about summary refs log blame commit diff stats
path: root/build3
blob: 09c7398d1bf6d80ff16a1249fd7917ab3f290b3e (plain) (tree)
<
<HTML>
<HEAD>
<TITLE>Computer Science Logo Style vol 3:Acknowledgments</TITLE>
</HEAD>
<BODY>
<CITE>Computer Science Logo Style</CITE> volume 3:
<CITE>Beyond Programming</CITE> 2/e Copyright (C) 1997 MIT
<H1>Acknowledgments</H1>

<TABLE width="100%"><TR><TD>
<IMG SRC="../csls3.jpg" ALT="cover photo">
<TD><TABLE>
<TR><TD align="right"><CITE><A HREF="http://www.cs.berkeley.edu/~bh/">Brian
Harvey</A><BR>University of California, Berkeley</CITE>
<TR><TD align="right"><BR>
<TR><TD align="right"><A HREF="../pdf/v3ch00.pdf">Download PDF version</A>
<TR><TD align="right"><A HREF="../v3-toc2.html">Back to Table of Contents</A>
<TR><TD align="right"><A HREF="preface.html"><STRONG>BACK</STRONG></A>
chapter thread <A HREF="../v3ch1/v3ch1.html"><STRONG>NEXT</STRONG></A>
<TR><TD align="right"><A HREF="https://mitpress.mit.edu/books/computer-science-logo-style-second-edition-volume-3">MIT
Press web page for <CITE>Computer Science Logo Style</CITE></A>
</TABLE></TABLE>

<HR>

<P>
As for the previous two volumes, my greatest debts are to Hal Abelson and
Paul Goldenberg.  Both of them read the manuscript carefully through several
drafts.  Hal is great at noticing the large problems; he makes comments like
&quot;throw out this whole chapter&quot; and &quot;you are putting the cart before the
horse here.&quot;  Paul's comments were generally on a more detailed level,
pointing out sections in which potentially valuable content was sabotaged by
a presentation that nobody would understand.  Together they have improved
the book enormously.

<P>Some of the examples in this book are ones that were posed to me by other
people in other contexts.  Horacio Reggini raised the issue of listing (not
merely counting) the combinations of <EM>r</EM> elements of a list;
Dick White
asked me to investigate just how secure the Simplex lock is;
Chris Anderson
taught the probability class where the question about multinomial expansions
arose.  I'm grateful to Anita Harnadek, whom I've never met, for
a logic
problem I use to demonstrate inference systems.  (She is, by the way, the
author of a fantastic textbook called <EM>Critical Thinking</EM> that I
recommend to teachers of almost any subject: math, English, or social
studies.)  Jim Davis's Logo interpreter in Logo (in the <EM>LogoWorks</EM>
anthology I co-edited) was an inspiration for the Pascal compiler.

<P>
I'm grateful to Dan Bobrow, Sherry Turkle, and
Terry Winograd for permission
to quote from their work here.  In particular, Bobrow's doctoral thesis
forms the basis for my chapter on artificial intelligence, and I'm grateful
for the program design as well as my extensive quotations from the thesis
itself.  He was also very patient in answering technical questions about
details of a program he wrote over 20 years ago.

<P>
Mike Clancy taught me about generating functions and used them to find the
closed form definition for the multinomial problem; Michael Somos,
via the <CODE>sci.math</CODE> newsgroup, provided the closed form solution to
the Simplex lock problem.  Paul Hilfinger straightened me out
about parser complexity.

