1<html><head><title>toybox news</title>
   2<!--#include file="header.html" -->
   5<li><h2><a href="#capitalize">Do you capitalize toybox?</a></h2></li>
   6<li><h2><a href="#why_toybox">Why toybox? (What was wrong with busybox?)</a></h2></li>
   7<li><h2><a href="#support_horizon">Why a 7 year support horizon?</a></h2></li>
   8<li><h2><a href="#code">Where do I start understanding the toybox source code?</a></h2></li>
  11<a name="capitalize" />
  12<h2>Q: Do you capitalize toybox?</h2>
  14<p>A: Only at the start of a sentence. The command name is all lower case so
  15it seems silly to capitalize the project name, but not capitalizing the
  16start of sentences is awkward, so... compromise. (It is _not_ "ToyBox".)</p>
  18<a name="why_toybox" />
  19<h2>Q: "Why is there toybox? What was wrong with busybox?"</h2>
  21<p>A: Toybox started back in 2006 when I
  22<a href=>handed off BusyBox maintainership</a>
  23and <a href=>started over from
  24scratch</a> on a new codebase after a
  25<a href=>protracted licensing argument</a> took all the fun out of working on BusyBox.</p>
  27<p>Toybox was just a personal project until it got
  28<a href=>relaunched
  29in November 2011</a> with a new goal to
  30<a href=>make Android
  31self-hosting</a>. This involved me relicensing my own
  32code, which <a href=>made people who had
  33never used or participated in the project loudly angry</a>. The switch came
  34after a lot of thinking <a href=>about
  35licenses</a> and <a href=>the
  36transition to smartphones</a>, which led to a
  37<a href=>2013</a>
  38<a href=>talk</a> laying
  39out a strategy to make Android self-hosting using toybox. This helped
  40<a href=>bring
  41it to Android's attention</a>, and they
  42<a href=>merged it</a> into Android M.</p>
  44<p>The answer to the second question is "licensing". BusyBox predates Android
  45by almost a decade but Android still doesn't ship with it because GPLv3 came
  46out around the same time Android did and caused many people to throw
  47out the GPLv2 baby with the GPLv3 bathwater.
  48Android <a href=>explicitly
  49discourages</a> use of GPL and LGPL licenses in its products, and has gradually
  50reimplemented historical GPL components such as its bluetooth stack under the
  51Apache license. Similarly, Apple froze xcode at the last GPLv2 releases
  52(GCC 4.2.1 with binutils 2.17) for over 5 years while it sponsored the
  53development of new projects (clang/llvm/lld) to replace them,
  54implemented its SMB server from scratch to replace samba,
  55<a href=>and so
  56on</a>. Toybox itself exists because somebody with in a legacy position
  57just wouldn't shut up about GPLv3, otherwise I would probably
  58still happily be maintaining BusyBox. (For more on how I wound
  59up working on busybox in the first place,
  60<a href=>see here</a>.)</p>
  62<h2><a name="support_horizon">Q: Why a 7 year support horizon?</a></h2>
  64<p>A: Our <a href=>longstanding rule of thumb</a> is to try to run and build on
  65hardware and distributions released up to 7 years ago, and feel ok dropping
  66support for stuff older than that. (This is a little longer than Ubuntu's
  67Long Term Support, but not by much.)</p>
  69<p>If a kernel or libc feature is less than 7 years old, I try to have a
  70build-time configure test for it and let the functionality cleanly drop out.
