NAME

dummydir – generate and manipulate a directory tree of files for
testing purposes

SYNOPSIS

dummydir operation directory [options]

dummydir [–replay=replay-file] directory

DESCRIPTION

dummydir (DummyDir) generates and manipulates a directory
tree of files for testing purposes. It uses a deterministic, seedable
pseudo-random number generator, so a run can be reproduced exactly later
from its recorded seed.

Every dummy directory maintains a manifest file,
DummyDir.manifest, in its root, which tracks the files and
subdirectories that dummydir created and is responsible for.
Operations only ever add, modify, or remove items recorded in this manifest;
files and directories not created by dummydir are never touched, even
if they live inside the target directory.

Every run is also logged to a replay file, DummyDir.replay,
in the root of the target directory. This file records each operation and
the flags (including the random seed) it was invoked with, so the exact
sequence of operations can be reproduced later with –replay.

OPERATIONS

add

Add new files to the dummy directory.

delete

Delete existing files from the dummy directory.

modify

Modify the contents of existing files.

rename

Rename existing files.

move

Move existing files to a different directory.

dadd

Add new subdirectories.

drename

Rename existing subdirectories.

dmove

Move existing subdirectories to a different parent directory.

ddelete

Delete existing subdirectories, along with everything inside them.

clean

Remove every file and directory this tool has created.

verify

Check that every file on disk matches what the manifest recorded.

rebuild

Rebuild the manifest from what is actually present on disk.

OPTIONS

–size=value

Total number of bytes the operation should affect. Accepts K/M/G/T/P/E
suffixes (e.g. 500M, 2.5G), a percentage of the relevant total (e.g. 25%),
or a range to pick randomly from (e.g. 1k-10m). Shorthand: -s.

–count=value

Total number of items (files or directories) the operation should affect.
Accepts the same percentage and range rules as –size. Shorthand:
-c.

–fileseed=value

A numerical value that acts as a seed for file add and modify operations.
Files with identical seeds will have identical data. Ranges may be
used.

–filesize=value

For add, the size of each new file. For modify, how much of
each file’s content to change. Same value rules as –size.

–filetype=type

For add only: random, binary, sparse, or
text. Default: random.

–modifytype=type

For modify only: append, truncate, overwrite,
chop, insert, or random. Default: random.

–maxdepth=N

For dadd and dmove only: how many levels deep the directory
structure is allowed to go. Default: 2.

–threads=N

Number of worker threads to use for file operations. 0 or 1 disables
threading. Default: the number of hardware threads available.

–seed=hex

Seed for the pseudo-random number generator, as a hex string. If not
given, a random seed is generated and recorded so the run can be
reproduced later.

–verbose

Print details about every individual error, conflict, or discrepancy
found, instead of just a summary count. Shorthand: -v.

–replay[=file]

Execute the operations, in order, from the provided replay file.

–help

Show a help message and exit.

EXIT
STATUS

0

Success.

1

Invalid arguments, an unrecognized operation or flag, or a missing/invalid
directory path, exception thrown, or files did not pass validation.

FILES

DummyDir.manifest

Located in the root of the target directory. Tracks every file and
directory dummydir has created there, along with each file’s size
and MD5 hash.

DummyDir.replay

Located in the root of the target directory. A log of every operation
performed against it, in the exact form needed to reproduce the run with
–replay.

EXAMPLES

Add 500 megabytes of new files,
each between 10 and 20 megabytes:

dummydir add ./mydir –size=500M –filesize=10M-20M

Delete 25% of the files
currently in the directory:

dummydir delete ./mydir –count=25%

Move 5 random subdirectories,
keeping the tree to 3 levels deep or less:

dummydir dmove ./mydir –count=5 –maxdepth=3

Check every file against the
manifest, listing any problems found:

dummydir verify ./mydir –verbose

Rebuild the manifest
from what is actually present on disk:

dummydir rebuild ./mydir

Replay a previously
recorded sequence of operations:

dummydir –replay ./mydir

NOTES

Every operation, once completed, reports how many bytes and how
many items were affected, along with the resulting MD5 hash of the manifest
file. Two runs performed with the same sequence of operations and the same
recorded seeds will always produce manifests with an identical hash, which
makes dummydir suitable for reproducible test fixtures and regression
baselines.

Numeric flags can be presented as a single number (eg 101), range
of numbers (eg 101-150), percentage (50%), or range of percentages
(25%-50%). A letter can be placed after the number to indicate a base 2
multiplier (eg 10k = 10240).

AUTHOR

Written by Chris Bartle.

LICENSE

BSD 3-Clause License. See the LICENSE file distributed with
the source code for the full text.