sas-linter 0.2.4 → 0.2.6

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: b10cf15e62fe7e0a81d8f9d2e2c9eb61af36ee00d10f32e9ce8354a7ebe3e40f
-  data.tar.gz: 9abe8249ab9d968fdc8aefe558dcdad321ba2879abe4bf08374b9ec998eecb5f
+  metadata.gz: 688281e326c33b3af52864a9d25c85f00967a2c4f8ac9233e57cf52a83407b18
+  data.tar.gz: a9f331a62f73a5bdb7b637bbd62783f3969b169746e7035adf687afc3030ceb4
 SHA512:
-  metadata.gz: 38e745b6f3ec4c8a7cc2ce3c23883569bcd931d88d6f9dc0a6c2288ed7f0b39f199e066340e780e8556d0ceeaa7bfeb149185ef2d458f58574281c54dbbf60d3
-  data.tar.gz: 8e6b5f2751eb5b3b9e50b80f0476f953519763a321fba4dd17ce219d891a84a3f0fe3c1ea1fbc3c18f4a698bc0655ef794ff2ef71d1d859aa23195a9119d41bd
+  metadata.gz: d952062a88c0bca6c33ef3f3e92c6cfeeec3f2142053e47c71dde3c04202ba0209b4672cad326161d62d05df90fb0940665cd0f65b73a62412d97b0ff6c8a1da
+  data.tar.gz: 7c41bac13b275a4dc9fc8bbf374339595b2b185fb6058e3ea27b412b18124a44bdb5b901eec7bdc9c0d7db3d9239ea0d56a30e6896ad88f0b9386dc74c2b1687

data/README.md

@@ -1,6 +1,6 @@
 # sas-linter
 
-A configurable lint engine for SAS source files. Built on the [`sas-lexer`](https://github.com/mes-amis/sas-lexer-rb) gem (a Ruby FFI binding to Misha Perlov's Rust [`sas-lexer`](https://github.com/mishamsk/sas-lexer)) and ships with fourteen pluggable rules covering structural defects, cosmetic issues, and source-header conventions.
+A configurable lint engine for SAS source files. Built on the [`sas-lexer`](https://github.com/mes-amis/sas-lexer-rb) gem (a Ruby FFI binding to Misha Perlov's Rust [`sas-lexer`](https://github.com/mishamsk/sas-lexer)) and ships with fifteen pluggable rules covering structural defects, cosmetic issues, and source-header conventions.
 
 ## Installation
 
@@ -63,6 +63,10 @@ rules:
     enabled: true
     autofix: false             # rule supports autofix; off by default
 
+  unterminated_comment:
+    enabled: true
+    autofix: false             # append `;` to `**…**` lines whose missing terminator eats the next line
+
   inconsistent_variable_case:
     enabled: true
     autofix: false             # rewrite every minority casing to the most-common form
@@ -141,6 +145,7 @@ findings = linter.lint_file("path/to/source.sas")
 | `encoding_issues` | Smart-quote / em-dash / Win-1252 byte sequences that confuse downstream tooling. |
 | `malformed_if_condition` | Empty conditions, missing operators, orphan `then`, unbalanced parens, etc. |
 | `missing_assignment_semicolon` | Assignment statements followed by an inline `**` comment but no terminating `;`. |
+| `unterminated_comment` | A standalone `** … **` comment whose missing `;` lets the SAS lexer extend the comment into the next line of real code, silently swallowing it. Autofix appends the `;`. |
 | `variable_value_out_of_known_range` | `if VAR = N` / `if VAR in (...)` literals fall outside the variable's documented acceptable values. Loads the catalog from one or more CSVs with configurable column names and column separator (`,`, `;`, tab). |
 | `inconsistent_variable_case` | Identifier appears with more than one casing in the same file (`myVar` vs `MyVar`). SAS treats both as the same variable; autofix rewrites every minority spelling to the most-common form. Skips proc-format definitions and `format.` / `lib.member` references. |
 | `format_for_unknown_variable` | `format` / `informat` / `attrib` statement assigns a format to a variable that's referenced nowhere else in the file — almost always a typo. Skipped on files that pull in columns via `set` / `merge` / `update` / `infile` / `input`. |

