From 294088fb273fc531b972d29ee39a52216a4d668b Mon Sep 17 00:00:00 2001
From: Jakob Ackermann <jakob.ackermann@overleaf.com>
Date: Mon, 31 May 2021 10:20:25 +0200
Subject: [PATCH] [ContentCacheManager] use PDF.js Xref table instead of stream
 detection (#242)

* make the content cache manager tests configurable

* extend stream content in unit tests

* [ContentCacheManagerTests] prepare for full object caching

* filesystem stream for pdfjs

* working??

* cleaning up

* handle overflow

* [misc] install pdfjs-dist

* [misc] move pdfjs code into app/lib/ and scripts/, also use CamelCase

* [misc] abstract the file loading and parsing of xRef tables into helper

* [misc] pdfjsTests: add snapshot based tests for the Xref table parser

* [misc] FSStream: throw proper error and drop commented code

* [misc] FSStream: integrate throwing of MissingDataException into getter

* [misc] pdfjs: fix eslint errors

* [misc] pdfjs: run format_fix

* [misc] pdfjs: allocate very small non empty dummy buffers explicitly

* [misc] install @overleaf/o-error

* [ContentCacheManager] use PDF.js Xref table instead of stream detection

Co-Authored-By: Brian Gough <brian.gough@overleaf.com>

* [pdfjs] parseXrefTable: handle empty PDF files gracefully

Co-authored-by: Brian Gough <brian.gough@overleaf.com>
---
 services/clsi/app/js/ContentCacheManager.js   | 174 ++++----
 services/clsi/app/js/OutputCacheManager.js    |  57 +--
 services/clsi/app/lib/pdfjs/FSPdfManager.js   |  43 ++
 services/clsi/app/lib/pdfjs/FSStream.js       | 138 +++++++
 services/clsi/app/lib/pdfjs/parseXrefTable.js |  24 ++
 services/clsi/package-lock.json               |  21 +-
 services/clsi/package.json                    |   4 +-
 services/clsi/scripts/demo-pdfjs-Xref.js      |  12 +
 .../clsi/test/acceptance/fixtures/minimal.pdf | Bin 0 -> 12313 bytes
 .../test/unit/js/ContentCacheManagerTests.js  | 371 ++++++------------
 ...71616916a7ba0434fcfb56d1fc7e27cd139eaa2f71 |   7 +
 ...17923e3714927bbf35e63ea88bd93c7a8076cf1fcd | Bin 0 -> 10161 bytes
 services/clsi/test/unit/lib/pdfjsTests.js     |  78 ++++
 .../lib/snapshots/asymptote/XrefTable.json    | 356 +++++++++++++++++
 .../biber_bibliography/XrefTable.json         | 128 ++++++
 .../lib/snapshots/epstopdf/XrefTable.json     | 125 ++++++
 .../unit/lib/snapshots/feynmf/XrefTable.json  | 115 ++++++
 .../unit/lib/snapshots/feynmp/XrefTable.json  | 102 +++++
 .../lib/snapshots/fontawesome/XrefTable.json  |  93 +++++
 .../fontawesome_xelatex/XrefTable.json        | 126 ++++++
 .../lib/snapshots/glossaries/XrefTable.json   |  94 +++++
 .../unit/lib/snapshots/gnuplot/XrefTable.json |  89 +++++
 .../unit/lib/snapshots/hebrew/XrefTable.json  |  81 ++++
 .../unit/lib/snapshots/knitr/XrefTable.json   | 145 +++++++
 .../lib/snapshots/knitr_utf8/XrefTable.json   | 179 +++++++++
 .../snapshots/latex_compiler/XrefTable.json   | 132 +++++++
 .../lualatex_compiler/XrefTable.json          |  74 ++++
 .../makeindex-custom-style/XrefTable.json     | 107 +++++
 .../lib/snapshots/makeindex/XrefTable.json    |  90 +++++
 .../unit/lib/snapshots/minted/XrefTable.json  |  77 ++++
 .../multibib_bibliography/XrefTable.json      | 133 +++++++
 .../lib/snapshots/nomenclature/XrefTable.json |  94 +++++
 .../references_in_include/XrefTable.json      |  90 +++++
 .../simple_bibliography/XrefTable.json        |  94 +++++
 .../snapshots/subdirectories/XrefTable.json   | 108 +++++
 .../lib/snapshots/tikz_feynman/XrefTable.json | 145 +++++++
 .../snapshots/xelatex_compiler/XrefTable.json |  73 ++++
 37 files changed, 3406 insertions(+), 373 deletions(-)
 create mode 100644 services/clsi/app/lib/pdfjs/FSPdfManager.js
 create mode 100644 services/clsi/app/lib/pdfjs/FSStream.js
 create mode 100644 services/clsi/app/lib/pdfjs/parseXrefTable.js
 create mode 100644 services/clsi/scripts/demo-pdfjs-Xref.js
 create mode 100644 services/clsi/test/acceptance/fixtures/minimal.pdf
 create mode 100644 services/clsi/test/unit/js/snapshots/minimalCompile/chunks/896749b8343851b0dc385f71616916a7ba0434fcfb56d1fc7e27cd139eaa2f71
 create mode 100644 services/clsi/test/unit/js/snapshots/minimalCompile/chunks/d7cfc73ad2fba4578a437517923e3714927bbf35e63ea88bd93c7a8076cf1fcd
 create mode 100644 services/clsi/test/unit/lib/pdfjsTests.js
 create mode 100644 services/clsi/test/unit/lib/snapshots/asymptote/XrefTable.json
 create mode 100644 services/clsi/test/unit/lib/snapshots/biber_bibliography/XrefTable.json
 create mode 100644 services/clsi/test/unit/lib/snapshots/epstopdf/XrefTable.json
 create mode 100644 services/clsi/test/unit/lib/snapshots/feynmf/XrefTable.json
 create mode 100644 services/clsi/test/unit/lib/snapshots/feynmp/XrefTable.json
 create mode 100644 services/clsi/test/unit/lib/snapshots/fontawesome/XrefTable.json
 create mode 100644 services/clsi/test/unit/lib/snapshots/fontawesome_xelatex/XrefTable.json
 create mode 100644 services/clsi/test/unit/lib/snapshots/glossaries/XrefTable.json
 create mode 100644 services/clsi/test/unit/lib/snapshots/gnuplot/XrefTable.json
 create mode 100644 services/clsi/test/unit/lib/snapshots/hebrew/XrefTable.json
 create mode 100644 services/clsi/test/unit/lib/snapshots/knitr/XrefTable.json
 create mode 100644 services/clsi/test/unit/lib/snapshots/knitr_utf8/XrefTable.json
 create mode 100644 services/clsi/test/unit/lib/snapshots/latex_compiler/XrefTable.json
 create mode 100644 services/clsi/test/unit/lib/snapshots/lualatex_compiler/XrefTable.json
 create mode 100644 services/clsi/test/unit/lib/snapshots/makeindex-custom-style/XrefTable.json
 create mode 100644 services/clsi/test/unit/lib/snapshots/makeindex/XrefTable.json
 create mode 100644 services/clsi/test/unit/lib/snapshots/minted/XrefTable.json
 create mode 100644 services/clsi/test/unit/lib/snapshots/multibib_bibliography/XrefTable.json
 create mode 100644 services/clsi/test/unit/lib/snapshots/nomenclature/XrefTable.json
 create mode 100644 services/clsi/test/unit/lib/snapshots/references_in_include/XrefTable.json
 create mode 100644 services/clsi/test/unit/lib/snapshots/simple_bibliography/XrefTable.json
 create mode 100644 services/clsi/test/unit/lib/snapshots/subdirectories/XrefTable.json
 create mode 100644 services/clsi/test/unit/lib/snapshots/tikz_feynman/XrefTable.json
 create mode 100644 services/clsi/test/unit/lib/snapshots/xelatex_compiler/XrefTable.json

diff --git a/services/clsi/app/js/ContentCacheManager.js b/services/clsi/app/js/ContentCacheManager.js
index d424681615..ab73c61d93 100644
--- a/services/clsi/app/js/ContentCacheManager.js
+++ b/services/clsi/app/js/ContentCacheManager.js
@@ -7,45 +7,97 @@ const fs = require('fs')
 const crypto = require('crypto')
 const Path = require('path')
 const Settings = require('settings-sharelatex')
+const OError = require('@overleaf/o-error')
 const pLimit = require('p-limit')
-
-const MIN_CHUNK_SIZE = Settings.pdfCachingMinChunkSize
-
-const START_OF_STREAM_MARKER = 'stream'
-const END_OF_STREAM_MARKER = 'endstream'
-const START_OF_STREAM_MARKER_LENGTH = START_OF_STREAM_MARKER.length
-const END_OF_STREAM_MARKER_LENGTH = END_OF_STREAM_MARKER.length
+const { parseXrefTable } = require('../lib/pdfjs/parseXrefTable')
 
 /**
  *
  * @param {String} contentDir path to directory where content hash files are cached
  * @param {String} filePath the pdf file to scan for streams
+ * @param {number} size the pdf size
  */
-async function update(contentDir, filePath) {
-  const stream = fs.createReadStream(filePath)
-  const extractor = new PdfStreamsExtractor()
+async function update(contentDir, filePath, size) {
   const ranges = []
   const newRanges = []
   // keep track of hashes expire old ones when they reach a generation > N.
   const tracker = await HashFileTracker.from(contentDir)
   tracker.updateAge()
 
-  for await (const chunk of stream) {
-    const pdfStreams = extractor.consume(chunk)
-    for (const pdfStream of pdfStreams) {
-      if (pdfStream.end - pdfStream.start < MIN_CHUNK_SIZE) continue
-      const hash = pdfStreamHash(pdfStream.buffers)
+  const rawTable = await parseXrefTable(filePath, size)
+  rawTable.sort((a, b) => {
+    return a.offset - b.offset
+  })
+  rawTable.forEach((obj, idx) => {
+    obj.idx = idx
+  })
 
-      const range = { start: pdfStream.start, end: pdfStream.end, hash }
+  const uncompressedObjects = []
+  for (const object of rawTable) {
+    if (!object.uncompressed) {
+      continue
+    }
+    const nextObject = rawTable[object.idx + 1]
+    if (!nextObject) {
+      // Ignore this possible edge case.
+      // The last object should be part of the xRef table.
+      continue
+    } else {
+      object.endOffset = nextObject.offset
+    }
+    const size = object.endOffset - object.offset
+    object.size = size
+    if (size < Settings.pdfCachingMinChunkSize) {
+      continue
+    }
+    uncompressedObjects.push(object)
+  }
+
+  const handle = await fs.promises.open(filePath)
+  try {
+    for (const object of uncompressedObjects) {
+      let buffer = Buffer.alloc(object.size, 0)
+      const { bytesRead } = await handle.read(
+        buffer,
+        0,
+        object.size,
+        object.offset
+      )
+      if (bytesRead !== object.size) {
+        throw new OError('could not read full chunk', {
+          object,
+          bytesRead
+        })
+      }
+      const idxObj = buffer.indexOf('obj')
+      if (idxObj > 100) {
+        throw new OError('objectId is too large', {
+          object,
+          idxObj
+        })
+      }
+      const objectIdRaw = buffer.subarray(0, idxObj)
+      buffer = buffer.subarray(objectIdRaw.byteLength)
+
+      const hash = pdfStreamHash(buffer)
+      const range = {
+        objectId: objectIdRaw.toString(),
+        start: object.offset + objectIdRaw.byteLength,
+        end: object.endOffset,
+        hash
+      }
       ranges.push(range)
 
       // Optimization: Skip writing of duplicate streams.
       if (tracker.track(range)) continue
 
-      await writePdfStream(contentDir, hash, pdfStream.buffers)
+      await writePdfStream(contentDir, hash, buffer)
       newRanges.push(range)
     }
+  } finally {
+    await handle.close()
   }
+
   const reclaimedSpace = await tracker.deleteStaleHashes(5)
   await tracker.flush()
   return [ranges, newRanges, reclaimedSpace]
@@ -142,94 +194,21 @@ class HashFileTracker {
   }
 }
 
-class PdfStreamsExtractor {
-  constructor() {
-    this.fileIndex = 0
-    this.inStream = false
-    this.streamStartIndex = 0
-    this.buffers = []
-    this.lastChunk = Buffer.alloc(0)
-  }
-
-  consume(chunk) {
-    let chunkIndex = 0
-    const pdfStreams = []
-    chunk = Buffer.concat([this.lastChunk, chunk])
-    while (true) {
-      if (!this.inStream) {
-        // Not in a stream, look for stream start
-        const index = chunk.indexOf(START_OF_STREAM_MARKER, chunkIndex)
-        if (index === -1) {
-          // Couldn't find stream start
-          break
-        }
-        // Found stream start, start a stream
-        this.inStream = true
-        this.streamStartIndex = this.fileIndex + index
-        chunkIndex = index
-      } else {
-        // In a stream, look for stream end
-        const index = chunk.indexOf(END_OF_STREAM_MARKER, chunkIndex)
-        if (index === -1) {
-          break
-        }
-        // add "endstream" part
-        const endIndex = index + END_OF_STREAM_MARKER_LENGTH
-        this.buffers.push(chunk.slice(chunkIndex, endIndex))
-        pdfStreams.push({
-          start: this.streamStartIndex,
-          end: this.fileIndex + endIndex,
-          buffers: this.buffers
-        })
-        this.inStream = false
-        this.buffers = []
-        chunkIndex = endIndex
-      }
-    }
-
-    const remaining = chunk.length - chunkIndex
-    const nextMarkerLength = this.inStream
-      ? END_OF_STREAM_MARKER_LENGTH
-      : START_OF_STREAM_MARKER_LENGTH
-    if (remaining > nextMarkerLength) {
-      const retainMarkerSection = chunk.length - nextMarkerLength
-      if (this.inStream) {
-        this.buffers.push(chunk.slice(chunkIndex, retainMarkerSection))
-      }
-      this.lastChunk = chunk.slice(retainMarkerSection)
-      this.fileIndex += retainMarkerSection
-    } else {
-      this.lastChunk = chunk.slice(chunkIndex)
-      this.fileIndex += chunkIndex
-    }
-    return pdfStreams
-  }
-}
-
-function pdfStreamHash(buffers) {
+function pdfStreamHash(buffer) {
   const hash = crypto.createHash('sha256')
-  for (const buffer of buffers) {
-    hash.update(buffer)
-  }
+  hash.update(buffer)
   return hash.digest('hex')
 }
 