  71I also keep old Ubuntu images around in VMs and perform the occasional
  72defconfig build there to see what breaks. (I'm not perfect about this,
  73but I accept bug reports.)</p>
  75<p>My original theory was "4 to 5 18-month cycles of moore's law should cover
  76the vast majority of the installed base of PC hardware", loosely based on some
  77research I did <a href=>back in 2003</a>
  78and <a href=>updated in 2006</a>
  79which said that low end systems were 2 iterations of moore's
  80law below the high end systems, and that another 2-3 iterations should cover
  81the useful lifetime of most systems no longer being sold but still in use and
  82potentially being upgraded to new software releases.</p>
  84<p>It turns out <a href=>I missed
  85industry changes</a> in the 1990's that stretched the gap
  86from low end to high end from 2 cycles to 4 cycles, and _that_ analysis
  87ignored the switch from PC to smartphone cutting off the R&D air supply of the
  88laptop market.  Meanwhile the Moore's Law s-curve started bending
  89down in 2000 and these days is pretty flat because the drive for faster clock
  90speeds <a href=>stumbled</a>
  91then <a href=>died</a>, and
  92the subsequent drive to go wide maxed out around 4x SMP with ~2 megabyte
  93caches for most applications. These days the switch from exponential to
  94linear growth in hardware capabilities is
  95<a href=>common</a>
  96<a href=>knowledge</a>.</p>
  98<p>But the 7 year rule of thumb stuck around anyway: if a kernel or libc
  99feature is less than 7 years old, I try to have a build-time configure test
 100for it and let the functionality cleanly drop out. I also keep old Ubuntu
 101images around in VMs and perform the occasional defconfig build there to
 102see what breaks.</p>
 104<h2><a name="code" />Where do I start understanding the source code?</h2>
 106<p>Toybox is written in C. There are longer writeups of the
 107<a href=design.html>design ideas</a> and a <a href=code.html>code walkthrough</a>,
 108and the <a href=about.html>about page</a> summarizes what we're trying to
 109accomplish, but here's a quick start:</p>
 111<p>Toybox uses the standard three stage configure/make/install
 112<a href=code.html#building>build</a>, in this case "<b>make defconfig;
 113make; make install</b>". Type "<b>make help</b>" to
 114see available make targets.</p>
 116<p><b>The configure stage is copied from the Linux kernel</b> (in the "kconfig"
 117directory), and saves your selections in the file ".config" at the top
 118level. The "defconfig" target selects the
 119maximum sane configuration (enabling all the commands and features that
 120aren't unfinished, only intended as examples, debug code, etc) and is
 121probably what you want. You can use "make menuconfig" to manually select
 122specific commands to include, through an interactive menu (cursor up and
 123down, enter to descend into a sub-menu, space to select an entry, ? to see
 124an entry's help text, esc to exit). The menuconfig help text is the
 125same as the command's --help output.</p>
 127<p><b>The "make" stage creates a toybox binary</b> (which is stripped, look in
 128generated/unstripped for the debug versions), and "install" adds a bunch of
 129symlinks to toybox under the various command names. Toybox determines which
 130command to run based on the filename, or you can use the "toybox" name in which case the first
 131argument is the command to run (ala "toybox ls -l"). <b>You can also build
 132individual commands as standalone executables</b>, ala "make sed cat ls".</p>
 134<p><b>The main() function is in main.c at the top level</b>,
 135along with setup plumbing and selecting which command to run this time.
 136The function toybox_main() implements the "toybox" multiplexer command.</p>
 138<p><b>The individual command implementations are under "toys"</b>, and are grouped
 139into categories (mostly based on which standard they come from, posix, lsb,
 140android...) The "pending" directory contains unfinished commands, and the
 141"examples" directory contains examples. Commands in those two directories
 142are _not_ selected by defconfig. (These days pending directory is mostly
 143third party submissions that have not yet undergone proper code review.)</p>
 145<p><b>Common infrastructure shared between commands is under "lib"</b>. Most
 146commands call lib/args.c to parse their command line arguments before calling
 147the command's own main() function, which uses the option string in
 148the command's NEWTOY() macro. This is similar to the libc function getopt(),
 149but more powerful, and is documented at the top of lib/args.c.</p>
 151<p>Most of the actual <b>build/install infrastructure is shell scripts under
 152"scripts"</b>. <b>These populate the "generated" directory</b> with headers
 153created from other files, which are <a href=code.html#generated>described</a>
 154in the code walkthrough. All the
 155build's temporary files live under generated, including the .o files built
 156from the .c files (in generated/obj). The "make clean" target deletes that
 157directory. ("make distclean" also deletes your .config and deletes the
 158kconfig binaries that process .config.)</p>
 160<p>Each command's file contains all the information for that command, so
 161<b>adding a command to toybox means adding a single file under "toys"</b>.
 162Usually you <a href=code.html#adding>start a new command</a> by copying an
 163existing command file to a new filename
 164(toys/examples/hello.c, toys/examples/skeleton.c, toys/posix/cat.c,
 165and toys/posix/true.c have all been used for this purpose) and then replacing
 166all instances of its old name with the new name (which should match the
 167new filename), and modifying the help text, argument string, and what the
 168code does. You might have to "make distclean" before you new command
 169shows up in defconfig or menuconfig.</p>
 171<p><b>The toybox test suite lives in the "tests" directory</b>. From the top
 172level you can "make tests" to test everything, or "make test_sed" test a
 173single command's standalone version (which should behave identically)
 174but that's why we test.</p>
 176<!--#include file="footer.html" -->