From: Paul E Murphy <murphyp@linux.ibm.com>
To: "Gabriel F. T. Gomes" <gabriel@inconstante.net.br>,
libc-alpha@sourceware.org
Subject: Re: [PATCH v2 23/30] Refactor *cvt functions implementation (5/5)
Date: Thu, 14 Nov 2019 23:12:00 -0000 [thread overview]
Message-ID: <c7c2a8c0-521c-2ffb-0a22-d322044d8715@linux.ibm.com> (raw)
In-Reply-To: <20191025153410.15405-24-gabriel@inconstante.net.br>
On 10/25/19 10:34 AM, Gabriel F. T. Gomes wrote:
> From: "Gabriel F. T. Gomes" <gabrielftg@linux.ibm.com>
>
> Changes since v1:
>
> - Adapted after patch v1 19/31 (remove hidden_* uses) was dropped.
> - Fixed unintended removal of cvt* symbols on alpha and s390x
> (now tested with build-many-glibcs.py).
>
> -- 8< --
> This patch is to be squashed with the other n/5 refactoring patches.
>
> This patch replaces the use of the APPEND macro with one new macro being
> defined for each of the cvt* functions. This makes it easier to define
> functions names for IEEE long double on powerpc64le, e.g. __ecvtieee128.
I think the non-trivial changes here to fixup the shuffling and the
preceding four patches should be squashed and resent as a single patch
to the list. I am guessing they should squash into a somewhat readable
patch self contained patch.
The slightly more substantial changes needed to support the new ldbl
would be more easily reviewed as a separate patch (e.g replacing the
__APPEND macro and reworking cvt_symbol)
> diff --git a/misc/efgcvt.c b/misc/efgcvt.c
> index 81ac60415e..93a0269d5c 100644
> --- a/misc/efgcvt.c
> +++ b/misc/efgcvt.c
> @@ -16,19 +16,27 @@
> License along with the GNU C Library; if not, see
> <http://www.gnu.org/licenses/>. */
>
> -#include "efgcvt.c"
I see some noise in the refactoring :). (and in efgcvt_r.c),
> diff --git a/misc/efgcvt_r-template.c b/misc/efgcvt_r-template.c
...
> -#define FLOOR APPEND(floor, FLOAT_NAME_EXT)
> -#define FABS APPEND(fabs, FLOAT_NAME_EXT)
> -#define LOG10 APPEND(log10, FLOAT_NAME_EXT)
> -#define EXP APPEND(exp, FLOAT_NAME_EXT)
These macros don't even appear to be used... OK.
> diff --git a/misc/efgcvt_r.c b/misc/efgcvt_r.c
> index aa0eb8ca43..5b48ca5b9b 100644
> --- a/misc/efgcvt_r.c
> +++ b/misc/efgcvt_r.c
> @@ -16,22 +16,24 @@
> License along with the GNU C Library; if not, see
> <http://www.gnu.org/licenses/>. */
>
> -#include "efgcvt_r.c"
> +#define ECVT_R ecvt_r
> +#define FCVT_R fcvt_r
> +#define __ECVT_R __ecvt_r
> +#define __FCVT_R __fcvt_r
> +#include <efgcvt-dbl-macros.h>
> +#include <efgcvt_r-template.c>
>
> #if LONG_DOUBLE_COMPAT (libc, GLIBC_2_0)
> -# define cvt_symbol(symbol) \
> - cvt_symbol_1 (libc, __APPEND (FUNC_PREFIX, symbol), \
> - APPEND (q, symbol), GLIBC_2_0); \
> - weak_alias (__APPEND (FUNC_PREFIX, symbol), APPEND (FUNC_PREFIX, symbol))
> +# define cvt_symbol(local, symbol) \
> + cvt_symbol_1 (libc, local, APPEND (q, symbol), GLIBC_2_0); \
> + weak_alias (local, symbol)
> # define cvt_symbol_1(lib, local, symbol, version) \
> libc_hidden_def (local) \
> compat_symbol (lib, local, symbol, version)
> #else
> -# define cvt_symbol(symbol) \