-async function writePdfStream(dir, hash, buffers) {
+async function writePdfStream(dir, hash, buffer) {
   const filename = Path.join(dir, hash)
   const atomicWriteFilename = filename + '~'
-  const file = await fs.promises.open(atomicWriteFilename, 'w')
   if (Settings.enablePdfCachingDark) {
     // Write an empty file in dark mode.
-    buffers = []
+    buffer = Buffer.alloc(0)
   }
   try {
-    try {
-      for (const buffer of buffers) {
-        await file.write(buffer)
-      }
-    } finally {
-      await file.close()
-    }
+    await fs.promises.writeFile(atomicWriteFilename, buffer)
     await fs.promises.rename(atomicWriteFilename, filename)
   } catch (err) {
     try {
@@ -247,5 +226,8 @@ function promiseMapWithLimit(concurrency, array, fn) {
 
 module.exports = {
   HASH_REGEX: /^[0-9a-f]{64}$/,
-  update: callbackify(update)
+  update: callbackify(update),
+  promises: {
+    update
+  }
 }
diff --git a/services/clsi/app/js/OutputCacheManager.js b/services/clsi/app/js/OutputCacheManager.js
index e6167570f8..4d2e64132b 100644
--- a/services/clsi/app/js/OutputCacheManager.js
+++ b/services/clsi/app/js/OutputCacheManager.js
@@ -275,36 +275,39 @@ module.exports = OutputCacheManager = {
           outputDir,
           OutputCacheManager.path(outputFile.build, outputFile.path)
         )
+        const pdfSize = outputFile.size
         const timer = new Metrics.Timer('compute-pdf-ranges')
-        ContentCacheManager.update(contentDir, outputFilePath, function (
-          err,
-          result
-        ) {
-          if (err) return callback(err, outputFiles)
-          const [contentRanges, newContentRanges, reclaimedSpace] = result
+        ContentCacheManager.update(
+          contentDir,
+          outputFilePath,
+          pdfSize,
+          function (err, result) {
+            if (err) return callback(err, outputFiles)
+            const [contentRanges, newContentRanges, reclaimedSpace] = result
 
