gitlab-org--gitlab-foss/lib/gitlab/git/diff.rb

251 lines
7.6 KiB
Ruby
Raw Normal View History

# frozen_string_literal: true
2017-01-04 13:43:06 -05:00
module Gitlab
module Git
class Diff
TimeoutError = Class.new(StandardError)
include Gitlab::EncodingHelper
2017-01-04 13:43:06 -05:00
# Diff properties
attr_accessor :old_path, :new_path, :a_mode, :b_mode, :diff
# Stats properties
attr_accessor :new_file, :renamed_file, :deleted_file
alias_method :new_file?, :new_file
alias_method :deleted_file?, :deleted_file
alias_method :renamed_file?, :renamed_file
attr_accessor :expanded
attr_writer :too_large
2017-01-04 13:43:06 -05:00
2017-06-06 17:28:06 -04:00
alias_method :expanded?, :expanded
# The default maximum content size to display a diff patch.
#
# If this value ever changes, make sure to create a migration to update
# current records, and default of `ApplicationSettings#diff_max_patch_bytes`.
DEFAULT_MAX_PATCH_BYTES = 100.kilobytes
2017-05-30 15:50:02 -04:00
# This is a limitation applied on the source (Gitaly), therefore we don't allow
# persisting limits over that.
MAX_PATCH_BYTES_UPPER_BOUND = 500.kilobytes
SERIALIZE_KEYS = %i(diff new_path old_path a_mode b_mode new_file renamed_file deleted_file too_large).freeze
2017-01-04 13:43:06 -05:00
class << self
2017-01-04 13:43:06 -05:00
def between(repo, head, base, options = {}, *paths)
straight = options.delete(:straight) || false
common_commit = if straight
base
else
# Only show what is new in the source branch
# compared to the target branch, not the other way
# around. The linex below with merge_base is
# equivalent to diff with three dots (git diff
# branch1...branch2) From the git documentation:
# "git diff A...B" is equivalent to "git diff
# $(git-merge-base A B) B"
repo.merge_base(head, base)
2017-01-04 13:43:06 -05:00
end
options ||= {}
actual_options = filter_diff_options(options)
repo.diff(common_commit, head, actual_options, *paths)
end
# Return a copy of the +options+ hash containing only recognized keys.
# Allowed options are:
2017-01-04 13:43:06 -05:00
#
# :ignore_whitespace_change ::
# If true, changes in amount of whitespace will be ignored.
#
# :max_files ::
# Limit how many files will patches be allowed for before collapsing
#
# :max_lines ::
# Limit how many patch lines (across all files) will be allowed for
# before collapsing
2017-01-04 13:43:06 -05:00
#
# :limits ::
# A hash with additional limits to check before collapsing patches.
# Allowed keys are: `max_bytes`, `safe_max_files`, `safe_max_lines`
# and `safe_max_bytes`
#
# :expanded ::
# If false, patch raw data will not be included in the diff after
# `max_files`, `max_lines` or any of the limits in `limits` are
# exceeded
2017-01-04 13:43:06 -05:00
def filter_diff_options(options, default_options = {})
allowed_options = [:ignore_whitespace_change, :max_files, :max_lines,
:limits, :expanded]
2017-01-04 13:43:06 -05:00
if default_options
actual_defaults = default_options.dup
actual_defaults.keep_if do |key|
allowed_options.include?(key)
end
else
actual_defaults = {}
end
if options
filtered_opts = options.dup
filtered_opts.keep_if do |key|
allowed_options.include?(key)
end
filtered_opts = actual_defaults.merge(filtered_opts)
else
filtered_opts = actual_defaults
end
filtered_opts
end
2017-09-03 07:45:44 -04:00
# Return a binary diff message like:
#
2017-09-03 07:45:44 -04:00
# "Binary files a/file/path and b/file/path differ\n"
2017-09-05 13:16:08 -04:00
# This is used when we detect that a diff is binary
# using CharlockHolmes.
2017-09-03 07:45:44 -04:00
def binary_message(old_path, new_path)
"Binary files #{old_path} and #{new_path} differ\n"