> - cvt_symbol_1 (__APPEND (FUNC_PREFIX, symbol), APPEND (FUNC_PREFIX, symbol))
> -# define cvt_symbol_1(local, symbol) \
> +# define cvt_symbol(local, symbol) \
> libc_hidden_def (local) \
> weak_alias (local, symbol)
> #endif
> -cvt_symbol (fcvt_r);
> -cvt_symbol (ecvt_r);
> +cvt_symbol (__fcvt_r, fcvt_r);
> +cvt_symbol (__ecvt_r, ecvt_r);
> diff --git a/misc/qefgcvt.c b/misc/qefgcvt.c
> index ea48c6b48a..903bf93aa4 100644
> --- a/misc/qefgcvt.c
> +++ b/misc/qefgcvt.c
> @@ -16,18 +16,24 @@
> License along with the GNU C Library; if not, see
> <https://www.gnu.org/licenses/>. */
>
> -#include "efgcvt.c"
> +#define ECVT qecvt
> +#define FCVT qfcvt
> +#define GCVT qgcvt
> +#define __ECVT __qecvt
> +#define __FCVT __qfcvt
> +#define __GCVT __qgcvt
> +#define __ECVT_R __qecvt_r
> +#define __FCVT_R __qfcvt_r
> +#include <efgcvt-ldbl-macros.h>
> +#include <efgcvt-template.c>
>
> #if LONG_DOUBLE_COMPAT (libc, GLIBC_2_0)
> -# define cvt_symbol(symbol) \
> - cvt_symbol_1 (libc, __APPEND (FUNC_PREFIX, symbol), \
> - APPEND (FUNC_PREFIX, symbol), GLIBC_2_4)
> -# define cvt_symbol_1(lib, local, symbol, version) \
> - versioned_symbol (lib, local, symbol, version)
> +# define cvt_symbol(local, symbol) \
> + versioned_symbol (libc, local, symbol, GLIBC_2_4)
> #else
> -# define cvt_symbol(symbol) \
> - strong_alias (__APPEND (FUNC_PREFIX, symbol), APPEND (FUNC_PREFIX, symbol))
> +# define cvt_symbol(local, symbol) \
> + strong_alias (local, symbol)
> #endif
> -cvt_symbol(fcvt);
> -cvt_symbol(ecvt);
> -cvt_symbol(gcvt);
> +cvt_symbol (__qfcvt, qfcvt);
> +cvt_symbol (__qecvt, qecvt);
> +cvt_symbol (__qgcvt, qgcvt);
> diff --git a/misc/qefgcvt_r.c b/misc/qefgcvt_r.c
> index d2f36e6d2a..b790bf7837 100644
> --- a/misc/qefgcvt_r.c
> +++ b/misc/qefgcvt_r.c
> @@ -17,21 +17,21 @@
> License along with the GNU C Library; if not, see
> <https://www.gnu.org/licenses/>. */
>
> -#include "efgcvt_r.c"
> +#define ECVT_R qecvt_r
> +#define FCVT_R qfcvt_r
> +#define __ECVT_R __qecvt_r
> +#define __FCVT_R __qfcvt_r
> +#include <efgcvt-ldbl-macros.h>
> +#include <efgcvt_r-template.c>
>
> #if LONG_DOUBLE_COMPAT (libc, GLIBC_2_0)
> -# define cvt_symbol(symbol) \
> - cvt_symbol_1 (libc, __APPEND (FUNC_PREFIX, symbol), \
> - APPEND (FUNC_PREFIX, symbol), GLIBC_2_4)
> -# define cvt_symbol_1(lib, local, symbol, version) \
> +# define cvt_symbol(local, symbol) \
> libc_hidden_def (local) \
> - versioned_symbol (lib, local, symbol, version)
> + versioned_symbol (libc, local, symbol, GLIBC_2_4)
This isn't a bug with the patch, but should the version tested in
LONG_DOUBLE_COMPAT match that passed into versioned_symbol?
> #else
> -# define cvt_symbol(symbol) \
> - cvt_symbol_1 (__APPEND (FUNC_PREFIX, symbol), APPEND (FUNC_PREFIX, symbol))
> -# define cvt_symbol_1(local, symbol) \
> +# define cvt_symbol(local, symbol) \
> libc_hidden_def (local) \
> weak_alias (local, symbol)
> #endif
The two non-trivial versioning macros get duplicated in two places. Is
it possible to unify the four very similar instances into a single
shared macro?