-          if (Settings.enablePdfCachingDark) {
-            // In dark mode we are doing the computation only and do not emit
-            //  any ranges to the frontend.
-          } else {
-            outputFile.contentId = Path.basename(contentDir)
-            outputFile.ranges = contentRanges
+            if (Settings.enablePdfCachingDark) {
+              // In dark mode we are doing the computation only and do not emit
+              //  any ranges to the frontend.
+            } else {
+              outputFile.contentId = Path.basename(contentDir)
+              outputFile.ranges = contentRanges
+            }
+
+            timings['compute-pdf-caching'] = timer.done()
+            stats['pdf-caching-n-ranges'] = contentRanges.length
+            stats['pdf-caching-total-ranges-size'] = contentRanges.reduce(
+              (sum, next) => sum + (next.end - next.start),
+              0
+            )
+            stats['pdf-caching-n-new-ranges'] = newContentRanges.length
+            stats['pdf-caching-new-ranges-size'] = newContentRanges.reduce(
+              (sum, next) => sum + (next.end - next.start),
+              0
+            )
+            stats['pdf-caching-reclaimed-space'] = reclaimedSpace
+            callback(null, outputFiles)
           }
-
-          timings['compute-pdf-caching'] = timer.done()
-          stats['pdf-caching-n-ranges'] = contentRanges.length
-          stats['pdf-caching-total-ranges-size'] = contentRanges.reduce(
-            (sum, next) => sum + (next.end - next.start),
-            0
-          )
-          stats['pdf-caching-n-new-ranges'] = newContentRanges.length
-          stats['pdf-caching-new-ranges-size'] = newContentRanges.reduce(
-            (sum, next) => sum + (next.end - next.start),
-            0
-          )
-          stats['pdf-caching-reclaimed-space'] = reclaimedSpace
-          callback(null, outputFiles)
-        })
+        )
       } else {
         callback(null, outputFiles)
       }
diff --git a/services/clsi/app/lib/pdfjs/FSPdfManager.js b/services/clsi/app/lib/pdfjs/FSPdfManager.js
new file mode 100644
index 0000000000..a0450c1e85
--- /dev/null
+++ b/services/clsi/app/lib/pdfjs/FSPdfManager.js
@@ -0,0 +1,43 @@
+const { PDFDocument } = require('pdfjs-dist/lib/core/document')
+const { LocalPdfManager } = require('pdfjs-dist/lib/core/pdf_manager')
+const { MissingDataException } = require('pdfjs-dist/lib/core/core_utils')
+const { FSStream } = require('./FSStream')
+
+class FSPdfManager extends LocalPdfManager {
+  constructor(docId, options) {
+    const nonEmptyDummyBuffer = Buffer.alloc(1, 0)
+    super(docId, nonEmptyDummyBuffer)
+    this.stream = new FSStream(options.fh, 0, options.size)
+    this.pdfDocument = new PDFDocument(this, this.stream)
+  }
+
+  async ensure(obj, prop, args) {
+    try {
+      const value = obj[prop]
+      if (typeof value === 'function') {
+        return value.apply(obj, args)
+      }
+      return value
+    } catch (ex) {
+      if (!(ex instanceof MissingDataException)) {
+        throw ex
+      }
+      await this.requestRange(ex.begin, ex.end)
+      return this.ensure(obj, prop, args)
+    }
+  }
+
+  requestRange(begin, end) {
+    return this.stream.requestRange(begin, end)
+  }
+
+  requestLoadedStream() {}
+
+  onLoadedStream() {}
+
+  terminate(reason) {}
+}
+
+module.exports = {
+  FSPdfManager
+}
diff --git a/services/clsi/app/lib/pdfjs/FSStream.js b/services/clsi/app/lib/pdfjs/FSStream.js
new file mode 100644
index 0000000000..9179e83f0e
--- /dev/null
+++ b/services/clsi/app/lib/pdfjs/FSStream.js
@@ -0,0 +1,138 @@
+const { Stream } = require('pdfjs-dist/lib/core/stream')
+const { MissingDataException } = require('pdfjs-dist/lib/core/core_utils')
+
+const BUF_SIZE = 1024 // read from the file in 1024 byte pages
+
+class FSStream extends Stream {
+  constructor(fh, start, length, dict, cachedBytes) {
+    const nonEmptyDummyBuffer = Buffer.alloc(1, 0)
+    super(nonEmptyDummyBuffer, start, length, dict)
+    delete this.bytes
+    this.fh = fh
+    this.cachedBytes = cachedBytes || []
+  }
+
+  get length() {
+    return this.end - this.start
+  }
+
+  get isEmpty() {
+    return this.length === 0
+  }
+
+  // Manage cached reads from the file
+
+  requestRange(begin, end) {
+    // expand small ranges to read a larger amount
+    if (end - begin < BUF_SIZE) {
+      end = begin + BUF_SIZE
+    }
+    end = Math.min(end, this.length)
+    // keep a cache of previous reads with {begin,end,buffer} values
+    const result = {
+      begin: begin,
+      end: end,
+      buffer: Buffer.alloc(end - begin, 0)
+    }
+    this.cachedBytes.push(result)
+    return this.fh.read(result.buffer, 0, end - begin, begin)
+  }
+
+  _ensureGetPos(pos) {
+    const found = this.cachedBytes.find((x) => {
+      return x.begin <= pos && pos < x.end
+    })
+    if (!found) {
+      throw new MissingDataException(pos, pos + 1)
+    }
+    return found
+  }
+
+  _ensureGetRange(begin, end) {
+    end = Math.min(end, this.length) // BG: handle overflow case
+    const found = this.cachedBytes.find((x) => {
+      return x.begin <= begin && end <= x.end
+    })
+    if (!found) {
+      throw new MissingDataException(begin, end)
+    }
+    return found
+  }
+
+  _readByte(found, pos) {
+    return found.buffer[pos - found.begin]
+  }
+
+  _readBytes(found, pos, end) {
+    return found.buffer.subarray(pos - found.begin, end - found.begin)
+  }
+
+  // handle accesses to the bytes
+
+  ensureByte(pos) {
+    this._ensureGetPos(pos) // may throw a MissingDataException
+  }
+
+  getByte() {
+    const pos = this.pos
+    if (this.pos >= this.end) {
+      return -1
+    }
+    const found = this._ensureGetPos(pos)
+    return this._readByte(found, this.pos++)
+  }
+
+  // BG: for a range, end is not included (see Buffer.subarray for example)
+
+  ensureBytes(length, forceClamped = false) {
+    const pos = this.pos
+    this._ensureGetRange(pos, pos + length)
+  }
+
+  getBytes(length, forceClamped = false) {
+    const pos = this.pos
+    const strEnd = this.end
+
+    const found = this._ensureGetRange(pos, pos + length)
+    if (!length) {
+      const subarray = this._readBytes(found, pos, strEnd)
+      // `this.bytes` is always a `Uint8Array` here.
+      return forceClamped ? new Uint8ClampedArray(subarray) : subarray
+    }
+    let end = pos + length
+    if (end > strEnd) {
+      end = strEnd
+    }
+    this.pos = end
+    const subarray = this._readBytes(found, pos, end)
+    // `this.bytes` is always a `Uint8Array` here.
+    return forceClamped ? new Uint8ClampedArray(subarray) : subarray
+  }
+
+  getByteRange() {
+    // BG: this isn't needed as far as I can tell
+    throw new Error('not implemented')
+  }
+
+  reset() {
+    this.pos = this.start
+  }
+
+  moveStart() {
+    this.start = this.pos
+  }
+
+  makeSubStream(start, length, dict = null) {
+    // BG: had to add this check for null length, it is being called with only
+    // the start value at one point in the xref decoding. The intent is clear
+    // enough
+    // - a null length means "to the end of the file" -- not sure how it is
+    //   working in the existing pdfjs code without this.
+    if (!length) {
+      length = this.end - start
+    }
+    return new FSStream(this.fh, start, length, dict, this.cachedBytes)
+  }
+}
+
+module.exports = { FSStream }
diff --git a/services/clsi/app/lib/pdfjs/parseXrefTable.js b/services/clsi/app/lib/pdfjs/parseXrefTable.js
new file mode 100644
index 0000000000..4db1e0cfc7
--- /dev/null
+++ b/services/clsi/app/lib/pdfjs/parseXrefTable.js
@@ -0,0 +1,24 @@
+const fs = require('fs')
+const { FSPdfManager } = require('./FSPdfManager')
+
+async function parseXrefTable(path, size) {
+  if (size === 0) {
+    return []
+  }
+
+  const file = await fs.promises.open(path)
+  try {
+    const manager = new FSPdfManager(0, { fh: file, size })
+
+    await manager.ensureDoc('checkHeader')
+    await manager.ensureDoc('parseStartXRef')
+    await manager.ensureDoc('parse')
+    return manager.pdfDocument.catalog.xref.entries
+  } finally {
+    file.close()
+  }
+}
+
+module.exports = {
+  parseXrefTable
+}
diff --git a/services/clsi/package-lock.json b/services/clsi/package-lock.json
index 773d9f4319..52571c6cc7 100644
--- a/services/clsi/package-lock.json
+++ b/services/clsi/package-lock.json
@@ -1008,9 +1008,19 @@
       }
     },
     "@overleaf/o-error": {
-      "version": "3.0.0",
-      "resolved": "https://registry.npmjs.org/@overleaf/o-error/-/o-error-3.0.0.tgz",
-      "integrity": "sha512-LsM2s6Iy9G97ktPo0ys4VxtI/m3ahc1ZHwjo5XnhXtjeIkkkVAehsrcRRoV/yWepPjymB0oZonhcfojpjYR/tg=="
+      "version": "3.3.1",
+      "resolved": "https://registry.npmjs.org/@overleaf/o-error/-/o-error-3.3.1.tgz",
+      "integrity": "sha512-1FRBYZO0lbJ0U+FRGZVS8ou6RhEw3e2B86WW/NbtBw554g0h5iC8ESf+juIfPMU/WDf/JDIFbg3eB/LnP2RSow==",
+      "requires": {
+        "core-js": "^3.8.3"
