# Functions for parsing control files
#
# Copyright (C) 2012, 2018, 2019 Patrick McDermott
#
# This file is part of opkbuild.
#
# opkbuild is free software: you can redistribute it and/or modify it
# under the terms of the GNU General Public License as published by
# the Free Software Foundation, either version 3 of the License, or
# (at your option) any later version.
#
# opkbuild is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
# GNU General Public License for more details.
#
# You should have received a copy of the GNU General Public License
# along with opkbuild.  If not, see <http://www.gnu.org/licenses/>.

_OB_SUBSTVARS_MAX_DEPTH=50
_OB_SUBSTVAR_TRIM_SED='
	H;             # Store each input line in the hold space.
	${             # At the last line of input:
	  g;           # restore the hold space into the pattern space,
	  s/^[\n]*//;  # remove leading newline characters,
	  s/[\n]*$//;  # remove trailing newline characters, and
	  p;           # print the results.
	};
'

_ob_parse_control_error()
{
	local file="${1}"
	local line_nr="${2}"
	local msg_id="${3}"
	shift 3 || _ob_abort
	local file_info=

	if [ ${line_nr} -eq 0 ]; then
		file_info="$(printf '%s' "${file}")"
	else
		file_info="$(printf '%s(l%d)' "${file}" "${line_nr}")"
	fi

	_ob_warn_msg "${msg_id}" "${file_info}" "${@}"

	return 0
}

## @brief Parse a control file
## @details \fBob_parse_control\fP() parses a control file of field names and
##          values formatted like RFC 822 (or RFC 2822 or RFC 5322) headers. 
##          For each field, \fBob_parse_control\fP calls \fIfield_cb\fP with the
##          field name, the field value, and \fIuser_data\fP as arguments.  If
##          \fIreq_fields\fP and \fIopt_fields\fP are given,
##          \fBob_parse_control\fP verifies that the input control file contains
##          all fields listed in \fIreq_fields\fP and no fields that are listed
##          in neither \fIreq_fields\fP nor \fIopt_fields\fP.  If \fIfield_cb\fP
##          returns non-zero, \fBob_parse_control\fP stops parsing and
##          immediately returns, without verifying that all required fields were
##          found.
## @operand file       req The control file to parse, or "-" for standard input.
## @operand field_cb   req Callback to run for each field.  Must accept three
##                     arguments: the field name, the field value, and
##                     \fIuser_data\fP.
## @operand user_data  req Data to pass to \fIfield_cb\fP.
## @operand req_fields opt Required fields that must appear in the control file.
## @operand opt_fields opt Optional fields that may appear in the control file.
## @return Returns 0 after parsing.
## @stderr Prints error messages on parse errors.
## @pure maybe This function has no side effects.  Whether this function is
##             subshell-safe in practice depends on whether \fIfield_cb\fP is
##             subshell-safe.
ob_parse_control()
{
	local file="${1}"
	local field_cb="${2}"
	local user_data="${3}"
	shift 3 || _ob_abort
	local check_fields=
	local req_fields=
	local opt_fields=
	local got_fields=
	local line_nr=
	local line=
	local name=
	local value=
	local sep=

	check_fields='false'
	if [ ${#} -eq 2 ]; then
		req_fields=" ${1} "
		opt_fields=" ${2} "
		shift 2
		check_fields='true'
	fi

	got_fields=' '

	line_nr=0

	while IFS='' read -r line; do
		line_nr=$((${line_nr} + 1))
		case "${line}" in
			'')
				_ob_parse_control_error "${file}" "${line_nr}" \
					'control_empty_line'
				;;
			'#'*)  # Comment.
				;;
			[!\ ]*':'*)  # "Name: Value" line.
				if [ -n "${name}" ]; then
					if ! "${field_cb}" "${name}" "${value}"\
							"${user_data}"; then
						return 0
					fi
				fi
				IFS=': 	' read name value <<-EOF
					${line}
					EOF
				if [ -z "${name}" ]; then
					# Badly formatted control field.
					_ob_parse_control_error "${file}" \
						"${line_nr}" 'control_bad_nv'
					continue
				fi
				case "${got_fields}" in *" ${name} "*)
					# Duplicate field.
					_ob_parse_control_error \
						"${file}" "${line_nr}" \
						'control_duplicate_field' \
						"${name}"
					continue
				esac
				got_fields="${got_fields}${name} "
				case "${req_fields}" in *" ${name} "*)
					# Required field: remove from list.
					req_fields="${req_fields% ${name} *}$(:\
						) ${req_fields#* ${name} }"
					continue
				esac
				if ${check_fields}; then
					case "${opt_fields}" in *" ${name} "*)
						# Optional field.
						continue
					esac
					# Unknown field.
					_ob_parse_control_error \
						"${file}" "${line_nr}" \
						'control_unknown_field' \
						"${name}"
				fi
				;;
			' '*)  # Continuation line.
				if [ -z "${name}" ]; then
					# Expecting a "Name: Value" line.
					_ob_parse_control_error "${file}" \
						"${line_nr}" \
						'control_found_continuation'
					continue
				fi
				value="${value}${OB_LF}${line# }"
				;;
		esac
	done <<-EOF
		$(cat -- "${file}")
		EOF

	if [ -n "${name}" ]; then
		if ! "${field_cb}" "${name}" "${value}" "${user_data}"; then
			return 0
		fi
	fi

	if ${check_fields}; then
		case "${req_fields}" in *[!\ ]*)
			# Missing required control fields.
			sep="$(_ob_get_msg 'list_item_separator')"
			req_fields="$(printf "%s${sep}" ${req_fields}; \
				printf 'x')"
			req_fields="${req_fields%${sep}x}"
			_ob_parse_control_error "${file}" '0' \
				'control_missing_fields' "${req_fields}"
		esac
	fi

	return 0
}

## @brief Set a source package substitution variable
## @details \fBob_set_source_substvar\fP() sets a source package substitution
##          variable for later use by \fBob_substvars\fP(3).
## @operand name  req The name of the substitution variable.  May only consist
##                    of uppercase and lowercase Latin letters, digits, and
##                    hyphens and must be at least one character long.
## @operand value req The value of the substitution variable.
## @return Returns 0 on success, or 1 if \fIname\fP is empty or contains invalid
##         characters.
## @pure no This function sets an internal global variable.
ob_set_source_substvar()
{
	local name="${1}"
	local value="${2}"
	shift 2 || _ob_abort