data/lib/sas_linter/rules/malformed_label_statement.rb

@@ -0,0 +1,127 @@
+# frozen_string_literal: true
+
+require_relative "../../sas_linter"
+require "sas_lexer"
+
+class SasLinter
+  module Rules
+    # Flag `label` statements where the `=` between the variable name
+    # and the string literal is missing.
+    #
+    # Motivating bug: `label aHSDELIRIUM 'Delirium Screener';` in
+    # SUITE9_HS_DELIRIUM_SCREENER_2014-04-15.TXT — SAS rejects this
+    # with "ERROR 22-322: Syntax error, expecting one of the following:
+    # =, ?", and the label is silently never attached. The treatment
+    # variant of the same algorithm shipped the same typo, and several
+    # other interRAI sources have shipped it over time.
+    #
+    # Detection: every `label` statement is a `KW_LABEL` keyword
+    # followed by one or more `IDENT '=' STRING_LITERAL` triples
+    # separated by whitespace, terminated by `;`. We walk each label
+    # statement and, for each IDENT inside it, require the next
+    # default-channel token to be `ASSIGN` (`=`). If instead the next
+    # token is a STRING_LITERAL, the `=` was dropped.
+    class MalformedLabelStatement < Rule
+      rule_id :malformed_label_statement
+      description "`label` statement missing `=` between variable and string literal."
+      severity :warning
+
+      TT = SasLexer::Lexer::TokenType
+
+      def self.supports_autofix?
+        true
+      end
+
+      def check(tokens, path:, all_tokens: nil, source: nil) # rubocop:disable Lint/UnusedMethodArgument
+        findings = []
+        each_label_violation(tokens) do |ident_t, string_t|
+          findings << finding(
+            line: ident_t[:start_line],
+            column: ident_t[:start_column] + 1,
+            message: "`label #{ident_t[:text]} #{shorten(string_t[:text])}` is missing the `=` " \
+                     "between the variable name and the label string.",
+            path: path
+          )
+        end
+        findings
+      end
+
+      def autofix(source)
+        return source if source.nil? || source.empty?
+
+        lexer = SasLexer::Lexer.new
+        begin
+          all_tokens = lexer.tokenize(source)
+        ensure
+          lexer.free
+        end
+        tokens = all_tokens.reject do |t|
+          t[:channel] == SasLexer::Lexer::TokenChannel::HIDDEN ||
+            t[:channel] == SasLexer::Lexer::TokenChannel::COMMENT
+        end
+
+        source_lines = source.split("\n", -1)
+        # Collect (line_idx, col_after_ident) for each malformed label,
+        # then apply edits right-to-left within each line so earlier
+        # column offsets stay valid.
+        edits_by_line = Hash.new { |h, k| h[k] = [] }
+
+        each_label_violation(tokens) do |ident_t, _string_t|
+          edits_by_line[ident_t[:start_line] - 1] << ident_t[:end_column]
+        end
+
+        edits_by_line.each do |line_idx, cols|
+          line = source_lines[line_idx]
+          next if line.nil?
+
+          # Right-to-left so earlier insertions don't shift later columns.
+          cols.sort.reverse.each do |col|
+            # Insert ` =` immediately after the IDENT (consuming the
+            # following space if there is one, preserving alignment).
+            replacement =
+              if col < line.length && line[col] == " "
+                # `aHSDELIRIUM 'Delirium Screener'` → `aHSDELIRIUM = 'Delirium Screener'`.
+                # Replace the single space with ` = ` (one space before, one after).
+                " = #{line[(col + 1)..]}"
+              else
+                " = #{line[col..]}"
+              end
+            line = "#{line[0...col]}#{replacement}"
+          end
+          source_lines[line_idx] = line
+        end
+
+        source_lines.join("\n")
+      end
+
+      private
+
+      def each_label_violation(tokens)
+        tokens.each_with_index do |t, i|
+          next unless t[:type] == TT::KW_LABEL
+
+          # Walk forward through the label statement until SEMI.
+          j = i + 1
+          while j < tokens.length && tokens[j][:type] != TT::SEMI
+            cur = tokens[j]
+            if cur[:type] == TT::IDENTIFIER
+              nxt = tokens[j + 1]
+              if nxt && nxt[:type] == TT::STRING_LITERAL
+                yield cur, nxt
+                j += 2
+                next
+              end
+            end
+            j += 1
+          end
+        end
+      end
+
+      def shorten(text)
+        return text if text.nil? || text.length <= 40
+
+        "#{text[0, 37]}..."
+      end
+    end
+  end
+end