+      },
+      "dependencies": {
+        "core-js": {
+          "version": "3.13.0",
+          "resolved": "https://registry.npmjs.org/core-js/-/core-js-3.13.0.tgz",
+          "integrity": "sha512-iWDbiyha1M5vFwPFmQnvRv+tJzGbFAm6XimJUT0NgHYW3xZEs1SkCAcasWSVFxpI2Xb/V1DDJckq3v90+bQnog=="
+        }
+      }
     },
     "@protobufjs/aspromise": {
       "version": "1.1.2",
@@ -5594,6 +5604,11 @@
       "integrity": "sha1-uULm1L3mUwBe9rcTYd74cn0GReA=",
       "dev": true
     },
+    "pdfjs-dist": {
+      "version": "2.7.570",
+      "resolved": "https://registry.npmjs.org/pdfjs-dist/-/pdfjs-dist-2.7.570.tgz",
+      "integrity": "sha512-/ZkA1FwkEOyDaq11JhMLazdwQAA0F9uwrP7h/1L9Akt9KWh1G5/tkzS+bPuUELq2s2GDFnaT+kooN/aSjT7DXQ=="
+    },
     "performance-now": {
       "version": "2.1.0",
       "resolved": "https://registry.npmjs.org/performance-now/-/performance-now-2.1.0.tgz",
diff --git a/services/clsi/package.json b/services/clsi/package.json
index dac3ba9dbb..51870b025a 100644
--- a/services/clsi/package.json
+++ b/services/clsi/package.json
@@ -10,7 +10,7 @@
     "start": "node $NODE_APP_OPTIONS app.js",
     "test:acceptance:_run": "mocha --recursive --reporter spec --timeout 15000 --exit $@ test/acceptance/js",
     "test:acceptance": "npm run test:acceptance:_run -- --grep=$MOCHA_GREP",
-    "test:unit:_run": "mocha --recursive --reporter spec $@ test/unit/js",
+    "test:unit:_run": "mocha --recursive --reporter spec $@ test/unit/js test/unit/lib",
     "test:unit": "npm run test:unit:_run -- --grep=$MOCHA_GREP",
     "nodemon": "nodemon --config nodemon.json",
     "lint": "node_modules/.bin/eslint --max-warnings 0 .",
@@ -20,6 +20,7 @@
   "author": "James Allen <james@sharelatex.com>",
   "dependencies": {
     "@overleaf/metrics": "^3.5.1",
+    "@overleaf/o-error": "^3.3.1",
     "async": "3.2.0",
     "body-parser": "^1.19.0",
     "diskusage": "^1.1.3",
@@ -32,6 +33,7 @@
     "logger-sharelatex": "^2.2.0",
     "mysql": "^2.18.1",
     "p-limit": "^3.1.0",
+    "pdfjs-dist": "^2.7.570",
     "request": "^2.88.2",
     "sequelize": "^5.21.5",
     "settings-sharelatex": "^1.1.0",
diff --git a/services/clsi/scripts/demo-pdfjs-Xref.js b/services/clsi/scripts/demo-pdfjs-Xref.js
new file mode 100644
index 0000000000..86be134ef5
--- /dev/null
+++ b/services/clsi/scripts/demo-pdfjs-Xref.js
@@ -0,0 +1,12 @@
+const fs = require('fs')
+const { parseXrefTable } = require('../app/lib/pdfjs/parseXrefTable')
+
+const pdfPath = process.argv[2]
+
+async function main() {
+  const size = (await fs.promises.stat(pdfPath)).size
+  const xRefEntries = await parseXrefTable(pdfPath, size)
+  console.log('Xref entries', xRefEntries)
+}
+
+main().catch(console.error)
diff --git a/services/clsi/test/acceptance/fixtures/minimal.pdf b/services/clsi/test/acceptance/fixtures/minimal.pdf
new file mode 100644
index 0000000000000000000000000000000000000000..d578e90567e15149674581235b0d67a2f9340924
GIT binary patch
literal 12313
zcmch-Wl)^awyul2y9F9|Y24l2f&`bw-5r9vy9WvG?m>gQI|&xtT@G1muU+fhwNKUk
zaqexYuV>FO=N#`lMtxmf^-wB{OELnPI1ng%?hD=!*Z?d5dt)mEetrP6yrmt;$jQ<R
zWC{R&i~vAZHXs{-SsI`N;N${+3|Rnr0A?is7l2s;00eSzvjdnF|AwmnonvDK2nr&A
z>`eb|2l)T{fx!N6LexDSKmcYfRgl?-43M3<iv@uFBWX!X8yAq%$JNHj1tbnKu{ZsQ
z`sY>I$jR0j@Ye(}dmC3<J7>U$Hf1M}sildFz0-%h4{ZQu%?~Gl0M?K9hYvD#rXY`x
zFW3MdCjDb<X8$q60#F4otJ>SU066~nS?$A60Fd(^+qwP2D7ODFhgn7(pu-R3-~@rp
zc(_?jO+j2HCY&ZfP8K5*6Cevan+Z3Ev5_D@7nhMShY`COrzr=Q8M`qE#L4n831Ve8
zVFPlQaT^Qj{jGtsixbGm7QrLi#MpG(+|byN8=fNld!Ux2a9|z803A$SDj5826_R|k
z=7NYZQnUpWSid(k{om64*UJA}MMU^(!N0mo*~lE^{MW00IP}jFi5a;V+1Q)^TjOy4
z+lv3OUTOfg4=?{!Rkr_Kuc4u#k)dJc6d2SO68vqjQ8P2Jv3_vGY=iM1u*s>N)2$KU
zU<KgNQU9rx>)*~5v$t~r*||6axc`}}!bckziQ4}yHp|D8gNK8Om5UR=%?@PZ=J{Lt
zkM8-1|8H%oAZL45CzHQw|EnMVsrIh}|0ebC?*B69f0X)PB6<FIk(8r+2mvfeVAEYC
z!^<c!RIf$Vfn<{_1691(P`EYBzg{*=zh@yLahp!&Q!6vj3;?<Sju2I~D|;5FPJ+RP
zBtMO7IzG1?BP5eaQeDs@;yiNsno@BvAHQK2{--TG|L=1AxH+(M^8BkTtN<PsPS$^o
z{ucaS^#Zc6aQx#o@!u^P(+I7eJ%+)EL@Mpzj=#PU*x~K41P&95uN!3TfC`0r_H{dP
z8)yAfpmi-&nv~OI*ZunQyLDHKS(Uw)=4pBL(dI&9@=<0bhK4d?J_FXT(b171Qn+|k
z&;Akc?d@%|-`m@w--{O@+5*4&XC+v=223&mXv8l_plHqW+m0I)R@ZMcN+3eOur_{x
zDcXTi=K-jD0I;wDQ9)wc2I8d+0O8g3Po~hyMvye3q{6tSUua7U)7XZG7H>VX{29TK
zQlr5X!NL7fg?7N;Ve{FBXNe$`f$mMvy!_`DmMAb5qKqQKnl<mTy#Slzqa&K($+MlE
zUbD09J{Z?#bdV{yM}V#^xP1VVu(^9a_O^RJ*k@<hZG)UT(m9Am8xbb|6^?I<)x`Ug
z!ao|ok?NsDb#uwr0UE;QKd$$|6re!?ndE3!xZ01*5Qy974d4#-j&FP$TQfbFUjhDs
z6O4@CP-6Gdn9&Ac43Qxq!L8~?ZK`vt8Ni&w7wijomZ6`B?9c2_AhLA$G2hhfP~uaj
z!P14e9^Ac|QYQgYnz)>}Xb<jDCp~lgdG$yGvIs}pV8N~JfcI)8f?cFWA3-<fyB@nm
zVu;&-hZp2Z5j6VaEd{MMo_cFBVMRM(=s!KQQ&hX}%QgYkVEQO1D1P8}VAJqmH;s*;
z_Z<DPG4zL@Badq@>tAc)$mh`wU)Q0gaZQnZ-o)-4f$p#nx;YyG-Gc8fFI=MgyHHwR
z8T7z$gbC509)(u&CWMcBF0S|Zg80EhiKbS6Lpi*>y?TRQ7?|m4piXW{p7NeSnks&U
z`7{$Rnm^y~WMoAJzy&M$iNG|N^ijYL>=Q19n0?;)V%m8Q&spElE5PbPA)wyM?;g|N
zKG!{OK^Q)GbK`-(=g46{Rek{k><E!@{o3BE^?7r66OwyX8h+2~^EZF&iFolxOmu||
zdT+&ldw3VJbB0A*KDO^`u-MSegB6B;;)jWUFFt~JC-{P$k3BrG<KKl)y+0KM0Ivgo
zf5sDGss||Lr|=Fh%-%6Ky?~b91VDypP^BoY;IDofV3kvv*FxO~Gxd%Be!RN5l#f55
zeR8wkH&tPb;hUZWjCDJGz?{v^otbwJ<oqN%!AK?#v>PG3-q3*oImi&Mz9``QEeD9T
zp}gep-70=eU>z^^b0jw%5N;Lz!ir#g7bF)P0iRw-9x6@T4tq%Bey|0;1Wk-U8oIy1
z`rLWHe0^UCD9&oQzD9aaSax=Q@AD@{0!(9CM0BzsHen&uRUlh-yith!>TMUGW0mNr
z${HFzYS>;sbBl^i@6*b`Z`}8xO<-LPY3rozzmW4u_|<z*+lB({q?9SVJv;5P?8}Nc
z)!2P>eCe(W6pNVZGOdx;ZW?gk6y)w&fNu=54t*_`&N|+`BhY&#pKOwf%RV^RtS{cF
zU9pYHu@a<Aey*2U#5=Vp{PD|jl|QNyUz83?Q-hU)>O@T``j0`H+A@!c5>_Pq_;9xr
za6Lp2=2*9mi*1u0`-?8qCvkRhm|#<fp4u?#Lm!HJRc!cuA6bO7=Y4^!5JXz1X{}A>
z2IKluN055Zrnm?}hxK|GZnuNo-ghw3LzyRs=0q!5LPv$1viXa09fd*%UF+7<ucqDx
zm6V#ROs>3Dr<+Mk2=ab@5b}k8)`(%)U7yw$c&d<4ZtL_zy6oGCZHQJ+ZWb7Ej<Ly`
zF1Jk+>o}ry^)bc)-lzuqyjKNcxTOJ6BlU;5S5L&7{OZfphw@jc>CS}#LH*ktK3pd(
ze^58qTg>aaQ*KMj4Hir4_J0sB5A`bb`jji-7;NR`;-IZKjirM^fxDn#m+mp0n8GEr
za_B`R*tEL5YRz8_GHG5aG}Lw%lX<suSXVT|uJ%)(ktGSBJy$m1b{k%&%jWH34kniJ
z*Lfq2Kdn@t62&8-oT9fjuCBE*FsHHiu7w?~XKQ$gOQQ_=E&W0r@xI?v5LOI}`qG{)
zJ|9(3F_Cemi8KH6OX68FN{l03`xmm7Eb>nh_rwe!gxSWM5Jb8z+2_6;d)(-QV#JOJ
zLFn+`J<khxKR&NjyG?)vWk$4N$Yr%PKuOuI3NDvA-$^sb6qLEXCaijY`tEJ#WSVrf
zHb51F+B#dBUQz-N7$D3+fmO^e%w9u!aHyl|L;Lh-@oA;8{FGFsMfyt;a06DCj{wh)
zRwib7a^fpJuNMU>4NdWs?q?>`A>T}e(dM;_NsON2aBGY*Ct$hW;43B37bfHRZ>Vw&
zRE#*^d@rng4sw|(&1cn^lptpf6t74VEfCce__DGUc(}e+jzfzno20*jA0L=~6CwvS
zx|ZsrCGHVxU}21Zzd#Z=6SmboS|KEIwaefiMbqddVDPHE{pkqrA`xxH`4E!(g?3N{
zPhyD`uicKsFW94-!1PEs=kn|_wm(Dr5kIjgiso*X)MbT-5S31QbC)}gZ2~T{WkKny
zGV;#M#>_np4W^kNa-a5YHc_#C!fal!&iyCFBLe$<HcePtB~53sMY?;)_@iSX-%&I|
z3RcXHZ6$#LDd4Bc5H9&gr^{6}z4S2W&=b)C*OM>AX6TR#;mloaBTL#NG0u6<XF<(a
zNc@M8emR}ipwofm^I}}m$CJW5N}j5j%sH6=6}d5#ID$C7Vx!PMDWYRPm8kP7&M%+b
z7z)FA7KcZtK4aOdJKgZmh5lameOQD`GLNb50B6p2&V~_EcEpU(?WHugR5+rqW28qX
zMq+iVM;k__eN2~<#+M!iwU8$zc`^d?rRZkDT%F4t<^Ocn{8Cdc`3Amqx~0KhM}=NV
z=!QnD2q^_qSp6K45S|=*aM)^Lo5)z+)n9`Z7BQ>6p6^G8tt$|`Sz#CPr)_7>N1o(M
zM&z_t$Mh(5E>w4HYGuUjE<NRM9aaoK!j3097ptV-Uu38?+VS}oudh)8^IZ>TavSF)
zldOxaC>D}cSlUb7c=)LFOUL$P>h4z~R9CF(t&*U;4qFDx`iTVqv}F3%*LUfO<aKI7
zs7JnH&OU5@3nap8YLBkQXj8wRkKCC|EVS<l@<(q(<a_T3L)Y|ki~3Qup&)v_6kUoQ
z)#;07!I`@&@?!@-gJm#$k%9Zh8mxJ}9#oxt>u{vz+MIpw8u0`C+c22#FP#1hv8=wX
zsw3Nt%BfTWPB4KOOvRNAry+1uVs(Rg5ia3kCztJM^375{Obr<=<v+DHB4c&&FjFqp
zcROfx)XbKQ5gSjM?l`SE8hWAhO7Xf<fhuaM8NO1jNbj)BiW%z7;%}>;akNTD4|HG=
zAYI2giTTBGND~lf5vY)<cAKWMJT1sH%;Rvt`%}x#{vaYoNck7HC}U%(Wv`Uow`*C)
z6~Uic=<CpVEBS!$wQ*3LYVKFBU`hY7$`129(PpVP#~MLi*WaE0la-1`N9X+8sS5^`
zKGXpPTZwVgWTyTNuBNsZ#I>ic7g_N6N7^-SqGNkvQFjM#^d)l|2W+aaID;qk*NQ=E
z``XK_3&b@Ykjc}|=;}hl^4)5Rfg+TLeh&i~I|tFiT*MEu!=8t-1f7`g#cwxb0^Ij7
z_}l}DhYzjNL?<WqSuShIyP^I#Bp9|sNOz=??GRSF6GK^%I+z!Zm}50%Ssh9qn3i$d
z8GfVqy{Bw62NdRXYe#Xt)v;<uXDtp$aF|!xw1j@xT4lX4Jtc+Y?0HL^7{HL&P`%=z
z@2OoV#+Kf%?LMje52btz^Hfp;(x!UZvTm;$GIw9-r+zUp?nEgXGNgmss4ZkPSpc>1
z8$?$cH93=buO|+l`(t?GV5oabjD&>SeuO3|8mV16kOF1~2i_hEqdwcfKyC1S!-TJ6
z^=}b&NLA-q%Npj!Ic^-sr>ws)(XQ<%ze%IP=J5k3fCf#I>m?3!B9FMkrB5Mp>Rx`%
zuKe=aRy$j9Dt4Oy(WC_r-$#01OYy!Sb0O8f@nrpa+Q$17qJp<EczJ5r9vesWRW?A1
ztP}MlUo$^MrbQEHWk64LxaQfrlX4f)+^N+lp=)^(*;ycqOGB53a-D9hUilN{&@z(e
z6~{pv1Ak%$BY1T&w2=DOf#Bw%+5RSanlAM}$-90!d!pknv69flx#G)Dgiga$^qs7q
z$;Z59JeG@izK>0mIo)b*DcwWj4O1%j#&BZ_aZjWgbkym!0^gT)^V_@7uz5k3_r#T1
z`4XYc8Z~=feaH?)ZBRxNC}>8<PNQD2uj!lKJutPNF4Ey;_w$!epL@bN5+p$Kr>9a6
z-ODqtotTVK3=E{NPTeiR6WXfVpOA@*WDLqxBd?o(HjT?>Kb7xtjJI}KJ-8&#QZKL5
zetU@x`a|K6f5Hd}G3wKJBw>nPIsJ4mTJbZoy2E67Kz96@<9oC8PIp0P@`QdiS~e?p
zS}Ym(j#nEuJ1u^La?Wd9^YLkrK%)nNOksV9^6g!241A@be`fuITRFVlSu?%sSDvrr
zL3p@d9kjd}CS;u&$qytp2t^il`wGF+$_!pdVz7R}E%ugXP?@&Ve{<30CtfZ#v`_w&
zY00#Knf)a$WxAW313aqa#g?wAfGGPL`i32qw#{$sJ2JMKdW)D$P4u7<Ndj6uih6m_
z7=ZDmAzc9T^N->b%kD^R2a$BDszvh!9Ll#=D~f%tKMhi$lOg!`kWW2cj<L5n0o}G7
z*F~ivDw@|%J4Y=-2~r@VCXgVp9=~LsA}zH0xZ$wW+|puHAAv=9oG}eUWykEky+E;<
z8(u2Q)%~<@+f>u15YumqXs5?>O3WGV92Ty5Gsis7N;Sm|V)CT44w0jt4>Zz9Wxs}-
zI@-2|gaDe06^V~DBa1bRIho7mT9s}hC@&Xr6Ui0VF$t;Th>h2VS>C>9<h2`Iipfi<
zI&Okj6tBDTg_l1a-Qqo-b`k^E7Tg_CfLcK&vi8`tT?9^#t_Tx9n$YfLE;1!V3X&5G
z!mmhc#f|4~xh%MONw;<Y0r{0W+@Y(jr6t-kwTgIp8gYu|)3k_ue!dYK!Qew8zH29G
zD)(OJskXJ$O<z~M8n@Oe1_aW7by!vvxi2VtC@w280q_5CQUv3|;@bA?_jTMFPF&L(
zU`}9h14@`eFjqYol*!1*^{(y5x@U3kXW8%g5GOnn^gSU%Whc)xDY^=zC!9)lTUV^3
