forked from lwg/issues
-
Notifications
You must be signed in to change notification settings - Fork 28
Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
New issue from Matt Stephanson: "LWG 3809 changes behavior of some ex…
…isting std::subtract_with_carry_engine code"
- Loading branch information
Showing
1 changed file
with
201 additions
and
0 deletions.
There are no files selected for viewing
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,201 @@ | ||
| <?xml version='1.0' encoding='utf-8' standalone='no'?> | ||
| <!DOCTYPE issue SYSTEM "lwg-issue.dtd"> | ||
|
|
||
| <issue num="4014" status="New"> | ||
| <title>LWG 3809 changes behavior of some existing <tt>std::subtract_with_carry_engine</tt> code</title> | ||
| <section><sref ref="[rand.eng.sub]"/></section> | ||
| <submitter>Matt Stephanson</submitter> | ||
| <date>15 Nov 2023</date> | ||
| <priority>99</priority> | ||
|
|
||
| <discussion> | ||
| <p> | ||
| Issue <iref ref="3809"/> pointed out that <tt>subtract_with_carry_engine<T></tt> can be seeded with values | ||
| from a <tt>linear_congruential_engine<T, 40014u, 0u, 2147483563u></tt> object, which results in narrowing | ||
| when <tt>T</tt> is less than 32 bits. Part of the resolution was to modify the LCG seed sequence as follows: | ||
| </p> | ||
| <blockquote> | ||
| <pre>explicit subtract_with_carry_engine(result_type value);</pre> | ||
| <p>-7- <i>Effects</i>: | ||
| Sets the values of | ||
| <math> | ||
| <mrow> | ||
| <msub> <mi>X</mi> <mrow> <mo>-</mo> <mi>r</mi> </mrow> </msub> | ||
| <mo separator="true">,</mo> | ||
| <mo>…</mo> | ||
| <mo separator="true">,</mo> | ||
| <msub> <mi>X</mi> <mrow> <mo>-</mo> <mn>1</mn> </mrow> </msub> | ||
| </mrow> | ||
| </math>, | ||
| in that order, as specified below. If | ||
| <math> | ||
| <mrow> | ||
| <msub> <mi>X</mi> <mrow> <mo>-</mo> <mn>1</mn> </mrow> </msub> | ||
| </mrow> | ||
| </math> | ||
| is then <math> <mn>0</mn> </math>, | ||
| sets <math> <mi>c</mi> </math> to <math> <mn>1</mn> </math>; | ||
| otherwise sets <math> <mi>c</mi> </math> to <math> <mn>0</mn> </math>. | ||
| </p> | ||
| <p> | ||
| To set the values | ||
| <math> <mrow> <msub> <mi>X</mi> <mi>k</mi> </msub> </mrow> </math>, | ||
| first construct <tt>e</tt>, a <tt>linear_congruential_engine</tt> object, | ||
| as if by the following definition: | ||
| </p> | ||
| <blockquote><pre> | ||
| linear_congruential_engine<<del>result_type</del><ins>uint_least32_t</ins>, | ||
| 40014u,0u,2147483563u> e(value == 0u ? default_seed : value); | ||
| </pre></blockquote> | ||
| <p> | ||
| Then, to set each | ||
| <math> <mrow> <msub> <mi>X</mi> <mi>k</mi> </msub> </mrow> </math>, | ||
| obtain new values | ||
| <math> | ||
| <mrow> | ||
| <msub> <mi>z</mi> <mn>0</mn> </msub> | ||
| <mo separator="true">,</mo> | ||
| <mo>…</mo> | ||
| <mo separator="true">,</mo> | ||
| <msub> <mi>z</mi> <mrow> <mi>n</mi> <mo>-</mo> <mn>1</mn> </mrow> </msub> | ||
| </mrow> | ||
| </math> | ||
| from | ||
| <math> | ||
| <mrow> | ||
| <mi>n</mi> <mo>=</mo> | ||
| <mo form="prefix" stretchy="false">⌈</mo> | ||
| <mi>w</mi> <mo lspace="0em" rspace="0em">/</mo> <mn>32</mn> | ||
| <mo form="postfix" stretchy="false">⌉</mo> | ||
| </mrow> | ||
| </math> | ||
| successive invocations of <tt>e</tt>. | ||
| Set | ||
| <math> <mrow> <msub> <mi>X</mi> <mi>k</mi> </msub> </mrow> </math> | ||
| to | ||
| <math> | ||
| <mrow> | ||
| <mrow> | ||
| <mo fence="true" form="prefix">(</mo> | ||
| <msubsup> | ||
| <mo movablelimits="false">∑</mo> | ||
| <mrow> <mi>j</mi> <mo>=</mo> <mn>0</mn> </mrow> | ||
| <mrow> <mi>n</mi> <mo>-</mo> <mn>1</mn> </mrow> | ||
| </msubsup> | ||
| <msub> <mi>z</mi> <mi>j</mi> </msub> | ||
| <mo>∙</mo> | ||
| <msup> <mn>2</mn> <mrow> <mn>32</mn> <mi>j</mi> </mrow> </msup> | ||
| <mo fence="true" form="postfix">)</mo> | ||
| </mrow> | ||
