tests/pdiff

context: skip path conflicts by default when clearing unknown file (issue5776)

Prior to adding path conflict checking in 989e884d1be9, the test-audit-path.t
tests failed as shown here (but it was globbed away).  989e884d1be9 made it fail
with a message about the destination manifest containing a conflict (though the
no-symlink case wasn't updated).  When the path conflict checking was gated
behind an experimental config in 2a774cae3a03^::2a774cae3a03, the update started
erroneously succeeding here.  It turns out that the child of 989e884d1be9 is the
origin of this change when path conflict checking is disabled, as shown by
grafting the experimental config range on top of it.

What's happening here is merge.batchget() is writing the symlink 'back' to wdir
(but as a regular file for the no-symlink case), and then tries to write
'back/test', but calls wctx['back/test'].clearunknown() first.  The code that's
gated here was removing the newly written 'back' file, allowing 'back/test' to
succeed.  I tried checking for the dir components of 'back/test' in dirstate,
and skipping removal if present.  But that didn't work because the dirstate
isn't updated after each file is written out.

This is the last persistent test failure on Windows, so the testbot should start
turning green now. \o/

#!/bin/sh

# Script to get stable diff output on any platform.
#
# Output of this script is almost equivalent to GNU diff with "-Nru".
#
# Use this script as "hg pdiff" via extdiff extension with preparation
# below in test scripts:
#
#   $ cat >> $HGRCPATH <<EOF
#   > [extdiff]
#   > pdiff = sh "$RUNTESTDIR/pdiff"
#   > EOF

filediff(){
    # USAGE: filediff file1 file2 [header]

    # compare with /dev/null if file doesn't exist (as "-N" option)
    file1="$1"
    if test ! -f "$file1"; then
        file1=/dev/null
    fi
    file2="$2"
    if test ! -f "$file2"; then
        file2=/dev/null
    fi

    if cmp -s "$file1" "$file2" 2> /dev/null; then
        # Return immediately, because comparison isn't needed. This
        # also avoids redundant message of diff like "No differences
        # encountered" (on Solaris)
        return
    fi

    if test -n "$3"; then
        # show header only in recursive case
        echo "$3"
    fi

    # replace "/dev/null" by corresponded filename (as "-N" option)
    diff -u "$file1" "$file2" |
    sed "s@^--- /dev/null\(.*\)\$@--- $1\1@" |
    sed "s@^\+\+\+ /dev/null\(.*\)\$@+++ $2\1@"

    # in this case, files differ from each other
    return 1
}

if test -d "$1" -o -d "$2"; then
    # ensure comparison in dictionary order
    (
    if test -d "$1"; then (cd "$1" && find . -type f); fi
    if test -d "$2"; then (cd "$2" && find . -type f); fi
    ) |
    sed 's@^\./@@g' | sort | uniq |
    while read file; do
        filediff "$1/$file" "$2/$file" "diff -Nru $1/$file $2/$file"
    done

    # TODO: there is no portable way for current while-read based
    # implementation to return 1 at detecting changes.
    #
    # On bash and dash, assignment to variable inside while-block
    # doesn't affect outside, because inside while-block is executed
    # in sub-shell. BTW, it affects outside while-block on ksh (as sh
    # on Solaris).
else
    filediff "$1" "$2"
fi