data/lib/sas_linter/rules/unterminated_comment.rb

@@ -0,0 +1,101 @@
+# frozen_string_literal: true
+
+require_relative "../../sas_linter"
+require "sas_lexer"
+
+class SasLinter
+  module Rules
+    # Flag a `** ... **` comment line whose missing `;` causes the
+    # SAS lexer to extend the comment statement into the following
+    # line(s) of real code, silently swallowing it.
+    #
+    # SAS `*` / `**` comment statements are terminated by the next
+    # `;` — the closing `**` is just prose. So
+    #
+    #     ** SOME COMMENT **
+    #     y = x + 1;
+    #
+    # lexes as a single comment token covering both lines, and the
+    # `y = x + 1;` assignment never executes.
+    #
+    # Detection: a COMMENT-channel `PREDICTED_COMMENT_STAT` token
+    # whose `start_line != end_line` AND whose first source line,
+    # rstripped, ends with `**`. That shape is the boxed-comment
+    # closer the user clearly intended — they only forgot the `;`.
+    # Legitimate multi-line `*...;` prose ends its first line with
+    # plain text, not `**`, so it's left alone.
+    #
+    # Autofix: append `;` to the end of each flagged first line.
+    class UnterminatedComment < Rule
+      rule_id :unterminated_comment
+      description "`**` comment missing its terminating `;` — consumes following code."
+      severity :warning
+
+      TT = SasLexer::Lexer::TokenType
+      COMMENT_CHANNEL = SasLexer::Lexer::TokenChannel::COMMENT
+
+      def self.supports_autofix?
+        true
+      end
+
+      def check(_tokens, path:, all_tokens: nil, source: nil)
+        return [] unless all_tokens && source
+
+        lines = source.split("\n", -1)
+        unterminated_comment_lines(all_tokens, lines).map do |i|
+          finding_for_line(lines[i], i, path)
+        end
+      end
+
+      def autofix(source)
+        return source if source.nil? || source.empty?
+
+        lines = source.split("\n", -1)
+        bad = unterminated_comment_lines(tokenize(source), lines)
+        return source if bad.empty?
+
+        bad.each do |i|
+          lines[i] = "#{lines[i].rstrip};"
+        end
+        lines.join("\n")
+      end
+
+      private
+
+      def finding_for_line(line, idx, path)
+        finding(
+          line: idx + 1,
+          column: line.length - line.lstrip.length + 1,
+          message: "`**` comment missing `;` — consumes the next line of code as comment text.",
+          path: path
+        )
+      end
+
+      def tokenize(source)
+        lexer = SasLexer::Lexer.new
+        begin
+          lexer.tokenize(source)
+        ensure
+          lexer.free
+        end
+      end
+
+      # 0-indexed source lines that hold a `** ... **` comment whose
+      # missing `;` made the lexer extend it into the next line.
+      def unterminated_comment_lines(all_tokens, lines)
+        bad = []
+        all_tokens.each do |t|
+          next unless t[:channel] == COMMENT_CHANNEL
+          next unless t[:type] == TT::PREDICTED_COMMENT_STAT
+          next unless t[:start_line] < t[:end_line]
+
+          first = lines[t[:start_line] - 1] or next
+          next unless first.rstrip.end_with?("**")
+
+          bad << (t[:start_line] - 1)
+        end
+        bad
+      end
+    end
+  end
+end

data/lib/sas_linter/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 class SasLinter
-  VERSION = "0.2.4"
+  VERSION = "0.2.6"
 end

data/lib/sas_linter.rb

@@ -307,7 +307,9 @@ require_relative "sas_linter/rules/line_endings"
 require_relative "sas_linter/rules/encoding_issues"
 require_relative "sas_linter/rules/malformed_if_condition"
 require_relative "sas_linter/rules/missing_assignment_semicolon"
+require_relative "sas_linter/rules/malformed_label_statement"
 require_relative "sas_linter/rules/variable_value_out_of_known_range"
 require_relative "sas_linter/rules/invalid_numeric_literal"
 require_relative "sas_linter/rules/inconsistent_variable_case"
 require_relative "sas_linter/rules/format_for_unknown_variable"
+require_relative "sas_linter/rules/unterminated_comment"

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: sas-linter
 version: !ruby/object:Gem::Version
-  version: 0.2.4
+  version: 0.2.6
 platform: ruby
 authors:
 - Craig McNamara
@@ -69,11 +69,13 @@ files:
 - lib/sas_linter/rules/invalid_numeric_literal.rb
 - lib/sas_linter/rules/line_endings.rb
 - lib/sas_linter/rules/malformed_if_condition.rb
+- lib/sas_linter/rules/malformed_label_statement.rb
 - lib/sas_linter/rules/missing_assignment_semicolon.rb
 - lib/sas_linter/rules/source_headers.rb
 - lib/sas_linter/rules/tab_expansion.rb
 - lib/sas_linter/rules/trailing_whitespace.rb
 - lib/sas_linter/rules/unreachable_inner_branch_value.rb
+- lib/sas_linter/rules/unterminated_comment.rb
 - lib/sas_linter/rules/variable_value_out_of_known_range.rb
 - lib/sas_linter/version.rb
 homepage: https://github.com/mes-amis/sas-linter