zlNC@z^|U))Rp=Vb?C{V{eXGt!Faac)QTKOXD)-%r7BaCuIJjygVUapl)aQMJuamd#
zp5*0aPGZ5<lk+n-CUW`<)?ZnrRG9n*HAcdsyl~xGNfMAbk%Z&~F4?uF^(2rhIQ8Qw
zM%r#Tp%GL&zoE12_ZLUCx?YqULoHxRw4<J_x#}IIiqqbACa{xwlUt&^c^y8ku)8Y?
z$<?N67Pmg_txm>Kt|v*0XD=_xMtfnrC-S*<1Ww;>hVaX7eQww#37d6kF9L@I=`;{{
zO|ZjVSr2yL^;aW6r`XoKz?v=N+{jD~9FThKO_rs97m5Rnebtk;Fql8%(xX)0(N38G
zl&PE#ZYyYBD#(ka**IVxzt&$q{|Yu99~O!sNu!Yot;xLt_a@*?qb<kGjF(8d|6#yE
zgxME9CiJ8PbPMJ%-I&MCZTJaCYK1XVn92T*vF_wsre#9y7jUfo#-kPbRm@J>q7m9&
zrvHzGjirBm7mn|>CFghLW`|BpcvHVI#%b(IhQW-YCCKehq<(QH$t+91*|qU#nz1>~
zCTNRC)<rUFCRa|SP>6iKjc^3^=~btf*K=bfLYQ%9m6wQ+*@L54*);HRmtWCc_8$p?
zk#DiOF;CQFN{yCiR1>l*eZFoTw0QPAJzY6{(+&)%Ku*tpNiENFashJ|<3X+nV6W=Y
zV@(k&67MkArcv8)5%)_YC2M1;G?d54kNDvUo!^~q)mtw?!Z{D8*339(zJYQz=i+z6
zdH`KodOVgg8I72L*cnaA;it_jl8+ihf=xF`7~fjY*()6Flam36OkDW1&P~H;kNbx1
zJO@36&phnRgY6ksCFkPk)7;K$KmO*sK-MZ1&eW9o#l!uwzr@*)zwmY-wbDIQ+r+<c
z`KJQ39G;_sXS@0i^+WRj3G>%8wDcwJuiEt@MUq*(9H8|s-2D`cWPN?xtUz<M@z#~p
z<j35Si!u=$6S}yD^X>+nXQ!HCGa^6wqphMudWFIA_2L!{#|+`@@6&S&eivOmmG(BX
z=Vfs~5%~1n!Y0B$CCTqj7{yfY!R*`^az<WrXq1h(`g~qzd9~k`cXPw^2`xJ1W8@nG
zv_k8|nL+e8!&yv31u_r9i(W6MY6lq}s+2~H6{!2zuHcoee=y!s6O5h>OPOqP5c)ld
z0@>tEfVYV&>)hg7KkY`MpBW!Q__R~B5xXF`?|c2bLY+F!1kaa}y_TjEl&++pGL9NC
z6I7(G1u&C;c&>?Ycyl*Y4w2wz#V)m!rqpb7rdnoCscAB)uE$sTGUl{rc=C0<(2!_E
zo7daPRP)77H)+!u^|bD@o&`YdO*!|JhTJg#)iWR#Anr#@UAy^tW5~CoG+Ixq?rTTk
zC4GdBFHt5@CgQnR3S#C2-SW-@NTmqxH`z&b->L9_OAzLlH_godU=-CVc5h``+z+To
z_lTL|9sJ}Z0kfK&gliCs+DMG(S8MOp@0Naf+?ykqzx5S4%TL_7L9X4^1N_8QNn|)L
z_#OT8`%>ln!Ki-Nt5;H}6%UubS{u^9@e}l90Uwg)-Jxaj&)f8!7`0v3Tz8J_lz6x~
zK1&q?DX9HV$E<|lIt2A=iTYY5mzXJp4W3COW<14mq?=B`5OHt)i)lItlZw<Bl<SKa
zBGyk@fR@OB>XfRp>$`}gLKz!h^KoC+kH7DMj=(65-Eez^9StxUjB2gYuu9M+Lu)i4
z<30#Fh_RW_1R`{9&`wD@vm+QnHbR_KOW=!a@xB+k4l!qt46;>kC?PX{p>+0B&~=wS
z9ZQj5(~8j*9c~(}i`L?EcN67#zRX4H=8qZT<H&o9g*YQS|56q-B7v1teEP$Q-4`!?
z)H?$(xA>d>%r?EK9)RC-{jYEcdp92AnlrJy5i*hy)H|D+2tKy!IsTIZ@J8wWqZw)W
zh7sm|)|E#+;(?qCsr_AwQqN7kd}q)#8ib8O$2X;ppkMBf32WV5Z@GAU)a}R@;eXcR
z9nzNBQQ$D!QN1{{#`EW&$>>0Ee6o#}_qV_h@`@{{4dJ&>>k)|=G(Re<gPX?~Zhyz6
z)C4hsj2X|izC6oEpKSpte@zHlAo((@G}WSj^sS!j2@;Wuy|5dUD__eeynprW|1cZy
zq37JJQsi_e(`zM^I$yLdE~_+t8DvyjivrIYqb3FlF#WFf$i;X>B`0CqpA<@WiRGh(
z{MM9Vx%v~=uT(Q6BI5T;fmF0FH140P_=gi*M#>b77xibsStI-4l{Q@xb+2dt%~l&+
z==Bvun!{K5#!2Rh?0_{70`FFnA#Ix3Bpy1&K_WxG2gMcOrov!RYtcyDW?iY0evEiK
ztiN$fu%skc8=$Tvd{nm^^RlT88{QjTXIHb_YLgQ=&%27ME>&=m?})OGZtq@+zi?2<
zZIIP)=#V+Cy50T-1Yha0n{2L~?tZCvLAOhcxTvt4C7W~9Zh)fSw<>vWtiHaE7q96N
zCY{zES?FW{UMFO_dr~$Y6}1XPI9|=PEo+nr7o6}V^0|6O(G^-u(8%_D?k<)<?RPKG
z8s%{c$XBu!Ha0{Z%)Ho<_~K_=-(G{}L$RQqHM@s(YZ0hnN8Wz2P&LDJ*`G4vf<c52
zB#Kmis)%iYwFjG(6cDkvvYYc!BIrW>ZF!?junvhMnmeyhhqaE2mDwK%a>LR3Dc7h_
z9SI+PQXuAOr6j~UlI^Mi*%8hkBg8@87{O|nz3Z|_2piB~x9#!GfCbt8dI%K4qOJZ%
z_RvwdI6;!5oFE6`;A)Os=qD}jKqFM#-~mVp&h2xaG8)G%uO_!uFg>1*?UJiGI>ef1
zWfrKI(Drxg^1hIDmm&L1aGoJ0YKX^!U~RrQUMc6uc8px_TV=QCj8udtBx>dTS!Or?
zo?RBZs9n0dfS=F&`=m9F6yVrAb7ut~6R9NUl`~SGxvpD?UtEz$%o`_334Ab|Ou*Fr
zQT0aOLHXy+*<k*md;(Y!L>Tj%dKY@i6hHASe9K*d`_@YSTB^-=j`a&skwfO&rbJT7
zjEC?gJdyj=kc<}M=+7mb+zDG3+25g;s!mD~vntXq$!ySbo>UOn)^zhNDkpwL49hw=
zy*Os6r`U50^&6ZdZ`eU1O?K&RhQd7V-_uOZTmBNqL4eUpo8WH5c9(I;8<@zoDsZ?B
zN244k@N2fpR%+KAXaDjRt)wsG6>IO=D2?afA9RVclXZ;CHyP256rp$UQ@XP}Hlkdr
zz)0er35PPWrM71BSoEi=i^lKYYzcE`Z0w>(NNJiOO?<adG6pb=OcC0w2D8;moxB>h
zda3if*l-;vXD>*EYN-kf0&7W+D`ra8J+`2+fdcgwTzI31Z>bS5Xs2J59lL&p<KpKu
z;#&n)$o|n?sXEuA(Z!RL+J4cO?voXX9EcVmo5sypKmi*&u)RU{i3Pr(Wt0twjk9Bv
z4#={%6zG(;RS|QjQmRl%k*o)8HTcL~c=|YL?FU=?>FVe7B;{wmX!*;z$2Vy-n+(lf
ze@0|e>s5s4OK@}!M4*ixl%9#05qIdBckoI$=S(O`FVPn!<c^o`W8%&N!tCryo$MsL
zuSA+J&!O#TOCOe`w1+_|5&p^uJCz=)F=+8*GXFg-tPtvnE_SXC_o6d}<6Blo!@9J9
zNCb!DD5vlv>G}l&rQjR6g*Lp(^0)4?mSf8B=%_OpgAfGw1SwJVHd~F)1dg>m9Qc*W
zlUO;P8P%|In`PMg&eV$w6|yOst_vuL7d@Lq8|1>c3)JByI5naksx@>q!e{h>=j57T
z4rZL!1DO8VQn|j-M_JTPNWs<$*g*vhnMo8<{*T{(FaN%Ex(&*FB~-uqBc__QalQ?i
z(f!G~FB5eI>E@a4RMvY*%87Yx#~yQDaBH|S&4<_?-O{j0V$*=UICEKe>)3>WLSB1?
zook$w>tIHGm%V+rz|Nb_dDvGw@>?5WvC7X<I@Zi>hd-JlIar1yZuleeNyLbySI;<&
z2tlg~Wx;fU=G~2!IClPf=_1D17?S1-ad^9#H#c?je$*6NI`%uM*J}Gz$~V*<cn?v%
zawg+yZYA@QWW;hM>Xro5ErKZhJFqr<PWb_1;`!<~QdV{roC-Cr@lLyb`*>b<giOY#
zQBml>O!;@PYy}?JM6~TLD8L5=Y6NS2VW4rDB}R+ay|geu3d4y@T&*1>wzojrD&=jR
zIxUIw*5F;b!aP&`f>7gRC~z%cwnf4qqd(SlH@&wdi6I49Vvn*@Usom$*WO{M=+JoF
z_WamYp_m#)(G!0q5fu`ZvK?d~If7hVYtxwN)^!;^wTo%3-WoFosJhYP8)>s4kCyXv
ze&qTbBf3x#Hu*amw`0^*nco7CQur;o2RHh8x;g2B6>6U9eR6DztfaK}ZF6zb60zC)
zjN+YjGgnGZr2-9O=YS$l*l7-PN#CkOF;zqO@Y=euboxahaC8m#d(iHLP*5zwZ*j3R
z&d<Ni;%by@AMFoVp;wjX;OZZJA<x2?Jj**rXb3l<Bvb8&Zzr{q*ujoG_-bI$2qvw2
ze=tvspZnukUYqeOA#7lBkc;<fBO0$wRD?952tEj~4<!mq+h!!O?DmkcIm8bmWOick
z@^<$(Bs@7ta4$SW!1zx*ly+J$=hRr<mv-ZHfogZE`@0A?V~@P_o5+0cDHjkWs4GJX
zgyFg`b7V>0Tot7WspWr2OG`%(ca3WANL7DP5P5Qub)}ev5JeqqFIukB$oQ&g)`h!=
zPhs+Xf`4^z)UP(M>*!U4h^vBKgi~d<$h>3L&qCkl#MN1?;`jvSeD&VbF{b=7VRN88
z67dK(SKui&gpp`tYj8NOqL`M2H;C26hL?&O-OuYtU?CX>J2n)O%Vwl1b02$avaI4L
zX0|Y*D(WoWs}#;hXe!T%vE-G3p9x}FOps#wHMjN<)V|ODkx}%5-4D6pj`)bL^p;G*
z1?`cnnIhKm58D1sht<!Ham{9ySTge-vY#6M*s3lBonx6qKuR9bKScregH>?l<=B_)
z+MABhACTCSMVm!$u{9cN7&UJ2A?>QV>my+~zF6yha-2!GG{^}*5s1bjEV1oA5JgIE
zPTgaKRXN7T9uG$UBgC8SKyN*Esrht7t+v+rK7tN=V87Q%tBNrsB&@{-6B#JPO*xZn
z<78r|liabzk1EZ`@xmfIcQ_~cA{}f`;44KAcTc7DZU$HU-a8efC}&(o%XU!yYo018
zPwb(cQs4kWmXew%qMy<%yLU0Bdv+8}$GBe1XmSEmhMGVRYbB}nHgQjDS=L%%$WW;w
zp|ZU5d;eV)p_1tU3&|iYg#u3{L05od9gD7S7P(&Et0Zf9bcUMXpU#XqfBBi<UlXN2
zT!mE_SIlmbAZViLn|LQ+v6p1&A>%O1R^ffzBiTt0%XKZyL5bF-sXJ22j4QH(>n9{O
z*ivU4wykbS#fQDS25yF!l=ZCxtLOsD0wnG~<jb@&6#$OfE7x~joT1C!hB8EKc6sD&
zGc8up3Qhtihhjo=O&Gwciz59cpM(~2vStQpsft}1vtmVHgGAwsswypINghH%KUoi^
zP)bwsm0M+4Rn+pxCw~z_Ew8Z9CQ*}zz{bgHQV%wg*;@9Z%g<`^6Hftm^q4+U>8HNO
zQDL=8(IN(GIPu66c3va~ELwD!N$J`&jkXmmO8QZBx+V3$iK8GxwtUkwp{{)|NZ7s)
ziWuH6UF>C~zNRD&^W&mn+!a!O>-0_Rs_HYE4?H~%*^HHr#UIC)z^(E30}&%?T+5s6
zB%1S$>fSu>o?Iy_d!M}&kbhlm6zhH!2|I7Q8FRf&2tAWQz4gs=%$%3IQzHHfFvkEw
ztY^Xyl$Sb`l?U3iFRZ+>B-a1I#k*hG6?GbLhaW5FMDi39gvM5A=IU21ipZmg3XaCy
zOZ9HBU7R&AWI9+3qlmKp_I$SY$7%%Gpm8<h^1wKIkod+AGaq0i#vqW!Dx@6iP?i9P
zuwi%vUr+{kNI~pxtP__^Ds6zJ4-lMS4eDim;EahbsQB8bSpATz4`gI??$JxmulpXf
z;ESB52Q9_kARq6}b8Jyx9d^r79e|^+;P|VRkj^#ANSCIjN626KjYZh%hH2`v8#d=6
zbOzd{De>j53tV)bk}X`F+VHYo%^MX(9pfbU6839UVQ_NlP=b5JPWb{_V@rrF8hy>k
z5qn|}ABtR$zFXEK%je3HM$ubYXuQ29%iVQxD<pLpQ<mCH<M2<G94fupf+d-W-$4FB
zRymIeyw1jO6T+IVEwn~ou+q|ND_=w#jp`>$Z%Gw1CF}*4!9z!DB5yd(T|Q;ge;?BL
z_2@V$Hjn{6QZZP|PbxZl`#CbSdUAGkPGJ8qyoR}v+yCJ{Oe<sBO6IO9Pat#ND)uY1
z-%9Bhsqd!lp)=9ML(nK{$ED53-;#hsqiK_rH*vW);@E~z76=KL=LtY;_f+`x<7T=h
zH>c`J{Rr1`dN?gssEW{Ql)*IeBc-rFE=)`nF2)?BK#!9mZ@Gd3kq1-OX=N&^6-QFn
zR;0j`s#sw!Nb6o7EU7WBrnzBguN-`{#7VH3-;I&O1i*ngyq=%REy$;DjZv_E^$aW5
zCdN1#^QBXNmOXXZmJ0j*T2I|K6(K#iEdwRuPx$*6%rF>wRc$ceFV#t$O&S!etGZv#
zG2v7OqLz*N+m103kBY}wSz5+Z%yE!sjrMeqkg>Cg4(;P2^J1TAdAl&o%)>_w)m3#P
z_oQz<&hV$ToCLKQ)Y*PAd&|*ob?Y*n2DFxKowi`3B6KGMOXJ$0*)D)ZxPJI~exA&@
z`liPt<^9B#>*KBCo~rm{o;t|Sv?*Z4bBv;2yuG?q*}wDX<I4@pMc3nvoq}p!aOiox
zt4~`~&-t<f@K~qTKU3fbxNuDMdHF4VpDry+&s6LL11x06=<`KyD(%5gRHSvn^(Gg6
zSIcI&ltM;fesXZko=bW9x~ZmHn0+k5fOdif<6VGox(I+sTe^U%zJp@Q+z?;m67P`M
zztB!EBEq`2u-+UQFeEJLcwWEsAwaUG4OE*!h;K+Vqgz)#*pRB*6Ug69dejJgywRJG
zv1FsjuIE?^Kz^ohJKF#-5^PPuD?60{IqFR8#FP7FyCx|r(3+2rDI`l}7tpX4k6qC9
zpY;YJ4IlX|j8L637Fdo3S*&N^jp+&&)5LX`X!clG6<v314{Me@53@Faiu-9Ti6at9
zC239uvo#PYcYDjWx9+QTC1)|6j~y^r#z^@}L+C-zc@~vpmq{&rsN#K~)*aiRk{h9=
z40YxjhQ@UM_VMzRQnaX_=KNr>N#7Zw7*G)y@4rEqRFBv%het&?LCv52K)qOiy<-Fy
z56KkU>KGOCU-a2~4LpmBv0RF<2@7=GAIZy7lFHnZo)Lef3OG$014Lw}?RpGZaIyN%
zt>gBcGI2%ci+YRXBo{O8Jmx(SJIt@rIw89tY6_U%!tY%p0fnH6&E=RCaUNHgGs-;$
zb5Y|j_6G>#`DczL><WKk5Cirj+SzXy+zbdEZHZM;gnczx7$q~we>DFHw4pxdtUt;l
zA$)E;&RLiug$?8F6foXWVZDH)bNx-tBh*-Cg%Y7HTp;vm0@n~K(Ir3AjfAX474M93
z7@aUhd$8kk_zB5;-#OEoH8DXx_V3bj2M{xRC8@Gp8ng>Fjr0E2Ffo9umoc759g8Ex
zZQjKRFRDD0Gve3gp1OXe71F?3E=n)#V&~7Ubg*CqzPf}J1Er((019woBy_zc)3OAj
zMrZa*cvx|_-LNfdS!lQu)cmGQ+C-k434(Eg2$k)+T0O|2Ch>fu-}w<hr`x!qx;AWU
z-{2csTA~7xCj2Wh@A3eKCb6f|tC&hpdRowLO;iM*kaPC(lVw_YtXj6i=0!CQ{3T@i
z%r+8n*14M;SRE<qCKL6|5H%jS%r8l9&GYeg>hX-VV89#qZ?SS-6Z1VtiE4M(TMEjT
zm~Q^(#yd3A&Kamric*GNxTt07;t-{JCl^e<y#+>OGE?tIx!B}si&-!2^6pLib{=7b
z!L-U;Zg(g=M)3~NO>MM2n(ng**0hXjkFcR;F6(J)sQWdY4Dm8fMF6Nw^)pl5oV!Rm
z&Xz}yGNBL$>8C$TB7!_ftvpaLzAkVm6+7DmTyyl=MXBOhD<w|6ko#nQMemSMl@X%<
z3GD&@Qi}f~3;&58Dj8d;x!C?UhsXYZ#9<Xbu+<Mx*2&og@K2!kze&9iRTb+s7Bs)D
zn%18&=yKE}iAcrM6v1xRQZlkSG@ch233#*I%XG(YuV)lBGS8~Drhm?!1vg|!$fSj#
zi(!hUc!VKkNe$sOo|w(e0cNLyqF+2hWz+i~dBWGK^%zBislO<i1J3%agDHzKEOqZm
z(FZ4PUygo${N$U(CbDN%$0gZrd@ZZLVZW)U<{4=IEvBufy95c7ARLvpb$L!Y4bN`!
zhWt4q%<NN(0B+8hthv%E?kM_|dBa(mU|QnX`}mp{rq__52<={p7?0+IdxKnqW4UsF
zHhFdOb`di>cS?Z<F18jkB4V?8^u>1v_!34OI7(KshAO&mqI|nyTE9?^R#^)fcoJ{%
zr67yF_;N!fQYVXv2_Keat(YV5O%Sf_swivX!pHFnj!Ji$zMHY_hfl)kkv0`n6dedO
zd1g)^Gt4>b-BbvlgXqPSG)b7m8dhc(tydG_mrV=?i?yNvZUxw<g5ioUk1&Bb{p0AY
zuFVx(0_i8wm#x0VX(agY<SEtC8h@Npd4#Opz{^iKH{Z+i`lTxL2fdamneGS!XWh{I
zuLVAPTK&L0kNo+~p)g_M7S-!yjP-wLl7H(Fu@7F@#nRqR{DX`Bpd)!%Sy+J|peFYR
z)y>Vu{)P5~ivAn%0YcNLgS0+|3ihV|=U8QR2`wf`R~s8+BRjhf-kL_;!qOS=@p3RV
z`&iBZ&;U6(f1CzjV*;`RIe1t(IT%@)I6t6tb`~a9AVBx;Apm(xHxS?>5f)uqfVG2>
zi-j}D2;lZ#D><3im{>pZP<FC6bv5~W#y|J|*B&2J>P|)u4j&{tv!s!Y^S@)atp6tb
ze*ncl68OIiivI%u|KRBVzhb<>AgCk-P++KFhVqcWATPnW|FhWt58(WR5I1si@o)l}
QA+Q2D5hy7olq3=U7fM=w*#H0l

