lua,perf - avoid string.match() performance issue

Author: cscheid

src/resources/filters/common/tables.lua

@@ -131,7 +131,7 @@ end
 function hasRawLatexTable(raw)
   if _quarto.format.isRawLatex(raw) and _quarto.format.isLatexOutput() then
     for i,pattern in ipairs(_quarto.patterns.latexTablePatterns) do
-      if raw.text:match(pattern) then
+      if _quarto.modules.patterns.match_all_in_table(pattern)(raw.text) then
         return true
       end
     end

src/resources/filters/crossref/#crossref.lua#

@@ -444,12 +444,16 @@ end, function(float)
     local made_fix = false
     local function fix_raw(is_star_env)
       local function set_raw(el)
-        if _quarto.format.isRawLatex(el) and el.text:match(_quarto.patterns.latexLongtablePattern) then
+        if _quarto.format.isRawLatex(el) and _quarto.modules.patterns.match_all_in_table(_quarto.patterns.latexLongtablePattern)(el.text) then
           made_fix = true
           local raw = el
           -- special case for longtable floats in LaTeX
-          local extended_pattern = "(.-)" .. _quarto.patterns.latexLongtablePattern .. "(.*)"
-          local longtable_preamble, longtable_begin, longtable_content, longtable_end, longtable_postamble = raw.text:match(extended_pattern)
+          local extended_pattern = {".-"}
+          for _, pattern in ipairs(_quarto.patterns.latexLongtablePattern) do
+            table.insert(extended_pattern, pattern)
+          end
+          table.insert(extended_pattern, ".*")
+          local longtable_preamble, longtable_begin, longtable_content, longtable_end, longtable_postamble = _quarto.modules.patterns.match_all_in_table(extended_pattern)(raw.text)
           if longtable_preamble == nil or longtable_begin == nil or longtable_content == nil or longtable_end == nil or longtable_postamble == nil then
             warn("Could not parse longtable parameters. This could happen because the longtable parameters\n" ..
             "are not well-formed or because of a bug in quarto. Please consider filing a bug report at\n" ..

src/resources/filters/customnodes/floatreftarget.lua

@@ -42,6 +42,29 @@ local latex_tabular = "(\\begin{tabular}.*\\end{tabular})"
 local latex_table = "(\\begin{table})(.*)(\\end{table})"
 local latex_table_star = "(\\begin{table%*})(.*)(\\end{table%*})"
 
+local function combine_patterns(pattern_table)
+  local combined_pattern = {}
+  for i, v in ipairs(pattern_table) do
+    table.insert(combined_pattern, "(" .. v .. ")")
+  end
+  return table.concat(combined_pattern)
+end
+
+-- see https://github.com/quarto-dev/quarto-cli/issues/9729#issuecomment-2122907870
+-- for why this is necessary.
+local function match_all_in_table(pattern_table)
+  local function inner(text)
+    for i, v in ipairs(pattern_table) do
+      if text:match(v) == nil then
+        return nil
+      end
+    end
+    -- return the combined matches for the combined pattern
+    return text:match(combine_patterns(pattern_table))
+  end
+  return inner
+end
+
 return {
   attr_identifier = attr_identifier,
   engine_escape = engine_escape,
@@ -68,4 +91,7 @@ return {
   latex_tabular = latex_tabular,
   latex_table = latex_table,
   latex_table_star = latex_table_star,
+
+  match_all_in_table = match_all_in_table,
+  combine_patterns = combine_patterns
 }

src/resources/filters/modules/patterns.lua

@@ -72,6 +72,12 @@ function extract_latex_quartomarkdown_commands()
           return nil
         end
         local text = el.text
+        -- provide an early exit if the text does not contain the pattern
+        -- because Lua's pattern matching apparently takes a long time
+        -- to fail: https://github.com/quarto-dev/quarto-cli/issues/9729
+        if text:match("\\QuartoMarkdownBase64{") == nil then
+          return nil
+        end
         local pattern = "(.*)(\\QuartoMarkdownBase64{)([^}]*)(})(.*)"
         local pre, _, content, _, post = text:match(pattern)
         if pre == nil then
@@ -103,12 +109,17 @@ function inject_vault_content_into_rawlatex()
         return nil
         -- luacov: enable
       end
+      local text = el.text
+      -- provide an early exit if the text does not contain the pattern
+      -- because Lua's pattern matching apparently takes a long time
+      -- to fail: https://github.com/quarto-dev/quarto-cli/issues/9729
+      if el.text:match("3ab579b5%-63b4%-445d%-bc1d%-85bf6c4c04de") == nil then
+        return nil
+      end
 
       local pattern = "(.*)(3ab579b5%-63b4%-445d%-bc1d%-85bf6c4c04de%-[0-9]+)(.*)"
-      local text = el.text
       local pre, content_id, post = text:match(pattern)
 
