From 2cac075c0e12ec8f4eb9cf0546fafda08572ce7a Mon Sep 17 00:00:00 2001
From: Amit Haridas <amit.wh@gmail.com>
Date: Tue, 30 Jun 2026 19:57:40 +0530
Subject: [PATCH] fix(word): harden DOCX preprocessing and temp-file cleanup

- Make the HTML preprocessor code-block and inline-code aware so code

  examples containing <style> / <div> / comments are preserved.

- Strip all raw <div> tags (not just alignment attributes) to avoid

  malformed output from unmatched closing tags.

- Handle uppercase tags and single/unquoted attributes.

- Create temporary DOCX input files inside private mkdtemp directories

  instead of predictable names in the shared temp directory.

- Wrap batch DOCX preprocessing in try/catch so one unreadable file

  does not abort the entire batch.

- Add regression tests for the edge cases above.
---
 src/main.js                          |  57 ++++++++++++++++++++-------
 src/wordTemplateExporter.js          | Bin 25641 -> 27089 bytes
 tests/word-template-exporter.test.js |  56 ++++++++++++++++++++++++++
 3 files changed, 98 insertions(+), 15 deletions(-)

diff --git a/src/main.js b/src/main.js
index 3385f54..09a7e41 100644
--- a/src/main.js
+++ b/src/main.js
@@ -2563,17 +2563,16 @@ function performExportWithOptions(format, options) {
       // Use pandoc for export with advanced options
 
       let inputFile = currentFile;
+      let tempInputDir = null;
       let tempInputFile = null;
 
       // Pre-process markdown for Word output so that <style> blocks, HTML
-      // comments, and alignment <div> tags do not appear as visible text.
+      // comments, and <div> tags do not appear as visible text.
       if (format === 'docx') {
         const content = fs.readFileSync(currentFile, 'utf-8');
         const cleanedContent = WordTemplateExporter.preprocessMarkdownForWordExport(content);
-        tempInputFile = path.join(
-          require('os').tmpdir(),
-          `mc_export_${Date.now()}_${path.basename(currentFile)}`
-        );
+        tempInputDir = fs.mkdtempSync(path.join(require('os').tmpdir(), 'mc_export_'));
+        tempInputFile = path.join(tempInputDir, path.basename(currentFile));
         fs.writeFileSync(tempInputFile, cleanedContent, 'utf-8');
         inputFile = tempInputFile;
       }
@@ -2681,6 +2680,13 @@ function performExportWithOptions(format, options) {
               // Ignore cleanup errors
             }
           }
+          if (tempInputDir) {
+            try {
+              fs.rmdirSync(tempInputDir);
+            } catch {
+              // Ignore cleanup errors
+            }
+          }
         });
       } else if (format === 'pptx') {
         // Add PowerPoint footer if enabled
@@ -3777,19 +3783,33 @@ async function performBatchConversion(inputFolder, outputFolder, format, options
 
     // Build pandoc command for other formats
     let pandocInputFile = inputFile;
+    let batchTempInputDir = null;
     let batchTempInputFile = null;
 
     // Pre-process markdown for Word output so that <style> blocks, HTML
-    // comments, and alignment <div> tags do not appear as visible text.
+    // comments, and <div> tags do not appear as visible text.
     if (format === 'docx') {
-      const content = fs.readFileSync(inputFile, 'utf-8');
-      const cleanedContent = WordTemplateExporter.preprocessMarkdownForWordExport(content);
-      batchTempInputFile = path.join(
-        require('os').tmpdir(),
-        `mc_batch_export_${Date.now()}_${path.basename(inputFile)}`
-      );
-      fs.writeFileSync(batchTempInputFile, cleanedContent, 'utf-8');
-      pandocInputFile = batchTempInputFile;
+      try {
+        const content = fs.readFileSync(inputFile, 'utf-8');
+        const cleanedContent = WordTemplateExporter.preprocessMarkdownForWordExport(content);
+        batchTempInputDir = fs.mkdtempSync(path.join(require('os').tmpdir(), 'mc_batch_export_'));
+        batchTempInputFile = path.join(batchTempInputDir, path.basename(inputFile));
+        fs.writeFileSync(batchTempInputFile, cleanedContent, 'utf-8');
+        pandocInputFile = batchTempInputFile;
+      } catch (preprocessError) {
+        console.error(
+          `Batch: Failed to pre-process ${path.basename(inputFile)}:`,
+          preprocessError.message
+        );
+        mainWindow.webContents.send('batch-progress', {
+          completed: index + 1,
+          total: totalCount,
+          currentFile: path.basename(inputFile),
+          success: false,
+        });
+        processNextFile(index + 1);
+        return;
+      }
     }
 
     let pandocCmd = `${getPandocPath()} "${pandocInputFile}" -o "${outputFile}"`;
@@ -3893,7 +3913,7 @@ async function performBatchConversion(inputFolder, outputFolder, format, options
 
