Today I reached a pain threshold where I finally started using "typedef" in my code for a pointer type. Recall Linus Torvalds famous advice against doing this:<p>http://lkml.indiana.edu/hypermail/linux/kernel/0206.1/0402.html<p>In particular, he said:<p>~
PS. And never _ever_ make the "pointerness" part of the type. People who
write<p><pre><code> typedef struct urb_struct * urbp_t;
</code></pre>
(or whatever the name was) should just be shot.
~<p>I took this advice seriously and held out to the bitter end. But today I released this code:<p>https://github.com/chkoreff/Fexl/commit/9e361240168151706bcc289b400059a37c5245d8<p>Look at the difference in that sample code in the commit message. Using typedef is so obviously better that I just have to ignore Linus' advice here.
Here are the links:<p><a href="http://lkml.indiana.edu/hypermail/linux/kernel/0206.1/0402.html" rel="nofollow">http://lkml.indiana.edu/hypermail/linux/kernel/0206.1/0402.h...</a><p><a href="https://github.com/chkoreff/Fexl/commit/9e361240168151706bcc289b400059a37c5245d8" rel="nofollow">https://github.com/chkoreff/Fexl/commit/9e361240168151706bcc...</a>