00:00 1970 X-Msuck: nntp://news.gmane.org/gmane.linux.lib.musl.general/10434 Path: news.gmane.org!.POSTED!not-for-mail From: Rich Felker Newsgroups: gmane.linux.lib.musl.general Subject: Re: [PATCH] add missing *_unlocked and wcsftime_l prototypes to wchar.h Date: Fri, 9 Sep 2016 21:01:15 -0400 Message-ID: <20160910010115.GV15995@brightrain.aerifal.cx> References: <20160906210939.32637-1-dsabogalcc@gmail.com> <20160906210939.32637-2-dsabogalcc@gmail.com> <20160909230241.GT15995@brightrain.aerifal.cx> Reply-To: musl@lists.openwall.com NNTP-Posting-Host: blaine.gmane.org Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Trace: blaine.gmane.org 1473469293 26992 195.159.176.226 (10 Sep 2016 01:01:33 GMT) X-Complaints-To: usenet@blaine.gmane.org NNTP-Posting-Date: Sat, 10 Sep 2016 01:01:33 +0000 (UTC) User-Agent: Mutt/1.5.21 (2010-09-15) To: musl@lists.openwall.com Original-X-From: musl-return-10447-gllmg-musl=m.gmane.org@lists.openwall.com Sat Sep 10 03:01:30 2016 Return-path: Envelope-to: gllmg-musl@m.gmane.org Original-Received: from mother.openwall.net ([195.42.179.200]) by blaine.gmane.org with smtp (Exim 4.84_2) (envelope-from ) id 1biWfh-0006Sl-F6 for gllmg-musl@m.gmane.org; Sat, 10 Sep 2016 03:01:29 +0200 Original-Received: (qmail 21843 invoked by uid 550); 10 Sep 2016 01:01:28 -0000 Mailing-List: contact musl-help@lists.openwall.com; run by ezmlm Precedence: bulk List-Post: List-Help: List-Unsubscribe: List-Subscribe: List-ID: Original-Received: (qmail 21825 invoked from network); 10 Sep 2016 01:01:28 -0000 Content-Disposition: inline In-Reply-To: Original-Sender: Rich Felker Xref: news.gmane.org gmane.linux.lib.musl.general:10434 Archived-At: On Fri, Sep 09, 2016 at 08:42:45PM -0400, Daniel Sabogal wrote: > On Fri, Sep 9, 2016 at 7:02 PM, Rich Felker wrote: > > On Tue, Sep 06, 2016 at 05:09:39PM -0400, Daniel Sabogal wrote: > >> these functions had been implemented, but prototypes were not made available > >> --- > >> include/wchar.h | 12 ++++++++++++ > >> src/stdio/getwchar.c | 2 -- > >> src/stdio/getwchar_unlocked.c | 8 ++++++++ > >> src/stdio/putwchar.c | 2 -- > >> src/stdio/putwchar_unlocked.c | 8 ++++++++ > >> 5 files changed, 28 insertions(+), 4 deletions(-) > >> create mode 100644 src/stdio/getwchar_unlocked.c > >> create mode 100644 src/stdio/putwchar_unlocked.c > >> > >> diff --git a/include/wchar.h b/include/wchar.h > >> index 0167dce..58818f6 100644 > >> --- a/include/wchar.h > >> +++ b/include/wchar.h > >> @@ -136,6 +136,18 @@ size_t wcsftime (wchar_t *__restrict, size_t, const wchar_t *__restrict, const s > >> > >> #undef iswdigit > >> > >> +#if defined(_GNU_SOURCE) > >> +wint_t fgetwc_unlocked (FILE *); > >> +wint_t getwc_unlocked (FILE *); > >> +wint_t getwchar_unlocked (void); > >> +wint_t fputwc_unlocked (wchar_t, FILE *); > >> +wint_t putwc_unlocked (wchar_t, FILE *); > >> +wint_t putwchar_unlocked (wchar_t); > >> +wchar_t *fgetws_unlocked (wchar_t *__restrict, int, FILE *__restrict); > >> +int fputws_unlocked (const wchar_t *__restrict, FILE *__restrict); > >> +size_t wcsftime_l (wchar_t *__restrict, size_t, const wchar_t *__restrict, const struct tm *__restrict, locale_t); > >> +#endif > > > > I suspect BSDs also had these so they should possibly be exposed under > > _GNU_SOURCE || _BSD_SOURCE. Thoughts? > > Glibc doesn't provide any of these with _BSD_SOURCE. > FreeBSD and NetBSD provide wcsftime_l, but doesn't seem to have the > *_unlocked variants. > OpenBSD doesn't seem to provide either. > > Perhaps wcsftime_l should be moved under _GNU_SOURCE || _BSD_SOURCE. OK. > >> diff --git a/src/stdio/getwchar.c b/src/stdio/getwchar.c > >> index bd89e0e..77a9dc1 100644 > >> --- a/src/stdio/getwchar.c > >> +++ b/src/stdio/getwchar.c > >> @@ -5,5 +5,3 @@ wint_t getwchar(void) > >> { > >> return fgetwc(stdin); > >> } > >> - > >> -weak_alias(getwchar, getwchar_unlocked); > >> diff --git a/src/stdio/getwchar_unlocked.c b/src/stdio/getwchar_unlocked.c > >> new file mode 100644 > >> index 0000000..1d00567 > >> --- /dev/null > >> +++ b/src/stdio/getwchar_unlocked.c > >> @@ -0,0 +1,8 @@ > >> +#define _GNU_SOURCE > >> +#include "stdio_impl.h" > >> +#include > >> + > >> +wint_t getwchar_unlocked(void) > >> +{ > >> + return fgetwc_unlocked(stdin); > >> +} > > > > What is the motivation for replacing the aliases with wrappers? This > > does not seem like an improvement. > > Correct me if I'm wrong, but by using weak_alias(getwchar, getwchar_unlocked) > instead of the wrapper, wouldn't calls to getwchar_unlocked() really just invoke > getwchar() or fgetwc(stdin) instead of the intended fgetwc_unlocked(stdin)? As far as I can tell, the existence of a separate fgetwc_unlocked in musl is just a historical accident due to its use in implementing fgetws. It goes back to the initial git check-in and I didn't try to dig up any older history. In any case, trying to improve performance of these functions is orthogonal to exposing declarations for them, so if done at all it should be a separate commit. But I'd really like to avoid adding "junk" functions where aliases already work, and get rid of some we already have. Rich