diff --git a/src/build/build.v b/src/build/build.v
index 0de91a6..2784c26 100644
--- a/src/build/build.v
+++ b/src/build/build.v
@@ -9,9 +9,10 @@ import client
 import strings
 import util
 
-const container_build_dir = '/build'
-
-const build_image_repo = 'vieter-build'
+const (
+	container_build_dir = '/build'
+	build_image_repo    = 'vieter-build'
+)
 
 // create_build_image creates a builder image given some base image which can
 // then be used to build & package Arch images. It mostly just updates the
diff --git a/src/docker/README.md b/src/docker/README.md
new file mode 100644
index 0000000..4cc8971
--- /dev/null
+++ b/src/docker/README.md
@@ -0,0 +1,3 @@
+This module implements part of the Docker Engine API v1.41
+([documentation](https://docs.docker.com/engine/api/v1.41/)) using socket-based
+HTTP communication.
diff --git a/src/docker/containers.v b/src/docker/containers.v
index 7605452..62c8031 100644
--- a/src/docker/containers.v
+++ b/src/docker/containers.v
@@ -15,7 +15,7 @@ struct Container {
 }
 
 // containers returns a list of all containers.
-pub fn (mut d DockerDaemon) containers() ?[]Container {
+pub fn (mut d DockerConn) containers() ?[]Container {
 	d.send_request('GET', urllib.parse('/v1.41/containers/json')?)?
 	head, res := d.read_response()?
 
@@ -46,7 +46,7 @@ pub:
 }
 
 // create_container creates a new container with the given config.
-pub fn (mut d DockerDaemon) create_container(c NewContainer) ?CreatedContainer {
+pub fn (mut d DockerConn) create_container(c NewContainer) ?CreatedContainer {
 	d.send_request_with_json('POST', urllib.parse('/v1.41/containers/create')?, c)?
 	head, res := d.read_response()?
 
@@ -62,7 +62,7 @@ pub fn (mut d DockerDaemon) create_container(c NewContainer) ?CreatedContainer {
 }
 
 // start_container starts the container with the given id.
-pub fn (mut d DockerDaemon) start_container(id string) ? {
+pub fn (mut d DockerConn) start_container(id string) ? {
 	d.send_request('POST', urllib.parse('/v1.41/containers/$id/start')?)?
 	head, body := d.read_response()?
 
@@ -92,7 +92,7 @@ pub mut:
 }
 
 // inspect_container returns detailed information for a given container.
-pub fn (mut d DockerDaemon) inspect_container(id string) ?ContainerInspect {
+pub fn (mut d DockerConn) inspect_container(id string) ?ContainerInspect {
 	d.send_request('GET', urllib.parse('/v1.41/containers/$id/json')?)?
 	head, body := d.read_response()?
 
@@ -114,7 +114,7 @@ pub fn (mut d DockerDaemon) inspect_container(id string) ?ContainerInspect {
 }
 
 // remove_container removes the container with the given id.
-pub fn (mut d DockerDaemon) remove_container(id string) ? {
+pub fn (mut d DockerConn) remove_container(id string) ? {
 	d.send_request('DELETE', urllib.parse('/v1.41/containers/$id')?)?
 	head, body := d.read_response()?
 
@@ -127,7 +127,7 @@ pub fn (mut d DockerDaemon) remove_container(id string) ? {
 
 // get_container_logs returns a reader object allowing access to the
 // container's logs.
-pub fn (mut d DockerDaemon) get_container_logs(id string) ?&StreamFormatReader {
+pub fn (mut d DockerConn) get_container_logs(id string) ?&StreamFormatReader {
 	d.send_request('GET', urllib.parse('/v1.41/containers/$id/logs?stdout=true&stderr=true')?)?
 	head := d.read_response_head()?
 
diff --git a/src/docker/docker.v b/src/docker/docker.v
index f612d1f..4ce1ea6 100644
--- a/src/docker/docker.v
+++ b/src/docker/docker.v
@@ -15,17 +15,17 @@ const (
 	http_chunk_separator = [u8(`\r`), `\n`]
 )
 
-pub struct DockerDaemon {
+pub struct DockerConn {
 mut:
 	socket &unix.StreamConn
 	reader &io.BufferedReader
 }
 
 // new_conn creates a new connection to the Docker daemon.
-pub fn new_conn() ?&DockerDaemon {
+pub fn new_conn() ?&DockerConn {
 	s := unix.connect_stream(docker.socket)?
 