	# Validate variable name and convert to lower case.
	case "${name}" in *[!A-Za-z0-9-]* | '')
		return 1
	esac
	name="$(printf '%s' "${name}" | tr 'A-Z-' 'a-z_')"

	# Trim leading and trailing whitespace from value.
	value="$(printf '%s' "${value}" | sed -n "${_OB_SUBSTVAR_TRIM_SED}")"

	eval "_ob_substvar_src_${name}=\"\${value}\""

	return 0
}

## @brief Set a binary package substitution variable
## @details \fBob_set_source_substvar\fP() sets a binary package substitution
##          variable for later use by \fBob_substvars\fP(3).
## @operand name    req The name of the substitution variable.  May only consist
##                      of uppercase and lowercase Latin letters, digits, and
##                      hyphens and must be at least one character long.
## @operand value   req The value of the substitution variable.
## @operand package req The package for which to set the substitution variable.
## @return Returns 0 on success, or 1 if \fIname\fP is empty or contains invalid
##         characters.
## @pure no This function sets an internal global variable.
ob_set_binary_substvar()
{
	local name="${1}"
	local value="${2}"
	local package="${3}"
	shift 3 || _ob_abort

	# Validate variable name and convert to lower case.
	case "${name}" in *[!A-Za-z0-9-]* | '')
		return 1
	esac
	name="$(printf '%s' "${name}" | tr 'A-Z-' 'a-z_')"

	# Trim leading and trailing whitespace from value.
	value="$(printf '%s' "${value}" | sed -n "${_OB_SUBSTVAR_TRIM_SED}")"

	# Convert package name to clean form.
	package="$(printf '%s' "${package}" | tr 'A-Z' 'a-z' | \
		tr -C 'a-z0-9' '_')"

	# TODO: Maybe disallow variable names beginning with hyphens, and/or
	# otherwise make this safer.
	eval "_ob_substvar_bin_${package}__${name}=\"\${value}\""

	return 0
}

## @brief Substitute variables in text
## @details \fBob_substvars\fP() substitutes variables previously set with
##          \fBob_set_source_substvar\fP(3) in a string.  The format for
##          variable substitutions is \fI${var}\fP, and substitutions can be
##          nested.
## @operand string  req The string in which to substitute variables.
## @operand package opt The binary package for which to substitute variables.
## @return Returns 0 on success or 1 on possible recursion.
## @stderr Prints a warning on possible recursion.
## @pure yes This function has no side effects.
ob_substvars()
{
	local string="${1}"
	shift 1 || _ob_abort
	local check_bin=
	local package=
	local depth=
	local lhs=
	local name=
	local rhs=
	local old_rhs=
	local name_tr=
	local value=

	check_bin=false
	if [ ${#} -eq 1 ]; then
		package="${1}"
		shift 1
		check_bin=true
	fi

	# Logic inspired by that of dpkg's Dpkg::Substvars::substvars()
	# subroutine.

	depth=0

	while :; do
		lhs="${string%%\$\{*}"
		if [ ${#lhs} -eq ${#string} ]; then
			# No "${" was found.
			break
		fi
		string="${string#*\$\{}"
		name="${string%%\}*}"
		rhs="${string#*\}}"

		if [ ${#rhs} -lt ${#old_rhs} ]; then
			# Reset the nesting counter if we've advanced the right
			# side of the matched space.
			depth=0
		fi
		if [ ${depth} -ge ${_OB_SUBSTVARS_MAX_DEPTH} ]; then
			# Warn of possible recursion.
			_ob_error_msg 'substvar_deep_nesting'
			return 1
		fi
		old_rhs="${rhs}"

		# Validate the variable name.
		case "${name}" in *[!A-Za-z0-9-]* | '')
			_ob_warn_msg 'substvar_invalid' "${name}"
			# Remove the variable expansion altogether.  We can't
			# just leave the variable unexpanded, because the
			# presence of the "${" characters would cause this
			# parser to loop infinitely.
			string="${lhs}${rhs}"
			continue
		esac

		# Perform the substitution.
		name_tr="$(printf '%s' "${name}" | tr 'A-Z-' 'a-z_')"
		package="$(printf '%s' "${package}" | tr 'A-Z' 'a-z' | \
			tr -C 'a-z0-9' '_')"
		if eval "[ x\"\${_ob_substvar_src_${name_tr}:+set}\" = x'set' ]"
		then
			name_tr="_ob_substvar_src_${name_tr}"
		elif ${check_bin} && eval "[ x\"\${_ob_substvar_bin_$(: \
				)${package}__${name_tr}:+set}\" = x'set' ]"; then
			name_tr="_ob_substvar_bin_${package}__${name_tr}"
		else
			_ob_warn_msg 'substvar_unknown' "${name}"
			string="${lhs}${rhs}"
			continue
		fi
		value="$(eval "printf '%s' \"\${${name_tr}}\"")"
		string="${lhs}${value}${rhs}"
		depth=$((${depth} + 1))
	done

	printf '%s\n' "${string}"

	return 0
}