-
       while pre do
         local found = false
         vault.content = _quarto.ast.walk(vault.content, {

src/resources/filters/normalize/extractquartodom.lua

@@ -56,8 +56,10 @@ function compute_flags()
       end
 
       if _quarto.format.isRawLatex(el) then
-        if (el.text:match(_quarto.patterns.latexLongtablePattern) and
-            not el.text:match(_quarto.patterns.latexCaptionPattern)) then
+        local long_table_match = _quarto.modules.patterns.match_all_in_table(_quarto.patterns.latexLongtablePattern)
+        local caption_match = _quarto.modules.patterns.match_all_in_table(_quarto.patterns.latexCaptionPattern)
+        if (long_table_match(el.text) and
+            not caption_match(el.text)) then
             flags.has_longtable_no_caption_fixup = true
         end
       end

src/resources/filters/normalize/flags.lua

@@ -436,10 +436,11 @@ function render_latex_fixups()
   return {
     RawBlock = function(raw)
       if _quarto.format.isRawLatex(raw) then
-        if (raw.text:match(_quarto.patterns.latexLongtablePattern) and
-            not raw.text:match(_quarto.patterns.latexCaptionPattern)) then
+        local long_table_match = _quarto.modules.patterns.match_all_in_table(_quarto.patterns.latexLongtablePattern)
+        local caption_match = _quarto.modules.patterns.match_all_in_table(_quarto.patterns.latexCaptionPattern)
+        if long_table_match(raw.text) and caption_match(raw.text) then
           raw.text = raw.text:gsub(
-            _quarto.patterns.latexLongtablePattern, "\\begin{longtable*}%2\\end{longtable*}", 1)
+            _quarto.modules.patterns.combine_patterns(_quarto.patterns.latexLongtablePattern), "\\begin{longtable*}%2\\end{longtable*}", 1)
           return raw
         end
       end

src/resources/filters/quarto-post/latex.lua

@@ -26,11 +26,14 @@ function table_captions()
                 el = _quarto.ast.walk(el, {
                   RawBlock = function(raw)
                     if _quarto.format.isRawLatex(raw) then
-                      if raw.text:match(_quarto.patterns.latexTabularPattern) and not raw.text:match(_quarto.patterns.latexTablePattern) then
-                        raw.text = raw.text:gsub(_quarto.patterns.latexTabularPattern, 
-                                                "\\begin{table}\n\\centering\n%1%2%3\n\\end{table}\n",
-                                                1)
-                        return raw                       
+                      local tabular_match = _quarto.modules.patterns.match_all_in_table(_quarto.patterns.latexTabularPattern)
+                      local table_match = _quarto.modules.patterns.match_all_in_table(_quarto.patterns.latexTablePattern)
+                      if tabular_match(raw.text) and not table_match(raw.text) then
+                        raw.text = raw.text:gsub(
+                          _quarto.modules.patterns.combine_patterns(_quarto.patterns.latexTabularPattern),
+                          "\\begin{table}\n\\centering\n%1%2%3\n\\end{table}\n",
+                          1)
+                        return raw
                       end
                     end
                   end
@@ -169,8 +172,10 @@ function applyTableCaptions(el, tblCaptions, tblLabels)
           idx = idx + 1
         elseif hasRawLatexTable(raw) then
           for i,pattern in ipairs(_quarto.patterns.latexTablePatterns) do
-            if raw.text:match(pattern) then
-              raw.text = applyLatexTableCaption(raw.text, tblCaptions[idx], tblLabels[idx], pattern)
+            local match_fun = _quarto.modules.patterns.match_all_in_table(pattern)
+            if match_fun(raw.text) then
+              local combined_pattern = _quarto.modules.patterns.combine_patterns(pattern)
+              raw.text = applyLatexTableCaption(raw.text, tblCaptions[idx], tblLabels[idx], combined_pattern)
               break
             end
           end
@@ -198,20 +203,22 @@ end
 
 function applyLatexTableCaption(latex, tblCaption, tblLabel, tablePattern)
   local latexCaptionPattern = _quarto.patterns.latexCaptionPattern
+  local latex_caption_match = _quarto.modules.patterns.match_all_in_table(latexCaptionPattern)
   -- insert caption if there is none
-  local beginCaption, caption = latex:match(latexCaptionPattern)
+  local beginCaption, caption = latex_caption_match(latex)
   if not beginCaption then
     latex = latex:gsub(tablePattern, "%1" .. "\n\\caption{ }\\tabularnewline\n" .. "%2%3", 1)
   end
   -- apply table caption and label
-  local beginCaption, captionText, endCaption = latex:match(latexCaptionPattern)
+  local beginCaption, captionText, endCaption = latex_caption_match(latex)
   if #tblCaption > 0 then
     captionText = stringEscape(tblCaption, "latex")
   end
   if #tblLabel > 0 then
     captionText = captionText .. " {#" .. tblLabel .. "}"
   end
-  latex = latex:gsub(latexCaptionPattern, "%1" .. captionText:gsub("%%", "%%%%") .. "%3", 1)
+  assert(captionText)
+  latex = latex:gsub(_quarto.modules.patterns.combine_patterns(latexCaptionPattern), "%1" .. captionText:gsub("%%", "%%%%") .. "%3", 1)
   return latex
 end
 

src/resources/filters/quarto-pre/table-captions.lua

@@ -1705,32 +1705,32 @@ local function resolveServiceWorkers(serviceworkers)
    else
      return nil
    end
- end
+end
 
