From: Pierre-Marie de Rodat <derodat@adacore.com>
To: gcc-patches@gcc.gnu.org
Cc: Eric Botcazou <ebotcazou@adacore.com>
Subject: [Ada] Small improvement to System.Value_R.Scan_Raw_Real
Date: Fri, 27 Nov 2020 04:18:01 -0500 [thread overview]
Message-ID: <20201127091801.GA63309@adacore.com> (raw)
[-- Attachment #1: Type: text/plain, Size: 375 bytes --]
This makes System.Value_R.Scan_Raw_Real round the extra digit it returns
to the caller, so that it is precise to the last 'Aft digits for strings
that are the exact representation of a number, i.e. not already rounded.
Tested on x86_64-pc-linux-gnu, committed on trunk
gcc/ada/
* libgnat/s-valuer.adb (Scan_Decimal_Digits): Round Extra.
(Scan_Integral_Digits): Likewise.
[-- Attachment #2: patch.diff --]
[-- Type: text/x-diff, Size: 3944 bytes --]
diff --git a/gcc/ada/libgnat/s-valuer.adb b/gcc/ada/libgnat/s-valuer.adb
--- a/gcc/ada/libgnat/s-valuer.adb
+++ b/gcc/ada/libgnat/s-valuer.adb
@@ -144,6 +144,9 @@ package body System.Value_R is
-- Set to True if addition of a digit will cause Value to be superior
-- to Precision_Limit.
+ Precision_Limit_Just_Reached : Boolean := False;
+ -- Set to True if Precision_Limit_Reached was just set to True
+
Digit : Char_As_Digit;
-- The current digit
@@ -185,11 +188,24 @@ package body System.Value_R is
-- If precision limit has been reached, just ignore any remaining
-- digits for the computation of Value and Scale, but store the
- -- first in Extra. The scanning should continue only to assess the
- -- validity of the string.
+ -- first in Extra and use the second to round Extra. The scanning
+ -- should continue only to assess the validity of the string.
+
+ if Precision_Limit_Reached then
+ if Precision_Limit_Just_Reached then
+ if Digit >= Base / 2 then
+ if Extra = Base - 1 then
+ Extra := 0;
+ Value := Value + 1;
+ else
+ Extra := Extra + 1;
+ end if;
+ end if;
- if not Precision_Limit_Reached then
+ Precision_Limit_Just_Reached := False;
+ end if;
+ else
-- Trailing '0' digits are ignored until a non-zero digit is found
if Digit = 0 then
@@ -218,7 +234,7 @@ package body System.Value_R is
Temp := Value * Uns (Base) + Uns (Digit);
if Value <= Umax
- or else (Value <= UmaxB and then Temp <= Precision_Limit)
+ or else (Value <= UmaxB and then Temp <= Precision_Limit)
then
Value := Temp;
Scale := Scale - 1;
@@ -226,6 +242,7 @@ package body System.Value_R is
else
Extra := Digit;
Precision_Limit_Reached := True;
+ Precision_Limit_Just_Reached := True;
end if;
end if;
end if;
@@ -289,6 +306,9 @@ package body System.Value_R is
-- Set to True if addition of a digit will cause Value to be superior
-- to Precision_Limit.
+ Precision_Limit_Just_Reached : Boolean := False;
+ -- Set to True if Precision_Limit_Reached was just set to True
+
Digit : Char_As_Digit;
-- The current digit
@@ -324,12 +344,25 @@ package body System.Value_R is
-- If precision limit has been reached, just ignore any remaining
-- digits for the computation of Value, but update Scale and store
- -- the first in Extra. The scanning should continue only to assess
- -- the validity of the string.
+ -- the first in Extra and use the second to round Extra. The scanning
+ -- should continue only to assess the validity of the string.
if Precision_Limit_Reached then
Scale := Scale + 1;
+ if Precision_Limit_Just_Reached then
+ if Digit >= Base / 2 then
+ if Extra = Base - 1 then
+ Extra := 0;
+ Value := Value + 1;
+ else
+ Extra := Extra + 1;
+ end if;
+ end if;
+
+ Precision_Limit_Just_Reached := False;
+ end if;
+
else
Temp := Value * Uns (Base) + Uns (Digit);
@@ -341,6 +374,7 @@ package body System.Value_R is
else
Extra := Digit;
Precision_Limit_Reached := True;
+ Precision_Limit_Just_Reached := True;
Scale := Scale + 1;
end if;
end if;
reply other threads:[~2020-11-27 9:18 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=20201127091801.GA63309@adacore.com \
--to=derodat@adacore.com \
--cc=ebotcazou@adacore.com \
--cc=gcc-patches@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: 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).