| <mo lspace="0.2222em" rspace="0.2222em">mod</mo> | ||
| <mi>m</mi> | ||
| </mrow> | ||
| </math>. | ||
| </p> | ||
| </blockquote> | ||
| <p> | ||
| Inside <tt>linear_congruential_engine</tt>, the seed is reduced modulo 2147483563, so <tt>uint_least32_t</tt> | ||
| is fine from that point on. This resolution, however, forces <tt>value</tt>, the user-provided seed, to be | ||
| truncated from <tt>result_type</tt> to <tt>uint_least32_t</tt> before the reduction, which generally will | ||
| change the result. It also breaks the existing behavior that two seeds are equivalent if they're in the same | ||
| congruence class modulo the divisor. | ||
| </p> | ||
| </discussion> | ||
|
|
||
| <resolution> | ||
| <p> | ||
| This wording is relative to <paper num="N4964"/> <em>after the wording changes applied</em> by LWG <iref ref="3809"/>, | ||
| which had been accepted into the working paper during the Kona 2023-11 meeting. | ||
| </p> | ||
|
|
||
| <ol> | ||
|
|
||
| <li><p>Modify <sref ref="[rand.eng.sub]"/> as indicated:</p> | ||
|
|
||
| <blockquote> | ||
| <pre>explicit subtract_with_carry_engine(result_type value);</pre> | ||
| <p>-7- <i>Effects</i>: | ||
| Sets the values of | ||
| <math> | ||
| <mrow> | ||
| <msub> <mi>X</mi> <mrow> <mo>-</mo> <mi>r</mi> </mrow> </msub> | ||
| <mo separator="true">,</mo> | ||
| <mo>…</mo> | ||
| <mo separator="true">,</mo> | ||
| <msub> <mi>X</mi> <mrow> <mo>-</mo> <mn>1</mn> </mrow> </msub> | ||
| </mrow> | ||
| </math>, | ||
| in that order, as specified below. If | ||
| <math> | ||
| <mrow> | ||
| <msub> <mi>X</mi> <mrow> <mo>-</mo> <mn>1</mn> </mrow> </msub> | ||
| </mrow> | ||
| </math> | ||
| is then <math> <mn>0</mn> </math>, | ||
| sets <math> <mi>c</mi> </math> to <math> <mn>1</mn> </math>; | ||
| otherwise sets <math> <mi>c</mi> </math> to <math> <mn>0</mn> </math>. | ||
| </p> | ||
| <p> | ||
| To set the values | ||
| <math> <mrow> <msub> <mi>X</mi> <mi>k</mi> </msub> </mrow> </math>, | ||
| first construct <tt>e</tt>, a <tt>linear_congruential_engine</tt> object, | ||
| as if by the following definition: | ||
| </p> | ||
| <blockquote><pre> | ||
| linear_congruential_engine<uint_least32_t, | ||
| 40014u,0u,2147483563u> e(value == 0u ? default_seed : | ||
| <ins>static_cast<uint_least32_t>(</ins>value<ins> % 2147483563u)</ins>); | ||
| </pre></blockquote> | ||
| <p> | ||
| Then, to set each | ||
| <math> <mrow> <msub> <mi>X</mi> <mi>k</mi> </msub> </mrow> </math>, | ||
| obtain new values | ||
| <math> | ||
| <mrow> | ||
| <msub> <mi>z</mi> <mn>0</mn> </msub> | ||
| <mo separator="true">,</mo> | ||
| <mo>…</mo> | ||
| <mo separator="true">,</mo> | ||
| <msub> <mi>z</mi> <mrow> <mi>n</mi> <mo>-</mo> <mn>1</mn> </mrow> </msub> | ||
| </mrow> | ||
| </math> | ||
| from | ||
| <math> | ||
| <mrow> | ||
| <mi>n</mi> <mo>=</mo> | ||
| <mo form="prefix" stretchy="false">⌈</mo> | ||
| <mi>w</mi> <mo lspace="0em" rspace="0em">/</mo> <mn>32</mn> | ||
| <mo form="postfix" stretchy="false">⌉</mo> | ||
| </mrow> | ||
| </math> | ||
| successive invocations of <tt>e</tt>. | ||
| Set | ||
| <math> <mrow> <msub> <mi>X</mi> <mi>k</mi> </msub> </mrow> </math> | ||
| to | ||
| <math> | ||
| <mrow> | ||
| <mrow> | ||
| <mo fence="true" form="prefix">(</mo> | ||
| <msubsup> | ||
| <mo movablelimits="false">∑</mo> | ||
| <mrow> <mi>j</mi> <mo>=</mo> <mn>0</mn> </mrow> | ||
| <mrow> <mi>n</mi> <mo>-</mo> <mn>1</mn> </mrow> | ||
| </msubsup> | ||
| <msub> <mi>z</mi> <mi>j</mi> </msub> | ||
| <mo>∙</mo> | ||
| <msup> <mn>2</mn> <mrow> <mn>32</mn> <mi>j</mi> </mrow> </msup> | ||
| <mo fence="true" form="postfix">)</mo> | ||
| </mrow> | ||
| <mo lspace="0.2222em" rspace="0.2222em">mod</mo> | ||
| <mi>m</mi> | ||
| </mrow> | ||
| </math>. | ||
| </p> | ||
| </blockquote> | ||
|
|
||
| </li> | ||
|
|
||
| </ol> | ||
| </resolution> | ||
|
|
||
| </issue> |