-	d := &DockerDaemon{
+	d := &DockerConn{
 		socket: s
 		reader: io.new_buffered_reader(reader: s)
 	}
@@ -34,12 +34,12 @@ pub fn new_conn() ?&DockerDaemon {
 }
 
 // close closes the underlying socket connection.
-pub fn (mut d DockerDaemon) close() ? {
+pub fn (mut d DockerConn) close() ? {
 	d.socket.close()?
 }
 
 // send_request sends an HTTP request without body.
-pub fn (mut d DockerDaemon) send_request(method string, url urllib.URL) ? {
+pub fn (mut d DockerConn) send_request(method string, url urllib.URL) ? {
 	req := '$method $url.request_uri() HTTP/1.1\nHost: localhost\n\n'
 
 	d.socket.write_string(req)?
@@ -49,7 +49,7 @@ pub fn (mut d DockerDaemon) send_request(method string, url urllib.URL) ? {
 }
 
 // send_request_with_body sends an HTTP request with the given body.
-pub fn (mut d DockerDaemon) send_request_with_body(method string, url urllib.URL, content_type string, body string) ? {
+pub fn (mut d DockerConn) send_request_with_body(method string, url urllib.URL, content_type string, body string) ? {
 	req := '$method $url.request_uri() HTTP/1.1\nHost: localhost\nContent-Type: $content_type\nContent-Length: $body.len\n\n$body\n\n'
 
 	d.socket.write_string(req)?
@@ -60,7 +60,7 @@ pub fn (mut d DockerDaemon) send_request_with_body(method string, url urllib.URL
 
 // send_request_with_json<T> is a convenience wrapper around
 // send_request_with_body that encodes the input as JSON.
-pub fn (mut d DockerDaemon) send_request_with_json<T>(method string, url urllib.URL, data &T) ? {
+pub fn (mut d DockerConn) send_request_with_json<T>(method string, url urllib.URL, data &T) ? {
 	body := json.encode(data)
 
 	return d.send_request_with_body(method, url, 'application/json', body)
@@ -70,17 +70,17 @@ pub fn (mut d DockerDaemon) send_request_with_json<T>(method string, url urllib.
 // '\r\n\r\n', after which it parses the response as an HTTP response.
 // Importantly, this function never consumes the reader past the HTTP
 // separator, so the body can be read fully later on.
-pub fn (mut d DockerDaemon) read_response_head() ?http.Response {
+pub fn (mut d DockerConn) read_response_head() ?http.Response {
 	mut res := []u8{}
 
-	util.read_until_separator(mut d.reader, mut res, http_separator) ?
+	util.read_until_separator(mut d.reader, mut res, docker.http_separator)?
 
 	return http.parse_response(res.bytestr())
 }
 
 // read_response_body reads `length` bytes from the stream. It can be used when
 // the response encoding isn't chunked to fully read it.
-pub fn (mut d DockerDaemon) read_response_body(length int) ?string {
+pub fn (mut d DockerConn) read_response_body(length int) ?string {
 	if length == 0 {
 		return ''
 	}
@@ -101,7 +101,7 @@ pub fn (mut d DockerDaemon) read_response_body(length int) ?string {
 // read_response is a convenience function which always consumes the entire
 // response & returns it. It should only be used when we're certain that the
 // result isn't too large.
-pub fn (mut d DockerDaemon) read_response() ?(http.Response, string) {
+pub fn (mut d DockerConn) read_response() ?(http.Response, string) {
 	head := d.read_response_head()?
 