literal 0
HcmV?d00001

diff --git a/services/clsi/test/unit/js/ContentCacheManagerTests.js b/services/clsi/test/unit/js/ContentCacheManagerTests.js
index 35f22fffae..76c517d8ce 100644
--- a/services/clsi/test/unit/js/ContentCacheManagerTests.js
+++ b/services/clsi/test/unit/js/ContentCacheManagerTests.js
@@ -1,201 +1,123 @@
+const fs = require('fs')
 const Path = require('path')
-const crypto = require('crypto')
-const { Readable } = require('stream')
-const SandboxedModule = require('sandboxed-module')
-const sinon = require('sinon')
 const { expect } = require('chai')
 
 const MODULE_PATH = '../../../app/js/ContentCacheManager'
 
-class FakeFile {
-  constructor() {
-    this.closed = false
-    this.contents = []
-  }
-
-  async write(blob) {
-    this.contents.push(blob)
-    return this
-  }
-
-  async close() {
-    this.closed = true
-    return this
-  }
-
-  toJSON() {
-    return {
-      contents: Buffer.concat(this.contents).toString(),
-      closed: this.closed
-    }
-  }
-}
-
-function hash(blob) {
-  const hash = crypto.createHash('sha256')
-  hash.update(blob)
-  return hash.digest('hex')
-}
-
 describe('ContentCacheManager', function () {
   let contentDir, pdfPath
-  let ContentCacheManager, fs, files, Settings
-  function load() {
-    ContentCacheManager = SandboxedModule.require(MODULE_PATH, {
-      requires: {
-        fs,
-        'settings-sharelatex': Settings
-      }
-    })
-  }
+  let ContentCacheManager, files, Settings
+  before(function () {
+    Settings = require('settings-sharelatex')
+    ContentCacheManager = require(MODULE_PATH)
+  })
   let contentRanges, newContentRanges, reclaimed
-  function run(filePath, done) {
-    ContentCacheManager.update(contentDir, filePath, (err, ranges) => {
-      if (err) return done(err)
-      let newlyReclaimed
-      ;[contentRanges, newContentRanges, newlyReclaimed] = ranges
-      reclaimed += newlyReclaimed
-      done()
-    })
-  }
+  async function run(filePath, size) {
+    const result = await ContentCacheManager.promises.update(
+      contentDir,
+      filePath,
+      size
+    )
+    let newlyReclaimed
+    ;[contentRanges, newContentRanges, newlyReclaimed] = result
+    reclaimed += newlyReclaimed
 