 
-local latexTableWithOptionsPattern = "(\\begin{table}%[[^%]]+%])(.*)(\\end{table})"
-local latexTablePattern = "(\\begin{table})(.*)(\\end{table})"
-local latexLongtablePatternwWithPosAndAlign = "(\\begin{longtable}%[[^%]]+%]{[^\n]*})(.*)(\\end{longtable})"
-local latexLongtablePatternWithPos = "(\\begin{longtable}%[[^%]]+%])(.*)(\\end{longtable})"
-local latexLongtablePatternWithAlign = "(\\begin{longtable}{[^\n]*})(.*)(\\end{longtable})"
-local latexLongtablePattern = "(\\begin{longtable})(.*)(\\end{longtable})"
-local latexTabularPatternWithPosAndAlign = "(\\begin{tabular}%[[^%]]+%]{[^\n]*})(.*)(\\end{tabular})"
-local latexTabularPatternWithPos = "(\\begin{tabular}%[[^%]]+%])(.*)(\\end{tabular})"
-local latexTabularPatternWithAlign = "(\\begin{tabular}{[^\n]*})(.*)(\\end{tabular})"
-local latexTabularPattern = "(\\begin{tabular})(.*)(\\end{tabular})"
-local latexCaptionPattern =  "(\\caption{)(.-)(}[^\n]*\n)"
+local latexTableWithOptionsPattern_table = { "\\begin{table}%[[^%]]+%]", ".*", "\\end{table}" }
+local latexTablePattern_table = { "\\begin{table}", ".*", "\\end{table}" }
+local latexLongtablePatternWithPosAndAlign_table = { "\\begin{longtable}%[[^%]]+%]{[^\n]*}", ".*", "\\end{longtable}" }
+local latexLongtablePatternWithPos_table = { "\\begin{longtable}%[[^%]]+%]", ".*", "\\end{longtable}" }
+local latexLongtablePatternWithAlign_table = { "\\begin{longtable}{[^\n]*}", ".*", "\\end{longtable}" }
+local latexLongtablePattern_table = { "\\begin{longtable}", ".*", "\\end{longtable}" }
+local latexTabularPatternWithPosAndAlign_table = { "\\begin{tabular}%[[^%]]+%]{[^\n]*}", ".*", "\\end{tabular}" }
+local latexTabularPatternWithPos_table = { "\\begin{tabular}%[[^%]]+%]", ".*", "\\end{tabular}" }
+local latexTabularPatternWithAlign_table = { "\\begin{tabular}{[^\n]*}", ".*", "\\end{tabular}" }
+local latexTabularPattern_table = { "\\begin{tabular}", ".*", "\\end{tabular}" }
+local latexCaptionPattern_table = { "\\caption{", ".-", "}[^\n]*\n" }
 
 local latexTablePatterns = pandoc.List({
-  latexTableWithOptionsPattern,
-  latexTablePattern,
-  latexLongtablePatternwWithPosAndAlign,
-  latexLongtablePatternWithPos,
-  latexLongtablePatternWithAlign,
-  latexLongtablePattern,
-  latexTabularPatternWithPosAndAlign,
-  latexTabularPatternWithPos,
-  latexTabularPatternWithAlign,
-  latexTabularPattern,
+  latexTableWithOptionsPattern_table,
+  latexTablePattern_table,
+  latexLongtablePatternWithPosAndAlign_table,
+  latexLongtablePatternWithPos_table,
+  latexLongtablePatternWithAlign_table,
+  latexLongtablePattern_table,
+  latexTabularPatternWithPosAndAlign_table,
+  latexTabularPatternWithPos_table,
+  latexTabularPatternWithAlign_table,
+  latexTabularPattern_table,
 })
 
 -- global quarto params
@@ -1876,11 +1876,11 @@ _quarto = {
    processDependencies = processDependencies,
    format = format,
    patterns = {
-      latexTabularPattern = latexTabularPattern,
-      latexTablePattern = latexTablePattern,
-      latexLongtablePattern = latexLongtablePattern,
+      latexTabularPattern = latexTabularPattern_table,
+      latexTablePattern = latexTablePattern_table,
+      latexLongtablePattern = latexLongtablePattern_table,
       latexTablePatterns = latexTablePatterns,
-      latexCaptionPattern = latexCaptionPattern
+      latexCaptionPattern = latexCaptionPattern_table
    },
    utils = utils,
    withScriptFile = function(file, callback)