 	if head.header.get(http.CommonHeader.transfer_encoding) or { '' } == 'chunked' {
@@ -121,7 +121,7 @@ pub fn (mut d DockerDaemon) read_response() ?(http.Response, string) {
 
 // get_chunked_response_reader returns a ChunkedResponseReader using the socket
 // as reader.
-pub fn (mut d DockerDaemon) get_chunked_response_reader() &ChunkedResponseReader {
+pub fn (mut d DockerConn) get_chunked_response_reader() &ChunkedResponseReader {
 	r := new_chunked_response_reader(d.reader)
 
 	return r
@@ -129,7 +129,7 @@ pub fn (mut d DockerDaemon) get_chunked_response_reader() &ChunkedResponseReader
 
 // get_stream_format_reader returns a StreamFormatReader using the socket as
 // reader.
-pub fn (mut d DockerDaemon) get_stream_format_reader() &StreamFormatReader {
+pub fn (mut d DockerConn) get_stream_format_reader() &StreamFormatReader {
 	r := new_chunked_response_reader(d.reader)
 	r2 := new_stream_format_reader(r)
 
diff --git a/src/docker/images.v b/src/docker/images.v
index 5866905..cab7f34 100644
--- a/src/docker/images.v
+++ b/src/docker/images.v
@@ -10,7 +10,7 @@ pub:
 }
 
 // pull_image pulls the given image:tag.
-pub fn (mut d DockerDaemon) pull_image(image string, tag string) ? {
+pub fn (mut d DockerConn) pull_image(image string, tag string) ? {
 	d.send_request('POST', urllib.parse('/v1.41/images/create?fromImage=$image&tag=$tag')?)?
 	head := d.read_response_head()?
 
@@ -33,7 +33,7 @@ pub fn (mut d DockerDaemon) pull_image(image string, tag string) ? {
 }
 
 // create_image_from_container creates a new image from a container.
-pub fn (mut d DockerDaemon) create_image_from_container(id string, repo string, tag string) ?Image {
+pub fn (mut d DockerConn) create_image_from_container(id string, repo string, tag string) ?Image {
 	d.send_request('POST', urllib.parse('/v1.41/commit?container=$id&repo=$repo&tag=$tag')?)?
 	head, body := d.read_response()?
 
@@ -49,7 +49,7 @@ pub fn (mut d DockerDaemon) create_image_from_container(id string, repo string,
 }
 
 // remove_image removes the image with the given id.
-pub fn (mut d DockerDaemon) remove_image(id string) ? {
+pub fn (mut d DockerConn) remove_image(id string) ? {
 	d.send_request('DELETE', urllib.parse('/v1.41/images/$id')?)?
 	head, body := d.read_response()?
 
diff --git a/src/docker/stream.v b/src/docker/stream.v
index ed73098..02fb972 100644
--- a/src/docker/stream.v
+++ b/src/docker/stream.v
@@ -58,9 +58,9 @@ pub fn (mut r ChunkedResponseReader) read(mut buf []u8) ?int {
 // This function should only be called if the previous chunk has been
 // completely consumed.
 fn (mut r ChunkedResponseReader) read_chunk_size() ?u64 {
-	mut buf := []u8{len: 2}
-
 	if r.started {
+		mut buf := []u8{len: 2}
+
 		// Each chunk ends with a `\r\n` which we want to skip first
 		r.reader.read(mut buf)?
 	}
@@ -68,7 +68,7 @@ fn (mut r ChunkedResponseReader) read_chunk_size() ?u64 {
 	r.started = true
 
 	mut res := []u8{}
-	util.read_until_separator(mut r.reader, mut res, http_chunk_separator) ?
+	util.read_until_separator(mut r.reader, mut res, http_chunk_separator)?
 