-  beforeEach(function () {
-    reclaimed = 0
+    const fileNames = await fs.promises.readdir(contentDir)
+    files = {}
+    for (const fileName of fileNames) {
+      const path = Path.join(contentDir, fileName)
+      files[path] = await fs.promises.readFile(path)
+    }
+  }
+  before(function () {
     contentDir =
       '/app/output/602cee6f6460fca0ba7921e6/content/1797a7f48f9-5abc1998509dea1f'
     pdfPath =
       '/app/output/602cee6f6460fca0ba7921e6/generated-files/1797a7f48ea-8ac6805139f43351/output.pdf'
-    Settings = {
-      pdfCachingMinChunkSize: 1024,
-      enablePdfCachingDark: false
-    }
-    files = {}
-    fs = {
-      createReadStream: sinon.stub().returns(Readable.from([])),
-      promises: {
-        async writeFile(name, blob) {
-          const file = new FakeFile()
-          await file.write(Buffer.from(blob))
-          await file.close()
-          files[name] = file
-        },
-        async readFile(name) {
-          if (!files[name]) {
-            throw new Error()
-          }
-          return files[name].toJSON().contents
-        },
-        async open(name) {
-          files[name] = new FakeFile()
-          return files[name]
-        },
-        async stat(name) {
-          if (!files[name]) {
-            throw new Error()
-          }
-        },
-        async rename(oldName, newName) {
-          if (!files[oldName]) {
-            throw new Error()
-          }
-          files[newName] = files[oldName]
-          delete files[oldName]
-        },
-        async unlink(name) {
-          if (!files[name]) {
-            throw new Error()
-          }
-          delete files[name]
-        }
-      }
-    }
+
+    reclaimed = 0
+    Settings.pdfCachingMinChunkSize = 1024
   })
 
-  describe('with a small minChunkSize', function () {
-    beforeEach(function () {
-      Settings.pdfCachingMinChunkSize = 1
-      load()
+  before(async function () {
+    await fs.promises.rmdir(contentDir, { recursive: true })
+    await fs.promises.mkdir(contentDir, { recursive: true })
+    await fs.promises.mkdir(Path.dirname(pdfPath), { recursive: true })
+  })
+
+  describe('minimal', function () {
+    const PATH_MINIMAL = 'test/acceptance/fixtures/minimal.pdf'
+    const OBJECT_ID_1 = '9 0 '
+    const HASH_LARGE =
+      'd7cfc73ad2fba4578a437517923e3714927bbf35e63ea88bd93c7a8076cf1fcd'
+    const OBJECT_ID_2 = '10 0 '
+    const HASH_SMALL =
+      '896749b8343851b0dc385f71616916a7ba0434fcfb56d1fc7e27cd139eaa2f71'
+    function getChunkPath(hash) {
+      return Path.join('test/unit/js/snapshots/minimalCompile/chunks', hash)
+    }
+    let MINIMAL_SIZE, RANGE_1, RANGE_2, h1, h2, START_1, START_2, END_1, END_2
+    before(async function () {
+      await fs.promises.copyFile(PATH_MINIMAL, pdfPath)
+      const MINIMAL = await fs.promises.readFile(PATH_MINIMAL)
+      MINIMAL_SIZE = (await fs.promises.stat(PATH_MINIMAL)).size
+      RANGE_1 = await fs.promises.readFile(getChunkPath(HASH_LARGE))
+      RANGE_2 = await fs.promises.readFile(getChunkPath(HASH_SMALL))
+      h1 = HASH_LARGE
+      h2 = HASH_SMALL
+      START_1 = MINIMAL.indexOf(RANGE_1)
+      END_1 = START_1 + RANGE_1.byteLength
+      START_2 = MINIMAL.indexOf(RANGE_2)
+      END_2 = START_2 + RANGE_2.byteLength
     })
+    async function runWithMinimal() {
+      await run(pdfPath, MINIMAL_SIZE)
+    }
 
-    describe('when the ranges are split across chunks', function () {
-      const RANGE_1 = 'stream123endstream'
-      const RANGE_2 = 'stream(||)endstream'
-      const RANGE_3 = 'stream!$%/=endstream'
-      const h1 = hash(RANGE_1)
-      const h2 = hash(RANGE_2)
-      const h3 = hash(RANGE_3)
-      function runWithSplitStream(done) {
-        fs.createReadStream
-          .withArgs(pdfPath)
-          .returns(
-            Readable.from([
-              Buffer.from('abcstr'),
-              Buffer.from('eam123endstreamABC'),
-              Buffer.from('str'),
-              Buffer.from('eam(||'),
-              Buffer.from(')end'),
-              Buffer.from('stream-_~stream!$%/=endstream')
-            ])
-          )
-        run(pdfPath, done)
-      }
-      beforeEach(function (done) {
-        runWithSplitStream(done)
+    describe('with two ranges qualifying', function () {
+      before(function () {
+        Settings.pdfCachingMinChunkSize = 500
       })
-
-      it('should produce three ranges', function () {
-        expect(contentRanges).to.have.length(3)
+      before(async function () {
+        await runWithMinimal()
+      })
+      it('should produce two ranges', function () {
+        expect(contentRanges).to.have.length(2)
       })
 
       it('should find the correct offsets', function () {
         expect(contentRanges).to.deep.equal([
           {
-            start: 3,
-            end: 21,
-            hash: hash(RANGE_1)
+            objectId: OBJECT_ID_1,
+            start: START_1,
+            end: END_1,
+            hash: h1
           },
           {
-            start: 24,
-            end: 43,
-            hash: hash(RANGE_2)
-          },
-          {
-            start: 46,
-            end: 66,
-            hash: hash(RANGE_3)
+            objectId: OBJECT_ID_2,
+            start: START_2,
+            end: END_2,
+            hash: h2
           }
         ])
       })
 
       it('should store the contents', function () {
-        expect(JSON.parse(JSON.stringify(files))).to.deep.equal({
-          [Path.join(contentDir, h1)]: {
-            contents: RANGE_1,
-            closed: true
-          },
-          [Path.join(contentDir, h2)]: {
-            contents: RANGE_2,
-            closed: true
-          },
-          [Path.join(contentDir, h3)]: {
-            contents: RANGE_3,
-            closed: true
-          },
-          [Path.join(contentDir, '.state.v0.json')]: {
-            contents: JSON.stringify({
+        expect(files).to.deep.equal({
+          [Path.join(contentDir, h1)]: RANGE_1,
+          [Path.join(contentDir, h2)]: RANGE_2,
+          [Path.join(contentDir, '.state.v0.json')]: Buffer.from(
+            JSON.stringify({
               hashAge: [
                 [h1, 0],
-                [h2, 0],
-                [h3, 0]
+                [h2, 0]
               ],
               hashSize: [
-                [h1, 18],
-                [h2, 19],
-                [h3, 20]
+                [h1, RANGE_1.byteLength],
+                [h2, RANGE_2.byteLength]
               ]
-            }),
-            closed: true
-          }
+            })
+          )
         })
       })
 
@@ -203,71 +125,46 @@ describe('ContentCacheManager', function () {
         expect(contentRanges).to.deep.equal(newContentRanges)
       })
 
-      describe('when re-running with one stream removed', function () {
-        function runWithOneSplitStreamRemoved(done) {
-          fs.createReadStream
-            .withArgs(pdfPath)
-            .returns(
-              Readable.from([
-                Buffer.from('abcstr'),
-                Buffer.from('eam123endstreamABC'),
-                Buffer.from('stream!$%/=endstream')
-              ])
-            )
-          run(pdfPath, done)
-        }
-        beforeEach(function (done) {
-          runWithOneSplitStreamRemoved(done)
+      describe('when re-running with one range too small', function () {
+        before(function () {
+          Settings.pdfCachingMinChunkSize = 1024
         })
 
-        it('should produce two ranges', function () {
-          expect(contentRanges).to.have.length(2)
+        before(async function () {
+          await runWithMinimal()
+        })
+
+        it('should produce one range', function () {
+          expect(contentRanges).to.have.length(1)
         })
 
         it('should find the correct offsets', function () {
           expect(contentRanges).to.deep.equal([
             {
-              start: 3,
-              end: 21,
-              hash: hash(RANGE_1)
-            },
-            {
-              start: 24,
-              end: 44,
-              hash: hash(RANGE_3)
+              objectId: OBJECT_ID_1,
+              start: START_1,
+              end: END_1,
+              hash: h1
             }
           ])
         })
 
         it('should update the age of the 2nd range', function () {
-          expect(JSON.parse(JSON.stringify(files))).to.deep.equal({
-            [Path.join(contentDir, h1)]: {
-              contents: RANGE_1,
-              closed: true
-            },
-            [Path.join(contentDir, h2)]: {
-              contents: RANGE_2,
-              closed: true
-            },
-            [Path.join(contentDir, h3)]: {
-              contents: RANGE_3,
-              closed: true
-            },
-            [Path.join(contentDir, '.state.v0.json')]: {
-              contents: JSON.stringify({
+          expect(files).to.deep.equal({
+            [Path.join(contentDir, h1)]: RANGE_1,
+            [Path.join(contentDir, h2)]: RANGE_2,
+            [Path.join(contentDir, '.state.v0.json')]: Buffer.from(
+              JSON.stringify({
                 hashAge: [
                   [h1, 0],
-                  [h2, 1],
-                  [h3, 0]
+                  [h2, 1]
                 ],
                 hashSize: [
-                  [h1, 18],
-                  [h2, 19],
-                  [h3, 20]
+                  [h1, RANGE_1.byteLength],
+                  [h2, RANGE_2.byteLength]
                 ]
-              }),
-              closed: true
-            }
+              })
+            )
           })
         })
 
@@ -277,53 +174,35 @@ describe('ContentCacheManager', function () {
 
         describe('when re-running 5 more times', function () {
           for (let i = 0; i < 5; i++) {
-            beforeEach(function (done) {
-              runWithOneSplitStreamRemoved(done)
+            before(async function () {
+              await runWithMinimal()
             })
           }
 
-          it('should still produce two ranges', function () {
-            expect(contentRanges).to.have.length(2)
+          it('should still produce one range', function () {
+            expect(contentRanges).to.have.length(1)
           })
 
           it('should still find the correct offsets', function () {
             expect(contentRanges).to.deep.equal([
               {
-                start: 3,
-                end: 21,
-                hash: hash(RANGE_1)
-              },
-              {
-                start: 24,
-                end: 44,
-                hash: hash(RANGE_3)
+                objectId: OBJECT_ID_1,
+                start: START_1,
+                end: END_1,
+                hash: h1
               }
             ])
           })
 
           it('should delete the 2nd range', function () {
-            expect(JSON.parse(JSON.stringify(files))).to.deep.equal({
-              [Path.join(contentDir, h1)]: {
-                contents: RANGE_1,
-                closed: true
-              },
-              [Path.join(contentDir, h3)]: {
-                contents: RANGE_3,
-                closed: true
-              },
-              [Path.join(contentDir, '.state.v0.json')]: {
-                contents: JSON.stringify({
-                  hashAge: [
-                    [h1, 0],
-                    [h3, 0]
-                  ],
-                  hashSize: [
-                    [h1, 18],
-                    [h3, 20]
-                  ]
-                }),
-                closed: true
-              }
+            expect(files).to.deep.equal({
+              [Path.join(contentDir, h1)]: RANGE_1,
+              [Path.join(contentDir, '.state.v0.json')]: Buffer.from(
+                JSON.stringify({
+                  hashAge: [[h1, 0]],
+                  hashSize: [[h1, RANGE_1.byteLength]]
+                })
+              )
             })
           })
 