> -cvt_symbol(fcvt_r);
> -cvt_symbol(ecvt_r);
> +cvt_symbol (__qfcvt_r, qfcvt_r);
> +cvt_symbol (__qecvt_r, qecvt_r);
>
next prev parent reply other threads:[~2019-11-14 23:12 UTC|newest]
Thread overview: 72+ messages / expand[flat|nested] mbox.gz Atom feed top
2019-10-25 15:34 [PATCH v2 00/30] Add IEEE long double <-> string functions for powerpc64le Gabriel F. T. Gomes
2019-10-25 15:34 ` [PATCH v2 01/30] ldbl-128ibm-compat: Add regular character printing functions Gabriel F. T. Gomes
2019-10-29 15:42 ` Paul E Murphy
2019-10-30 16:10 ` Gabriel F. T. Gomes
2019-10-25 15:34 ` [PATCH v2 03/30] ldbl-128ibm-compat: Add regular character, fortified " Gabriel F. T. Gomes
2019-11-04 15:43 ` Paul E Murphy
2019-11-05 19:37 ` Gabriel F. T. Gomes
2019-11-08 16:44 ` Paul E Murphy
2019-10-25 15:34 ` [PATCH v2 02/30] ldbl-128ibm-compat: Add wide character " Gabriel F. T. Gomes
2019-11-04 15:08 ` Paul E Murphy
2019-11-04 18:10 ` Gabriel F. T. Gomes
2019-11-04 20:11 ` Paul E Murphy
2019-10-25 15:35 ` [PATCH v2 11/30] ldbl-128ibm-compat: Add error.h functions Gabriel F. T. Gomes
2019-11-08 22:21 ` Paul E Murphy
2019-10-25 15:35 ` [PATCH v2 13/30] ldbl-128ibm-compat: Add ISO C99 versions of scanf functions Gabriel F. T. Gomes
2019-11-08 23:26 ` Paul E Murphy
2019-11-21 13:44 ` Gabriel F. T. Gomes
2019-11-22 22:49 ` Gabriel F. T. Gomes
2019-11-27 19:13 ` Gabriel F. T. Gomes
2019-10-25 15:35 ` [PATCH v2 06/30] ldbl-128ibm-compat: Test positional arguments Gabriel F. T. Gomes
2019-11-04 16:55 ` Paul E Murphy
2019-11-06 17:28 ` Gabriel F. T. Gomes
2019-10-25 15:35 ` [PATCH v2 15/30] ldbl-128ibm-compat: Add syslog functions Gabriel F. T. Gomes
2019-11-14 17:47 ` Paul E Murphy
2019-11-14 18:25 ` Florian Weimer
2019-11-27 19:12 ` Gabriel F. T. Gomes
2019-10-25 15:35 ` [PATCH v2 08/30] ldbl-128ibm-compat: Add wide character scanning functions Gabriel F. T. Gomes
2019-11-04 17:47 ` Paul E Murphy
2019-11-11 14:33 ` Gabriel F. T. Gomes
2019-11-22 22:34 ` Gabriel F. T. Gomes
2019-10-25 15:35 ` [PATCH v2 14/30] ldbl-128ibm-compat: Add obstack printing functions Gabriel F. T. Gomes
2019-11-08 23:42 ` Paul E Murphy
2019-11-27 19:11 ` Gabriel F. T. Gomes
2019-10-25 15:35 ` [PATCH v2 09/30] ldbl-128ibm-compat: Add argp_error and argp_failure Gabriel F. T. Gomes
2019-11-08 19:58 ` Paul E Murphy
2019-10-25 15:35 ` [PATCH v2 10/30] ldbl-128ibm-compat: Add err.h functions Gabriel F. T. Gomes
2019-11-08 20:22 ` Paul E Murphy
2019-11-13 14:24 ` Gabriel F. T. Gomes
2019-11-27 19:11 ` Gabriel F. T. Gomes
2019-10-25 15:35 ` [PATCH v2 05/30] ldbl-128ibm-compat: Test double values Gabriel F. T. Gomes
2019-11-04 16:16 ` Paul E Murphy
2019-10-25 15:35 ` [PATCH v2 07/30] ldbl-128ibm-compat: Add regular character scanning functions Gabriel F. T. Gomes