 	// The length of the next chunk is provided as a hexadecimal
 	mut num_data := hex.decode(res#[..-2].bytestr())?
diff --git a/src/env/env.v b/src/env/env.v
index d145931..3d07d23 100644
--- a/src/env/env.v
+++ b/src/env/env.v
@@ -3,12 +3,13 @@ module env
 import os
 import toml
 
-// The prefix that every environment variable should have
-const prefix = 'VIETER_'
-
-// The suffix an environment variable in order for it to be loaded from a file
-// instead
-const file_suffix = '_FILE'
+const (
+	// The prefix that every environment variable should have
+	prefix      = 'VIETER_'
+	// The suffix an environment variable in order for it to be loaded from a file
+	// instead
+	file_suffix = '_FILE'
+)
 
 fn get_env_var(field_name string) ?string {
 	env_var_name := '$env.prefix$field_name.to_upper()'
diff --git a/src/util/README.md b/src/util/README.md
new file mode 100644
index 0000000..529e412
--- /dev/null
+++ b/src/util/README.md
@@ -0,0 +1,2 @@
+This module defines a few useful functions used throughout the codebase that
+don't specifically fit inside a module.
diff --git a/src/util/stream.v b/src/util/stream.v
new file mode 100644
index 0000000..06397aa
--- /dev/null
+++ b/src/util/stream.v
@@ -0,0 +1,95 @@
+// Functions for interacting with `io.Reader` & `io.Writer` objects.
+module util
+
+import io
+import os
+
+// reader_to_writer tries to consume the entire reader & write it to the writer.
+pub fn reader_to_writer(mut reader io.Reader, mut writer io.Writer) ? {
+	mut buf := []u8{len: 10 * 1024}
+
+	for {
+		bytes_read := reader.read(mut buf) or { break }
+		mut bytes_written := 0
+
+		for bytes_written < bytes_read {
+			c := writer.write(buf[bytes_written..bytes_read]) or { break }
+
+			bytes_written += c
+		}
+	}
+}
+
+// reader_to_file writes the contents of a BufferedReader to a file
+pub fn reader_to_file(mut reader io.BufferedReader, length int, path string) ? {
+	mut file := os.create(path)?
+	defer {
+		file.close()
+	}
+
+	mut buf := []u8{len: reader_buf_size}
+	mut bytes_left := length
+
+	// Repeat as long as the stream still has data
+	for bytes_left > 0 {
+		// TODO check if just breaking here is safe
+		bytes_read := reader.read(mut buf) or { break }
+		bytes_left -= bytes_read
+
+		mut to_write := bytes_read
+
+		for to_write > 0 {
+			// TODO don't just loop infinitely here
+			bytes_written := file.write(buf[bytes_read - to_write..bytes_read]) or { continue }
+			// file.flush()
+
+			to_write = to_write - bytes_written
+		}
+	}
+}
+
+// match_array_in_array<T> returns how many elements of a2 overlap with a1. For
+// example, if a1 = "abcd" & a2 = "cd", the result will be 2. If the match is
+// not at the end of a1, the result is 0.
+pub fn match_array_in_array<T>(a1 []T, a2 []T) int {
+	mut i := 0
+	mut match_len := 0
+
+	for i + match_len < a1.len {
+		if a1[i + match_len] == a2[match_len] {
+			match_len += 1
+		} else {
+			i += match_len + 1
+			match_len = 0
+		}
+	}
+
+	return match_len
+}
+
+// read_until_separator consumes an io.Reader until it encounters some
+// separator array. The data read is stored inside the provided res array.
+pub fn read_until_separator(mut reader io.Reader, mut res []u8, sep []u8) ? {
+	mut buf := []u8{len: sep.len}
+
+	for {
+		c := reader.read(mut buf)?
+		res << buf[..c]
+
+		match_len := match_array_in_array(buf[..c], sep)
+
+		if match_len == sep.len {
+			break
+		}
+
+		if match_len > 0 {
+			match_left := sep.len - match_len
+			c2 := reader.read(mut buf[..match_left])?
+			res << buf[..c2]
+
+			if buf[..c2] == sep[match_len..] {
+				break
+			}
+		}
+	}
+}
diff --git a/src/util/util.v b/src/util/util.v
index 9cf3011..266bcb5 100644
--- a/src/util/util.v
+++ b/src/util/util.v
@@ -1,13 +1,13 @@
 module util
 