<P>

<P><A HREF="../v3-toc2.html">(back to Table of Contents)</A>
<P><A HREF="preface.html"><STRONG>BACK</STRONG></A>
chapter thread <A HREF="../v3ch1/v3ch1.html"><STRONG>NEXT</STRONG></A>

<P>
<ADDRESS>
<A HREF="../index.html">Brian Harvey</A>, 
<CODE>bh@cs.berkeley.edu</CODE>
</ADDRESS>
</BODY>
</HTML>
="p">{ scheme = scheme[:strings.IndexRune(scheme, '+')] fmt.Println(scheme) } backend, err := handlers.GetHandlerForScheme(scheme, worker) if err != nil { return nil, err } worker.Backend = backend return worker, nil }
                                                                            
                                   
                                                                                      
                                   
                                                                        
                          
                                                                                                      

                                                       










                                                                             
                                                



               



                                                                                  
                                                                                                     
                                    
                                                                                             
                                                                                                    





                                                           
                                                                                        




































                                                                                                                                                          
                                                         
#!/bin/sh
# Alternative to build2 that can stop after any step. For example:
#   $ ./build3 mu.cc

set -e  # stop immediately on error

# Some environment variables that can be passed in. For example, to turn off
# optimization:
#   $ CFLAGS=-g ./build3
test "$CXX" || export CXX=c++
test "$CC" || export CC=cc
test "$CFLAGS" || export CFLAGS="-g -O2"
export CFLAGS="$CFLAGS -Wall -Wextra -ftrapv -fno-strict-aliasing"

# Outline:
# [0-9]*.cc -> mu.cc -> .build/*.cc -> .build/*.o -> .build/mu_bin
# (layers)   |        |              |             |
#          tangle  cleave          $CXX          $CXX

## arg parsing

# can be called with a target to stop after a partial build
#   $ ./build3 mu.cc
# can also be called with a layer to only build until
#   $ ./build3 --until 050
# scenarios:
#   ./build3              => TARGET=  UNTIL_LAYER=zzz
#   ./build3 x            => TARGET=x UNTIL_LAYER=zzz
#   ./build3 --until      => TARGET=  UNTIL_LAYER=zzz
#   ./build3 --until 050  => TARGET=  UNTIL_LAYER=050
TARGET=
UNTIL_LAYER=zzz
if [ $# -ge 1 ] && [ $1 != "--until" ]
then
  TARGET=$1
fi
if [ $# -ge 2 ] && [ $1 = "--until" ]
then
  UNTIL_LAYER=$2
fi

##

# there's two mechanisms for fast builds here:
# - if a command is quick to run, always run it but update the result only on any change
# - otherwise run it only if the output is 'older_than' the inputs
#
# avoid combining both mechanisms for a single file
# otherwise you'll see spurious messages about files being updated
# risk: a file may unnecessarily update without changes, causing unnecessary work downstream

# return 1 if $1 is older than _any_ of the remaining args
# also exit the entire script if previous invocation was to update $TARGET
older_than() {
  test $TARGET  &&  test "$last_target" = "$TARGET"  &&  exit 0
  local target=$1
  shift
  last_target=$target
  if [ ! -e $target ]
  then
#?     echo "$target doesn't exist"
    echo "updating $target" >&2
    return 0  # success
  fi
  local f
  for f in $*
  do
    if [ $f -nt $target ]
    then
      echo "updating $target" >&2
      return 0  # success
    fi
  done
  return 1  # failure
}

# redirect to $1, unless it's already identical
# no point checking for an early exit, because this usually runs in a pipeline/subshell
update() {
  if [ ! -e $1 ]
  then
    cat > $1
  else
    cat > $1.tmp
    diff -q $1 $1.tmp >/dev/null  &&  rm $1.tmp  ||  mv $1.tmp $1
  fi
}

# cp file $1 to directory $2, unless it's already identical
# also exit the entire script if previous invocation was to update $TARGET
update_cp() {
  test $TARGET  &&  test "$last_target" = "$TARGET"  &&  exit 0
  last_target=$2/$1
  if [ ! -e $2/$1 ]
  then
    cp $1 $2
  elif [ $1 -nt $2/$1 ]
  then
    cp $1 $2
  fi
}

noisy_cd() {
  cd $1
  echo "-- `pwd`" >&2
}

older_than enumerate/enumerate enumerate/enumerate.cc && {
  $CXX $CFLAGS enumerate/enumerate.cc -o enumerate/enumerate
}