2019-11-14 21:36 ` Paul E Murphy
2019-10-25 15:36 ` [PATCH v2 18/30] ldbl-128ibm-compat: Add tests for strfroml, strtold, and wcstold Gabriel F. T. Gomes
2019-11-14 18:26 ` Paul E Murphy
2019-11-27 17:47 ` Gabriel F. T. Gomes
2019-12-03 16:58 ` Gabriel F. T. Gomes
2019-10-25 15:36 ` [PATCH v2 20/30] Refactor *cvt functions implementation (2/5) Gabriel F. T. Gomes
2019-10-25 15:36 ` [PATCH v2 12/30] ldbl-128ibm-compat: Reuse tests for err.h and error.h functions Gabriel F. T. Gomes
2019-11-08 22:25 ` Paul E Murphy
2019-10-25 15:36 ` [PATCH v2 21/30] Refactor *cvt functions implementation (3/5) Gabriel F. T. Gomes
2019-10-25 15:36 ` [PATCH v2 23/30] Refactor *cvt functions implementation (5/5) Gabriel F. T. Gomes
2019-11-14 23:12 ` Paul E Murphy [this message]
2019-11-27 17:42 ` Gabriel F. T. Gomes
2019-10-25 15:36 ` [PATCH v2 22/30] Refactor *cvt functions implementation (4/5) Gabriel F. T. Gomes
2019-10-25 15:36 ` [PATCH v2 16/30] ldbl-128ibm-compat: Add strfmon_l with IEEE long double format Gabriel F. T. Gomes
2019-11-14 18:03 ` Paul E Murphy
2019-11-27 17:43 ` Gabriel F. T. Gomes
2019-12-03 16:57 ` Gabriel F. T. Gomes
2019-10-25 15:36 ` [PATCH v2 19/30] Refactor *cvt functions implementation (1/5) Gabriel F. T. Gomes
2019-10-25 15:36 ` [PATCH v2 04/30] ldbl-128ibm-compat: Add wide character, fortified printing functions Gabriel F. T. Gomes
2019-11-04 16:10 ` Paul E Murphy
2019-10-25 15:37 ` [PATCH v2 30/30] RFC: powerpc64le: Enable support for IEEE long double Gabriel F. T. Gomes
2019-10-25 15:37 ` [PATCH v2 26/30] ldbl-128ibm-compat: Compiler flags for stdio functions Gabriel F. T. Gomes
2019-10-25 15:37 ` [PATCH v2 29/30] powerpc64le: Require a compiler with -mno-gnu-attribute Gabriel F. T. Gomes
2019-10-25 15:37 ` [PATCH v2 27/30] Avoid compat symbols for totalorder in powerpc64le IEEE long double Gabriel F. T. Gomes
2019-10-25 15:37 ` [PATCH v2 24/30] ldbl-128ibm-compat: Add *cvt functions Gabriel F. T. Gomes
2019-10-25 15:37 ` [PATCH v2 25/30] Do not redirect calls to __GI_* symbols, when redirecting to *ieee128 Gabriel F. T. Gomes
2019-10-25 15:37 ` [PATCH v2 28/30] ldbl-128ibm-compat: Do not mix -mabi=*longdouble and -mlong-double-128 Gabriel F. T. Gomes
2019-10-25 15:37 ` [PATCH v2 17/30] ldbl-128ibm-compat: Add tests for strfmon and strfmon_l Gabriel F. T. Gomes
2019-11-14 18:11 ` Paul E Murphy
2019-12-03 16:58 ` Gabriel F. T. Gomes
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=c7c2a8c0-521c-2ffb-0a22-d322044d8715@linux.ibm.com \
--to=murphyp@linux.ibm.com \
--cc=gabriel@inconstante.net.br \
--cc=libc-alpha@sourceware.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: 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).