 import os
-import io
 import crypto.md5
 import crypto.sha256
 
-const reader_buf_size = 1_000_000
-
-const prefixes = ['B', 'KB', 'MB', 'GB']
+const (
+	reader_buf_size = 1_000_000
+	prefixes        = ['B', 'KB', 'MB', 'GB']
+)
 
 // Dummy struct to work around the fact that you can only share structs, maps &
 // arrays
@@ -23,50 +23,6 @@ pub fn exit_with_message(code int, msg string) {
 	exit(code)
 }
 
-// reader_to_writer tries to consume the entire reader & write it to the writer.
-pub fn reader_to_writer(mut reader io.Reader, mut writer io.Writer) ? {
-	mut buf := []u8{len: 10 * 1024}
-
-	for {
-		bytes_read := reader.read(mut buf) or { break }
-		mut bytes_written := 0
-
-		for bytes_written < bytes_read {
-			c := writer.write(buf[bytes_written..bytes_read]) or { break }
-
-			bytes_written += c
-		}
-	}
-}
-
-// reader_to_file writes the contents of a BufferedReader to a file
-pub fn reader_to_file(mut reader io.BufferedReader, length int, path string) ? {
-	mut file := os.create(path)?
-	defer {
-		file.close()
-	}
-
-	mut buf := []u8{len: util.reader_buf_size}
-	mut bytes_left := length
-
-	// Repeat as long as the stream still has data
-	for bytes_left > 0 {
-		// TODO check if just breaking here is safe
-		bytes_read := reader.read(mut buf) or { break }
-		bytes_left -= bytes_read
-
-		mut to_write := bytes_read
-
-		for to_write > 0 {
-			// TODO don't just loop infinitely here
-			bytes_written := file.write(buf[bytes_read - to_write..bytes_read]) or { continue }
-			// file.flush()
-
-			to_write = to_write - bytes_written
-		}
-	}
-}
-
 // hash_file returns the md5 & sha256 hash of a given file
 // TODO actually implement sha256
 pub fn hash_file(path &string) ?(string, string) {
@@ -108,49 +64,3 @@ pub fn pretty_bytes(bytes int) string {
 
 	return '${n:.2}${util.prefixes[i]}'
 }
-
-// match_array_in_array<T> returns how many elements of a2 overlap with a1. For
-// example, if a1 = "abcd" & a2 = "cd", the result will be 2. If the match is
-// not at the end of a1, the result is 0.
-pub fn match_array_in_array<T>(a1 []T, a2 []T) int {
-	mut i := 0
-	mut match_len := 0
-
-	for i + match_len < a1.len {
-		if a1[i + match_len] == a2[match_len] {
-			match_len += 1
-		} else {
-			i += match_len + 1
-			match_len = 0
-		}
-	}
-
-	return match_len
-}
-
-// read_until_separator consumes an io.Reader until it encounters some
-// separator array. The data read is stored inside the provided res array.
-pub fn read_until_separator(mut reader io.Reader, mut res []u8, sep []u8) ? {
-	mut buf := []u8{len: sep.len}
-
-	for {
-		c := reader.read(mut buf)?
-		res << buf[..c]
-
-		match_len := match_array_in_array(buf[..c], sep)
-
-		if match_len == sep.len {
-			break
-		}
-
-		if match_len > 0 {
-			match_left := sep.len - match_len
-			c2 := reader.read(mut buf[..match_left])?
-			res << buf[..c2]
-
-			if buf[..c2] == sep[match_len..] {
-				break
-			}
-		}
-	}
-}