end
# Returns the limit of bytes a single diff file can reach before it
# appears as 'collapsed' for end-users.
# By convention, it's 10% of the persisted `diff_max_patch_bytes`.
#
# Example: If we have 100k for the `diff_max_patch_bytes`, it will be 10k by
# default.
#
# Patches surpassing this limit should still be persisted in the database.
def patch_safe_limit_bytes
patch_hard_limit_bytes / 10
end
# Returns the limit for a single diff file (patch).
#
# Patches surpassing this limit shouldn't be persisted in the database
# and will be presented as 'too large' for end-users.
def patch_hard_limit_bytes
Gitlab::CurrentSettings.diff_max_patch_bytes
end
2017-01-04 13:43:06 -05:00
end
def initialize(raw_diff, expanded: true)
@expanded = expanded
2017-01-04 13:43:06 -05:00
case raw_diff
when Hash
2017-02-24 10:53:44 -05:00
init_from_hash(raw_diff)
prune_diff_if_eligible
2017-05-30 15:30:05 -04:00
when Gitlab::GitalyClient::Diff
2017-02-24 10:53:44 -05:00
init_from_gitaly(raw_diff)
prune_diff_if_eligible
when Gitaly::CommitDelta
init_from_gitaly(raw_diff)
2017-01-04 13:43:06 -05:00
when nil
raise "Nil as raw diff passed"
else
raise "Invalid raw diff type: #{raw_diff.class}"
end
end
def to_hash
hash = {}
SERIALIZE_KEYS.each do |key|
hash[key] = send(key) # rubocop:disable GitlabSecurity/PublicSend
2017-01-04 13:43:06 -05:00
end
hash
end
def mode_changed?
a_mode && b_mode && a_mode != b_mode
end
2017-01-04 13:43:06 -05:00
def submodule?
a_mode == '160000' || b_mode == '160000'
end
def line_count
@line_count ||= Util.count_lines(@diff)
end
def too_large?
2017-05-30 22:48:30 -04:00
if @too_large.nil?
@too_large = @diff.bytesize >= self.class.patch_hard_limit_bytes
2017-05-30 22:48:30 -04:00
else
@too_large
end
2017-01-04 13:43:06 -05:00
end
# This is used by `to_hash` and `init_from_hash`.
alias_method :too_large, :too_large?
def too_large!
2017-01-04 13:43:06 -05:00
@diff = ''
@line_count = 0
@too_large = true
end
def collapsed?
return @collapsed if defined?(@collapsed)
@collapsed = !expanded && @diff.bytesize >= self.class.patch_safe_limit_bytes
2017-01-04 13:43:06 -05:00
end
def collapse!
2017-01-04 13:43:06 -05:00
@diff = ''
@line_count = 0
@collapsed = true
end
def json_safe_diff
2017-09-05 13:16:08 -04:00
return @diff unless detect_binary?(@diff)
# the diff is binary, let's make a message for it
Diff.binary_message(@old_path, @new_path)
end
def has_binary_notice?
@diff.start_with?('Binary')
end
2017-01-04 13:43:06 -05:00
private
2017-02-24 10:53:44 -05:00
def init_from_hash(hash)
2017-01-04 13:43:06 -05:00
raw_diff = hash.symbolize_keys
SERIALIZE_KEYS.each do |key|
send(:"#{key}=", raw_diff[key.to_sym]) # rubocop:disable GitlabSecurity/PublicSend
2017-01-04 13:43:06 -05:00
end
2017-02-24 10:53:44 -05:00
end
2017-05-30 15:30:05 -04:00
def init_from_gitaly(diff)
2017-06-21 19:51:46 -04:00
@diff = encode!(diff.patch) if diff.respond_to?(:patch)
2017-05-30 15:30:05 -04:00
@new_path = encode!(diff.to_path.dup)
@old_path = encode!(diff.from_path.dup)
@a_mode = diff.old_mode.to_s(8)
@b_mode = diff.new_mode.to_s(8)
@new_file = diff.from_id == BLANK_SHA
@renamed_file = diff.from_path != diff.to_path
@deleted_file = diff.to_id == BLANK_SHA
@too_large = diff.too_large if diff.respond_to?(:too_large)
collapse! if diff.respond_to?(:collapsed) && diff.collapsed
2017-02-24 10:53:44 -05:00
end
2017-01-04 13:43:06 -05:00
def prune_diff_if_eligible
if too_large?
too_large!
elsif collapsed?
collapse!
end
2017-01-04 13:43:06 -05:00
end
end
end
end