     // Execute conversion (using runPandocCmd for safety)
     runPandocCmd(pandocCmd, async (error, _stdout, stderr) => {
-      // Clean up temporary pre-processed input file
+      // Clean up temporary pre-processed input file and directory
       if (batchTempInputFile) {
         try {
           fs.unlinkSync(batchTempInputFile);
@@ -3901,6 +3921,13 @@ async function performBatchConversion(inputFolder, outputFolder, format, options
           // Ignore cleanup errors
         }
       }
+      if (batchTempInputDir) {
+        try {
+          fs.rmdirSync(batchTempInputDir);
+        } catch {
+          // Ignore cleanup errors
+        }
+      }
 
       if (error) {
         console.error(`Batch: Failed to convert ${path.basename(inputFile)}:`, error.message, stderr);
diff --git a/src/wordTemplateExporter.js b/src/wordTemplateExporter.js
index 84c61de3800ac07f71af60db16d354794a7e095f..50df038453ba777450a547fcb2cf6f860793c99c 100644
GIT binary patch
delta 1352
zcmZvc&ui2`6vtb(Mcj*uXqSTU4%JQiBdZ?UZj07}g|-SUDD+pR$!vCPb|z$I`-8IP
ze=vWA2SL!YSG|fSPu|6g|AB8NlXMrngx#0-zVFxP<^8wu??1*r-mH=3ke#l8B9R1#
z@gVMpj7N0<m<1BEi1#R0K$J?^abHpCdax!!3dgYsP9zW>0^>2`6asCQ3E^l>DI_VC
zG(DwZK2uYf7KBy{c9NJW=#lg!6lWYml`w$|M$gD`!56faHAx!j5+q0gXB0wFN^zk;
zVp$11-`&~-)g=la6i8onMXIGSj0sNS_`-uECIQv8N`_^2C=aOgj8q`FM5QuK8}<(J
zKX;w1pQb3ViLFV{#-;j@dnfM8E}9(|+ADBr-T<%;IKfby7gHW7^lP`_p$jQh{gf|R
z=Q5+0pp;We0H}+EE&|KAp7T>ZW>|%$o=6iGp{eM(YR+2?^v1u7Fp~6*VBBf$^QK$6
z4cs#6u%qotZ)RYd>)_y=Vjz#rQbng)mM8V@3SoDo<<{Z8ob`P_2WaG9rtV#{S@K_v
z>FWSyR*svmH<qXE@3I`7d&oJ@8V~Z<{;eT_Mm8|L7SAf&*zD#s15Z-LIwU|$a3^{u
zBYQ#cUZz6PK>Y_+<jMeK(Ksf-iDH^<CN;#b3W^I>uv$Zl-o}nWy&yyPQNBxxOf=lZ
zGO!$7jh9jbpwsgA4*i2UzZK<gZ%k#CP60uWsGwWz3M??-k8NykZfvg~t-W}*el&A=
z6=nHafJcKdzZ|OGw&i1JLJ88}#)ixoisI_|=i1l%#Rotnl;hzSAA&Q6&=P)eb@%Tk
z@8lmQ>iN%!t*lZncrK0*<*OzuI9nryJNx0BJJyOyDgm6Md0a?H&&{@~w)W<X@xbcX
zfs~A9Wm<c&*30XW#BuCV8NKq4r!UIfpuM8i^4dMU8$2rW8L1GuiqjmqZkfQ|RSgd-
N|LfiUv^x3w@E=Tors4nq

delta 155
zcmcb3nQ`R_#tj0@B8fSf>3O-Sc_j)qDVb$<3MGl@#ghw}OQaQmpeVJZv?xzOgA2q_
zP{>Uz%1+5I&zoGuaYD|&AT=*DFCDB+0c?t`lAfNP64ZzV93qozIJso)W0Js<F~wT8
hKpd^49vdA8B*3i6jhsPjV8b;hPfQcu{G97m900YhGBW@G

diff --git a/tests/word-template-exporter.test.js b/tests/word-template-exporter.test.js
index 8cb8b7e..55f9d4b 100644
--- a/tests/word-template-exporter.test.js
+++ b/tests/word-template-exporter.test.js
@@ -58,4 +58,60 @@ Centered content
     const output = WordTemplateExporter.preprocessMarkdownForWordExport(input);
     expect(output).toBe(input);
   });
+
+  test('preserves HTML artifacts inside fenced code blocks', () => {
+    const input = `# Title
+
+\`\`\`
+<style>body{}</style>
+<!-- comment -->
+<div align="center">text</div>
+\`\`\`
+
+After code.
+`;
+    const output = WordTemplateExporter.preprocessMarkdownForWordExport(input);
+    expect(output).toContain('<style>body{}</style>');
+    expect(output).toContain('<!-- comment -->');
+    expect(output).toContain('<div align="center">text</div>');
+    expect(output).toContain('After code.');
+  });
+
+  test('preserves HTML artifacts inside inline code', () => {
+    const input = 'Use `<div align="center">` for alignment.';
+    const output = WordTemplateExporter.preprocessMarkdownForWordExport(input);
+    expect(output).toContain('`<div align="center">`');
+  });
+
+  test('handles uppercase tags and unquoted attributes', () => {
+    const input = `<DIV align=center>
+Centered
+</DIV>
+`;
+    const output = WordTemplateExporter.preprocessMarkdownForWordExport(input);
+    expect(output).not.toContain('<DIV');
+    expect(output).not.toContain('</DIV>');
+    expect(output).toContain('Centered');
+  });
+
+  test('handles single-quoted attributes', () => {
+    const input = `<div align='right'>Right</div>`;
+    const output = WordTemplateExporter.preprocessMarkdownForWordExport(input);
+    expect(output).not.toContain('<div');
+    expect(output).not.toContain('</div>');
+    expect(output).toContain('Right');
+  });
+
+  test('removes non-alignment div tags without leaving malformed HTML', () => {
+    const input = `<div class="note">Note text</div>`;
+    const output = WordTemplateExporter.preprocessMarkdownForWordExport(input);
+    expect(output).not.toContain('<div');
+    expect(output).not.toContain('</div>');
+    expect(output).toContain('Note text');
+  });
+
+  test('returns non-string input unchanged', () => {
+    expect(WordTemplateExporter.preprocessMarkdownForWordExport(null)).toBeNull();
+    expect(WordTemplateExporter.preprocessMarkdownForWordExport(123)).toBe(123);
+  });
 });