older_than tangle/tangle tangle/*.cc && {
  noisy_cd tangle
    # auto-generate various lists (ending in '_list' by convention) {
    # list of types
    {
      grep -h "^struct .* {" [0-9]*.cc  |sed 's/\(struct *[^ ]*\).*/\1;/'
      grep -h "^typedef " [0-9]*.cc
    }  |update type_list
    # list of function declarations, so I can define them in any order
    grep -h "^[^ #].*) {" [0-9]*.cc  |sed 's/ {.*/;/'  |update function_list
    # list of code files to compile
    ls [0-9]*.cc  |grep -v "\.test\.cc$"  |sed 's/.*/#include "&"/'  |update file_list
    # list of test files to compile
    ls [0-9]*.test.cc  |sed 's/.*/#include "&"/'  |update test_file_list
    # list of tests to run
    grep -h "^[[:space:]]*void test_" [0-9]*.cc  |sed 's/^\s*void \(.*\)() {$/\1,/'  |update test_list
    # }
    # Now that we have all the _lists, compile 'tangle'
    $CXX $CFLAGS boot.cc -o tangle
  noisy_cd ..  # no effect; just to show us returning to the parent directory
}

LAYERS=$(./enumerate/enumerate --until $UNTIL_LAYER  |grep '\.cc$')
older_than mu.cc $LAYERS enumerate/enumerate tangle/tangle && {
  # no update here; rely on 'update' calls downstream
  ./tangle/tangle $LAYERS  > mu.cc
}

older_than cleave/cleave cleave/cleave.cc && {
  $CXX $CFLAGS cleave/cleave.cc -o cleave/cleave
  rm -rf .build
}

mkdir -p .build
# auto-generate function declarations, so I can define them in any order
# functions start out unindented, have all args on the same line, and end in ') {'
#
#                                      \/ ignore methods
grep -h "^[^[:space:]#].*) {$" mu.cc  |grep -v ":.*("  |sed 's/ {.*/;/'  |update .build/function_list
# auto-generate list of tests to run
grep -h "^\s*void test_" mu.cc  |sed 's/^\s*void \(.*\)() {.*/\1,/'  |update .build/test_list
grep -h "^\s*void test_" mu.cc  |sed 's/^\s*void \(.*\)() {.*/"\1",/'  |update .build/test_name_list
mkdir -p .build/termbox
update_cp termbox/termbox.h .build/termbox

older_than mu_bin mu.cc *_list cleave/cleave termbox/* && {
  ./cleave/cleave mu.cc .build
  noisy_cd .build
    # create the list of global variable declarations from the corresponding definitions
    grep ';' global_definitions_list  |sed 's/[=(].*/;/'  |sed 's/^[^\/# ]/extern &/'  |sed 's/^extern extern /extern /'  |update global_declarations_list
    for f in mu_*.cc
    do
      older_than `echo $f  |sed 's/\.cc$/.o/'` $f header global_declarations_list function_list test_list && {
        $CXX $CFLAGS -c $f
      }
    done
  noisy_cd ../termbox
    older_than utf8.o utf8.c && {
      $CC $CFLAGS -c utf8.c
    }
    older_than termbox.o termbox.c termbox.h input.inl output.inl bytebuffer.inl && {
      $CC $CFLAGS -c termbox.c
    }
    older_than libtermbox.a *.o && {
      ar rcs libtermbox.a *.o
    }
  noisy_cd ..
  $CXX $CFLAGS .build/*.o termbox/libtermbox.a -o .build/mu_bin
  cp .build/mu_bin .
}

## [0-9]*.mu -> core.mu

MU_LAYERS=$(./enumerate/enumerate --until $UNTIL_LAYER  |grep '\.mu$') || exit 0  # ok if no .mu files
cat $MU_LAYERS  |update core.mu

exit 0

# scenarios considered:
#   0 status when nothing needs updating
#   no output when nothing needs updating
#     no output for mu.cc when .mu files modified
#     touch mu.cc but don't modify it; no output on second build
#     touch a .cc layer but don't modify it; no output on second build
#   only a single layer is recompiled when changing a C++ function
#   stop immediately after failure in tangle
#   stop immediately after target provided at commandline