<p>it performs exactly one computation</p><p>x = x * 1.064f<br />y = x / 65535</p><p>we don&#39;t know why.</p>

<p><span class="h-card" translate="no"><a href="https://mastodon.social/@whitequark" class="u-url mention">@<span>whitequark</span></a></span> <br />As included in Borland/Turbo C/C++</p>

<p><span class="h-card" translate="no"><a href="https://not.acu.lt/@ignaloidas" class="u-url mention">@<span>ignaloidas</span></a></span> no</p>

@whitequark@mastodon.social soft-float is for an embedded lisp implementation that uses NaN boxing

<p><span class="h-card" translate="no"><a href="https://rants.au/@sabik" class="u-url mention">@<span>sabik</span></a></span> ah, the CrackSmokingCompiler(float) routine?</p>

<p><span class="h-card" translate="no"><a href="https://mastodon.social/@whitequark" class="u-url mention">@<span>whitequark</span></a></span> &quot;i swear i didn&#39;t write %f - it was the gremlins!&quot;</p>

<p><span class="h-card" translate="no"><a href="https://mastodon.social/@whitequark" class="u-url mention">@<span>whitequark</span></a></span> my wild guess would be when you want to read FP registers of hardware float and need the values decoded</p>

<p>&quot;do you ever reverse engineer a JTAG adapter firmware and find out it uses a soft-float library?&quot; &quot;no, you?&quot; &quot;yes.&quot; &quot;what&#39;s it for?&quot; &quot;i am... not sure&quot;</p>

<p><span class="h-card"><a href="https://mastodon.social/@whitequark" class="u-url mention">@<span>whitequark</span></a></span> “everybody [thinks they understand lifetime rules] until they get punched in the face” – Bjarne Tyson</p>