@@ -332,7 +211,7 @@ describe('ContentCacheManager', function () {
           })
 
           it('should yield the reclaimed space', function () {
-            expect(reclaimed).to.equal(RANGE_2.length)
+            expect(reclaimed).to.equal(RANGE_2.byteLength)
           })
         })
       })
diff --git a/services/clsi/test/unit/js/snapshots/minimalCompile/chunks/896749b8343851b0dc385f71616916a7ba0434fcfb56d1fc7e27cd139eaa2f71 b/services/clsi/test/unit/js/snapshots/minimalCompile/chunks/896749b8343851b0dc385f71616916a7ba0434fcfb56d1fc7e27cd139eaa2f71
new file mode 100644
index 0000000000..bb9f891be7
--- /dev/null
+++ b/services/clsi/test/unit/js/snapshots/minimalCompile/chunks/896749b8343851b0dc385f71616916a7ba0434fcfb56d1fc7e27cd139eaa2f71
@@ -0,0 +1,7 @@
+obj
+<< /Type /ObjStm /Length 447 /Filter /FlateDecode /N 5 /First 32 >>
+stream
+x��RQk�0~߯�ǍK'ɒ��$�vkGIJ[(y���8�*����$�I�R�d���>�I"H@���9@J!`	V�/gg f�>BZx�J���9�ۮ]-B�'ZNg ��k�%�i\�!���������f��4����m݁���4���9Ķ�CY]\��@�����!�	4c�
U��f�=��JgO���g��>������zz�>A)�C9W�w�K�q�PÜ#���/48��VX/�T��p	�-%2"��*�B;�X���2�,9��G�z;��E�Υ��Jj/�c
+�n�%������ᵦf3��]!�����=�y��
,�s]��@�e+COW.C����kڒ	�c_ťX� v��>�N������2u��7=���}��	#��HV�r�9?kv
��6�����G�^�z׬.���v��=Uyj��ǡp����z��2
+endstream
+endobj
diff --git a/services/clsi/test/unit/js/snapshots/minimalCompile/chunks/d7cfc73ad2fba4578a437517923e3714927bbf35e63ea88bd93c7a8076cf1fcd b/services/clsi/test/unit/js/snapshots/minimalCompile/chunks/d7cfc73ad2fba4578a437517923e3714927bbf35e63ea88bd93c7a8076cf1fcd
new file mode 100644
index 0000000000000000000000000000000000000000..d503090d3a7389ff13f6c9a2106c66e0ecfbd73b
GIT binary patch
literal 10161
zcmV;iCr;RJVrmLJJRmPdX>4?5av(28Y+-a|L}g=dWMv>POl59obZ9XkF*G(gATLa1
zZfA68G9WoHHZmYDOl59obZ9dmFd#2XWo~D5Xdp2#Ff}qDK0XR_baG{3Z3=jtjkE<+
zoXZj|3?WExcNyH>g1dWg*8v8G8C-*Fa1R6z?t$Qv;1=8+f@`n@mq&8Wz4zSv|7*Rs
zX02~}SM92<>fY6~l<JzS5>`+Ppez*P%F4;cAqbFCdBe#8;Nak9<KW;zrlr*axq^Xz
zn~`aCfX*%;C`9ld1Sw~rx$Cn{+T8V-rvill6y3l8PHq4vpCBi{AO{D4i-SYpUqq<0
zAVAvO9b^SiVFM^aAwU;oS}CZbmov!5*7bRte=PwFmW%*S0ReuNKj8og2cR>^(i{R%
zF?Y2EIy~=aX$}TxLM=f+SFis{!60nw>gp)S&hFvi!DjB@!UlD=5oKfncz|4O0dIgV
zKxcQL72tQn09A7b;9s4wA=3i1Y(Xx6*)^fot{&#jK)|yB46+15T%IG`AXY$U!1Lw+
zO$B9unj;YM7g_l)0t?{pvjK3jasE5q-_gGlfgpc^%`GjV4vyv!FA&5AU=0ET0cx_!
zY_6WJEC6$e)o-FX*ai9=Z|-gm0-IYr6aG|g4v>}50GK~F_;-6Qmd+qYR~I%H5cqeC
z?7!1IpRx?ZN($=W0ED=@ApfpU8srSLd_HzB_P=h{9s>1%`1}p5K@cnJ-%VJ#IkIa*
zKu&Hz1?j&j&nD#m*ld8V0A3Cb4t@b{0MH2l^t7~P|D9gT%Mtj;$@$y-+<~u;Bh(RK
z{oDl57i0~5{zLY0F?R<7T%Fy3zCQnS{5L}8<OEoOEL{N>KpPMQ`9Im8%|PqF@bl$6
zgFFF-9M9V01aSQR{d;BnEH5i41nl)6`k(t{S5bbWpeW7!SI7VMNlHRJ0Y0oe8~|1>
zUJd}~Z;|moFMR(yjk-DL?=t=os{pZv0tEi5_W79pRk8cu0$})iI2Zx{ol6z^tXd#|
z;Xfib;^5`5eEx9$KTG{j$p4?he`oo>iTr;Hl63=v|F{|c`2Rm{a|aOE>u<udXx&_&
z2S5e-JO+^erRoCz8d?>g70Auuf4vH>=Ffv50kHx9`xrqkvLH{Ol{(1P()KU4{AJhv
zJ!N1J1gH*m0sVef0Ibhf^FO}l(XzCEemY#9CHcn%e4d>D-ckl)3AOq?VqCm@0CQ(&
zb1&rQOMXVY03XihS+oLr{t+>NoectYeGUOUx91D6hB_nvzEOTI0K5Ee(;xJ^0(Pr^
zKu!QV@E`CsfF1lF_^e}g=)WP4z%zpUgU!hUV0ZopeD-_%f&bNjrJJ+!^9=uy*mHOP
z!haqNAkY(NiM%iiwG;}ms|{(p{v<)-!MZ&vGDW+so5slMv*6tBc8iXf!dRUdwB~#%
zkuunYxx6XE@JoD=;=$*zxehVBB}t?8?(My4!kdZhR^<6{{Jyb*Ly6D5WUokBwZwPc
zJ~+MA3ABf6hW)BYo9pEE8eRP(&b3F6oag7>vZdj$*=>!TYCa{D`?5*a80~1Iz~X7z
zJd3OuLMjATRx%_e>{CyS>0g&q*g0cQ6p9JV$i7E0+=V^{>s+xnGu|r|T3jw2L^MPO
zgk*5Pu!kq;d?fc$6$$2iN(z_z<~^H*3n_BYSl%zQVzTXVWmJRSL_r$6YutwyzHn~k
zrT{|dN&0ZMl4?pDaZR16)bOfXO`~vJYq;%!WP4w%C0E)mbvmu>-jgm0On!a>Orz=B
z9tJdZ`rOqytqczJv0jC}Z@C_99NO&k)h!n56(hCr;AE1qHH=(d63ifb4`1dw@}ff*
zsDO*fvgn@j`X1gqRM;ozO!AhPbEANR!{9Z1H|#L!58m9jXR*GO^{1*|)~B)Nz#iDf
z!B4?_s!tYQ-=Cfq5ZQ9em}Q1B<YmZozRO#Uqt_6s1=ddom$9F&SK_ltmwHhsCvNc0
zpY_ue@+io4Z_iZ?Mj8cq+1&`_v;EMi)NYOE&DWsZI*MW)*{cs6kPHs(5xBJ4+qf(*
zmo&lKhw9v%R(eE84#s}h;t$Dt_vKAMPKS$Hx|~GRi=wN|neJB>)W=$q?vxIVaTvN<
zBDb3(9YyyZEM*GPwDg4wD!oek!r^ZhjOV2byNNyph~vNe)fm84+pc!a0E3!|xe`m8
zxw8aEZQ4HAsB`v6EJ>oNb^VdrdmW~GZgOOk^4r8FjSslesF<gx2_VE9oel?1prf4I
z4ES)dCc+RM`Dh*4w5shPQMX81lQG=~Tss;Vekn<f*v`#?D>`~44=5<5&s<e5WX64(
zP0F_0@XivyrHE@1sd6!@UB&$+99k}7)S(YcvnDMTp?&abeCM7oC1=u8E>8o}UrzEN
zl4uK5O*@;MO*uD#tjq;NP-B?=0qf^xp+6&Kv~{Rn5R&B{RumG<rtl0s?m%r^>e(9{
zb#9qF$`Dq;8Z3IO^~Z4ubVZD6Hu!~}S}Da*7)94I7`kpBeuQ|w8f5A~o$>DRj>DN*
z`5BU=izxQeA#~X}8xJa3-sCusGtCH^x7AO9Pz~YH-O=|bC=+IW4Z>OEoE)WZk<y=p
zTlXDK>Kbq6Ggk*~Pgiq8)++Y{km~C{eaa9UBr+4b;7>ioM=>4Fg*Qz3a`D<zU6_Y+
zhU`Mbb?jOmW)TBTh%ddl$=6xQjdP#-?t``z3_R!qex1B)W$nb2@1++Z`Rt>gB{{9p
zn$t;uQA^7Xj~b6VrD2BYl|swMPbZ(N@A3L}ETf1y*2v1wRTOVja`ZVWhU46Q_|^!M
zXN|7A0WX|$oCZHoaTAGL!6j#DP3kMLEh`Zo3^I1I5XcQF`ItvY8A!@yw<D92c`Y%Z
zM=H^Wb#tk4OzFI@;!{wkp#k6R-&QoSCK0R~b`Ty;14jm<uKS6Rh?Ius=(uQYk}aye
z!><&FiPBlzpnfV9Ts?%|t8R(ux#81%OdndAiqLwy(8?#C1ip@ztcmsHD<$JwG7^3p
zyZUZ)YLnwyNhemi89UbV^A3Wbb?DKaw9`hDYo%%=)s#^%x~O(JJ0@PJ%jHS2_u7e4
z*=n+ClLUI`x5cT$9z7r_B3|?J_L!0*u~R<}>U~3Vd^3L4A3)nv`E_Ct&*S^*cP?XS
zS@#;C%Je}@!S)-5+g_fe!w*>m0>OGtbfo!IE2Po^FTSgwj>S~~nJijK2%$2BSM%J1
zu9Wp~>Qi;LocDE!zyYDi0Da;X!|+3!!o98PZQZQTmKrtyJ&hAi+1+x+0S_KjU(`W2
zK&5hY;Ls!9s5=u@Nmx+nuxvrgv5*GObZhqE5V0pOX)TG}`d4=rxSdvAhAU5yTt|XY
zQ&O3IN4O0C2QN;URJKI^+Eo@<sLJPC2MPmq%e)>uL>3{D3MmRE4JPEZ&zpH@B35UQ
z76HSSsqX0u3yDX6rMFNO6jPkpsrR~AYmcM}?Py+GDtk|S0p+<hChb?|uGthw;%c0S
zdF)v*vS$>@4cuPj)ajd+7%D1r<8r+Q4=aXn4i!%=-eb|S{s^zI!DTn)R9%XqRlt|?
zI+Ag^lBB-7I*jozmo*2LKtwEgCxNTQCvULvoA3+UTV-SV;mX?8v)K09m0wN-cwWCO
zA~ZD|)zgW<BIv*PsgYZarltP$%RM;v1{pZSlIZxjNF40!Z<}=6l;nng79SFA#til$
zM!EuOT+POtidz%#aTCk0shhh`coS)l;F*5P8Nux{DCZ<+D%<Lh!LE)|VeYqZ3<wkQ
zSt%QS6<DdkjlZX(BQ&4aHWD#~j)q;O#-^6N4q|D0{knXXJou<PEYv1P#7JaaoJw~6
zR!R0+E6?ICE#Zq!U@VvcY*W>lwP-O}8M8v!v{yEhI`hrw`@@Ynj|L~fr(r)pxxj{#
zPGM8=a3LVk#l-&jql;B+1_a$Zp%V$QGJm%~aF$d#+ndNZ7VEUk86~ptV_C4fs`QsA
z6*+zZ8U<z0BVCfjybZ~B2$;_cHeB&loUB^?xv<i!?L+mE0<~BP3Bn=wuq6N0N$^Ly
z-TG_&?YTPog;5yY#qsT6x{i+=fl7c!BD@dmpjV)UNw-%P*~DE^$glf*yd~rdXL7h<
zk-gZ?4Rbx4H&$FZCEO~@vQQl*#@Gyb@-^qVEIg8#Edj2j1wT}Q#Dun_(!;eYD7{qa
zl;nO}<wDH-jz$F@o<!LC8*<1dE4(sQBg=b9c-W*lrpwK#a`ji=Pxk{D$R$w0jW`oO
zIL(${yRlukG5^?Hpt`*f6*^_{_a3a9phbqZR<GrH!VPewxddU&4iI7Ma>{y+{g~c+
zzXq^*bV!J)_tf$FRlkTekws-p?d?bSzSz<Iyb~>pEG!}Ya=y2O%~?|59St6&Nnff`
ziu1O|waiMK`l{qL%(%U3_;i%gC)nI6q5O`6=_GKV>@5QV%6zoyMPw1I(E9g6tH+A2
zyJOhIO3eE;rnX4ozM{O8&0d@koH96<jv@i!dbv0>DH*d+o&Art>+OR*w0Igxqq2oi
z_4b~P39Mj$nzHzIstIoHwkvgkIe{aC7#D$XSbDR~N^-O#=SAHcLDl5KqXCzxU;W9A
z6ygZh!Kj%gWV$P%bX+_h*ri}^lpUIBF5MHHT91{`z9TgOi>Lf;Dz%~<oS_ljZYEi5
z;}!NIGq18}jha^xgULo3DP1J8Ol8X;E&Q{XJrh;HrIl&Eidk?$m?o{(XVn%Z{<vx+
z<~QlHM~2RY8TSMFzj|?w^__seZ8h_xsD)8i^ZMcHw?C0bWnr~tJ|0~>MxRb81$WF~
z$Vb!H){DX#Xo!zuC@ieI(&leHrDk>*mN4@7(0#ejwH<|Iq1F)X>(fs!nRhj4b)V7e
zIr~qqrL#j!At`W*%6a%GNDQgs$hEt<-^M>6SJtbN`6$WOuPvRL*k@R*c0msO@Q=-u
ztMiSKmdp#Z^I)5MeeNT$-8W8@*Ops$KJp~}<V>US$8mO$c>3Xzg4@-1aSkz9gJVi>
z6)C+Ma`|-%&A_z~_et=YMM0vJlA?(6A+SVZ)AcuKI65KU;UIvZtXnvS+PJ8vS<$dg
z7+qG6PG`_53p;*&iQR+&#zB4C?3YpZdheEPuqU;F*?P3!+pEMqnBu!>t)zF=P<T$*
zPcAX{={6*U7Y7$@zvjZb^<d5ODPI%K6gVKKE`<}?cq~vNA|qY%Z_B>Y7x&UM_B#vB
zd68au1BOb@?pGweqDakhr{uWVYmAgl4nkd7yZuorvu17>5oCR>?#UA%ADWBz_5e}$
zzD{UKj(ipuQy+(lTk<FG!o~g0^>=wAs;P2Aglo>!(ez_WUe#-XGDlG^e#KVFhl_d_
zcDNsr4K@rvOFi*4SkPTX4XZX@k0i;t^fm<wQoNxNFy_Ohi@0_0s$vAy6GgfY?%Q=;
z>Xt+)_q>rbA$ucf4*q)R`PnpgPCrYqmRF^?`sLcrk0sobNJO02)=G?e691ArcDsVm
z_1=X%O5atp<R6F9bh@Mg17%yY8hXt%2=Z&iyBNc+3I&yIulxsQ*cS9j&&1~;c;(Ki
zn5I9EAj^SWNN8Ww?l)Z}RN+~b(IBZ&@5SIvSMg0uLzirD6YKr5@%!S0V$8@tjUSgN
zNrtbU@&S7qI+rP`6Pl1kllQ=1H5?Pdh|53vPce3cHDuk?7oM}n2q9_`(W9C){}SAC
zeW_SuCw_qIS^ha#hJizFNNBVLZ%M=HMQmw*vb`34^JzAwP_}Tq6A5H~%Pr6qtSrTu
zq}K!B9U*>??36Iq<1|?qD7F<f?#(Hs$=pGjug=-+M-Ci(p~*FZ!d<SIs<Jo~k^*Kp
zo2sWlB5wf>YHYJRII8k0@x$sq03+Wrb}!AZnn%i~R<0W~PgV1{#c2C}?d{p^{#k;6
zs|}c-{FbVpa&!Q5LpcqrfHbYYT{4wJq(r-CStwK8bVPoaAtJdjtYAzMpozeF1)#o|
zYQeHaA2!qoQ?@PBXWb6+({z6HGUo-TsO!s>&WsC@3%raWHGWw-L7)$V9~COj8<4oN
zHo-uQe4UvfijvhG+&CyLZx{X%bJGRS8866lITd+_txI%qe73qeZ_NIE)grh@h%U97
zT6lNy!>4m#Jfro*maKP%Sz~{r*vCx;OBiuYdAd~LC-}C+A1{IK5SZ6EfmyOaq(+-M
zHD%nr7v_}`lwMwKn}TOk%(&T>l=+^g@Tox-V=9lc@4mBJ`*N?PW*mMi>ffZ2D^10!
z+@-fxahX7zrqI*Xe(=4&tZ!`6@2QV5K?#_iqqQ69r<DJ45~U{pgfuu3OJRD`5GAx1
zUORg3pRl3W<erCK8)&>tjZCwEScb7gFJ&th$eS)4qDlBb)_VNyQ|Fm@QYB&5s}JTC
zbpfon=@R~ykzxB_s4i@s3d4CEf-_8GG4+z!+&Dzv$8O1t`z`o|J6V-k3%vq3_rZU?
zhH|^^KJV9*de_jAPx40und-C?kx@tUJrk6`dD}xZdpNVK#vd7*j@P%Sm9O2rmT8>N
zQ&%oh+>ot(EuFfVc{{!QC?8ggXR>Zdt~-v<wOJ}*zqsZy?tld4&vU=1h4w5lRG9+R
z0{6*eb-p}0jU(WTR=Dl0_pr(dMtnc(@ea-o&5$=RqK(r&cAxKnAx9hc-ZUXx|0WsZ
zMH`^1wb9hTEkanOcepOr=76i1c#Y3G#T{}*2HKpH7hjGKv>pq7uy1<9c9`+&!JR&!
z-+>LAenfM#OS*M<0qk{8LCBwk{}EOH*R0g%%3g>4dXt7~IX7NYxeUbX`UTFSI}BI$
z=xLP4^_by}Q{;7?cQu@qkO<Q|X;EKC1m+#<G8+L~8nWAxURY!C6O|jYd6UUzIi*V>
z-g1NjkN(5fms<+XPA3v2+}097Yx-Cqw~BzSm96gc_KB0DNo;**%zZNHUzdV$F^g8@
zh;ItvRscy7u2`sttYz^mxQrVu=4AzgEw&qD0*2GYZcj*>yMzKti9{h-Jz7eo=6W<+
z9yWtX4Qj!L8!tSgNSykE^SxdwFr+@U*<tc37ndn?Sl0W#wI%PMN${55z|kJ-iuY*M
zk6~_qs4x&5J=yIUE9lMS{CZvw_Fn*Yko3cvxm}CDAb$I@;)nuob~$3N?n9r+A|J^Q
z_H3_-J2UgsJnW(Yv`EA1(MU|Q$uGk)+4)@_#GUX*=Hw)(dEPsqb7k`o0&HKqp-;Pm
z;&=Ix+rGX2o)|kPx()D%>D!QSm)JB82ot&wdNo+gpw#;!DrE>eO0;SB^)ZDbtMUZh
zK>i)viISNpz^tx>w#zK_<ByfEgDz!aE$-i1`%H}P-!UcP%|2)heJ@e9unuKjYWuPp
zk`1ML6|+#R{i@A-;=<;@X2g6eHr}mHHg_UjxEn|B)@!AytYZAdEmPZz0h`Mw9x**G
z<F0s~68R4!A2a68KbUlmJ1GO9wV7$!#}|I6SA~g*<NTsWjC}<c>GF{H>=!L1l@k0^
z`##cPZ-m*oTpv_=`+wfJY!?OG*$XJ>{Y<pZFU_2Q+jts#xMRjyDAJQTDo({5U_1Cu
z*)iTt#iY2T$&cQ#s83#vM7jrmVz-1wMmM=2v8Ra2T)q?WwYe1udl6gYSE_2>HVvP9
zttPRSKGr+xCBiD@_fI^d#Xok%GGNB&FOROeZ~SEmtaRj*XIYqc@m=sL<Q@ywP1u`C
zo$9(@4qn3AM)zV=+}s$DS9pg=m%56ha$f<lKV;<WsmzOu+A#|2+L~>tv_*(M&3znv
z^7{^4Kh}{}oO%1cr9}_JccNIzIdXuYPisJ8U=PKb@Zm*Teqyq^uMm7B)l{3(<rMX3
zf>v%Ly6n}h(Jt}BmC1Ax92qejictEij<*MI0MbT1L1^-B(|k`Fy$|DQ^jR9*0~SJ_
z)J?Gz+!qv@!-8dY7Ffqiv`wyx35e{XLwRaXKQhUjbyfqrh&+uyH6yf%GH#sYbk-XO
zfU|Dkc%fe~4R`a#WrZ+VROw3SaX_V!Mm4G$oeJmj(<48}DLTZo1dqk%WlsopRi980
zHT61IIBJ9|7+Y=Eb!QO*+j*>;Wu+T!<CfUwKWn{UG|_~gU?mTQc=$fBphOs`Hp$?P
zOTqrEzCzJQiFpGLtoQDkzC7h_n>0dhn0GyXeANHWYZgZ!>t~wb*%=cIr=9&aie4|V
zzCS!fP8>sf7L!i_#fTz3WOw;e^j>gK$Mo*Spy*7I0FwfTFaA`$5tYw89?}W7_M&&+
z*`V8&Y^F8b@Ipc8FZH#OAx4?_h}RfF_u7S-w;qgDr#3i|-|(EK1=p?Ys0*8`m+>NO
z5uN%`3NzbWplGbk;)%#gaB}={n^cu=HO9kV?3CSZ0}Rf+UEYQU`R4a1&(vw+L>3AL
zSeMN>v=w(raG%7@o@%0S^@tE9%o={SYMf8HSIjh8{)|svNqUZNd9+9vaDRjG7U7%g
z7dy_0z6^<7=f_CAs*Z`ON1|ax_bwJCinp+@&g(+QQgl}1<1>-d?lUw(>I|1xg|$2N
z1W8{8EiQ>!GGEfJ*StDbZC#es<P`|J4(|L68&j4|)jivj`PpVqYrkz}%fzBow-*Uv
ziJ_K>1`zFmP;tG-h!+{1v>9rGt4iry*{$zgC|npuN8tQkNWw}%io}dPBG4C^)eZp5
z=WX;2e2y{v5SgjOL(DW1sKiP%x1w99xvd^GQYBF)M<3jS-?My6@OgZ4SmuOleq3Ii
zzmuSv{8)cWcaXJKwqwT9^HmEoQ^8IG!jW-vf(j{&#YoYK(L`{+)Np!{?>3RAn5SMq
z8#s_m!Y(+QF$Uq}N9^I0ciD<(*wYZ<SxD%omAZ!mPaEQyhwVtluV1%$E@$J=Kuv~u
z5kv1(2>e^m7JaF)C^Fa83mga*aZ63Wlk@x%4x*tWXjutS*rC3ux9cT{jEn9`Uxf;H
zkw-#QxouWe8ga0{H5sf>&J>+_nXU&*-l-K{b0^l-t4ftub=3|F@W0+1-6KF3)hCFj
z7Oz5hQm-nnK<+Do?;}?LaAr31#1ns<N1lC*>YFEW41{Y<6@#KInv*0)fBB~4*yHtb
z^@E!I8&vY?LsFaF@8AQOz8!1Anh)6w^!qC9N_*EwaxdHAZxhr$-^i?&d>(HRX<)TQ
z-d`i7n%F?!>tie=Oj+49H_Rb7=h0N;G`i%XZhI<o$bDIgp}8BSQOBq%GMeCU=~u~}
z6ks2A8OcnO9t)`Q`xdkcgW650gepE~zO-o;Zhz&NL1GmWMrYNJ804Dt^s&@_CnQ)~
zZ{e2nQ|2a6^e4L*_=~|yW6V=_r_@Fw9!sZGX&NWDJ`Tg-2W&fXpnyX}pj7lnYHnyY
zO|Ny3a^!yI7<z6SE@JJAixJ{vJmC~=J@_+0S#Izo0mVJ9KCpcT1s5+x%B9?TSQ2GT
z$dcDLxVs;1^)YR&P~cm!SVix+gybrtpJZO5hGLWi0a!7dXpzM&e-!h+D|=}lU?d`v
zzff|*y_yy`<>-b&#>|i4)XVatLsKNA`aG;0CJZJ}ZUcfN3QLc;waaAR^QjppbY!qy
z+b<GjuC#ceA+}&SSaZ;zirmvJ5kI6ABmWRMaf|Y)e$_E0K>n1!7mWMRwv+HO1k@(~
z&dc8-r>MdH-qy}(3$}aiB>ytro<~bjs}K_5=Omv%a?=ynUTUXKmR3OM^J}!I(ELq;
z%G(#FgXGOWgN_R0L__X2RpVxluTZf0Z|5=v+ECL7viW@j?uRaUs=Uc48{Py)mT$=Q
z&RCN)0P1)<uLlqs&TGNIFU`#Fe;8@=W;xdi-4it<M8R1LV%yEDg|rSn_&qenl0Zmp
znUgT&zalemkjM&}yb?M(zQePT`fx=!)%b}9f6e%)yl5|-uW9$Fz81Y@u;H%5y&K++
z>Uv(aB7OHI)eA)@tc9XL2)^;1N+14pPDmRkOu$G;NQoZ3%30w@u3Ak&`gBTlB+>#x
z55>Bq*sWHXflg+<7v&iwW2VhK+QrI#u!6nn{XrZzt29A2QPQMmyV8DWUVQ9zb5pD9
z>;~`J_jz%Rs_~KD#JY+L>KC3p`i_My9Npi=$d9X~DKI*NGIVS@CMOYodg?vZlm-=!
zh66Wj$*r2^72nROtLlx?qlvAH?vQ$@2z)=!pK>jy{VY5#0@y=7B+&hyu;^vF<}`s$
zW{lNtz)Q%xkcojSX)Zvbx{5NIL2FF-S%0<YuKRJ1wq|J*Nz`~s9kc0eQq_a+6k~}4
zM(Qj2i!tVeQ5Q^0G}vyswdxW03^bHNwnFt4R<y4zR`eMIZmqtuiU&1XYh4|2(U@sp
zorvr`j1(KEZNBG143xH(<t+zM>lqcygc0dKI-GDTYs>Lh`sycB+r0nD5eMgQ<-93U
z62?D3SThESf<HJV(Ufd*V{TiNyWczyNG&z|FiO+t(?<MAgl`&sM<WRLCRqPw2&MnQ
z&x1}&VyP)J=c?k=CX=5-__`%M=K@M4Cme~xC1#w#){VZ>$`D&(vO{6c%@au{8eJ6G
zld$!Y<yhECYfZ*rPpgrvs=TJd_M00|WW+EZ#VI6BIjkDJfN`-fTz#7(UBdlFGKh?s
zQ$FdunbUtv(S+j7sK9kVQ7zeK^pgT8j4QP|%?B0NN-G186RFw>e0Pd8A?T`HX=i1U
zYp9mtM^G)<O2yplA8Zvz?lo<=c9W&(!Q@|dU=t;>xWw8KJ=i@Tcfd@kSei{Baaq~(
z_PsWS*n40}95Zg8Be~JHYK%>CJ?!X1Khw1mF|F{VUe|n)w<999EJ#PI<S1sPPBF7Z
zqtQ}QDJ7?$3X#JizY{-|wUqL9tcR_O*vTD#K^s_lhlaI6WB7u!&aNSNGau5hG^FuW
zQ%ueJfc7gBd`Fn=ed~*duuqIZU)vUt4b5(PA73z7Dif2KStzt^t1wSqiz-~Fvg00$
z1BKsxn3@FheefSMro};E_t)^{nU?*PlE^$45EAyIC84~1lD)0MVbp@{>xJHqNRAoI
z8ATVbe|}{i3s&<?W8so#JIY-2`{eBMP*8jB{Gua)+O$Ky{XvKCx%A6*^^u0|Ne}gX
zpK+ShOZHD5fgooRF#_D01{$iUaH*<-Y`WFi{V<ZU=@%IH+2lfU#CHkHsx}OHLp}u+
zO}00~QlyEWB#VTM6XlkBvu)PWUtlih)`ujEYoYt@<>_k44PUg{nepdhoW&mWeiNV|
zVM8oEmoh(4j&P}w2nyX`>ItH$Ao!IF!?dwPMw6(s2P=R+%`$_*GWa%)jH0W7v`()0
zo?bC6EpxwJl%TPugVlWvmt6%%G_y>QcRA~5vaW~qFs^_WUQKc0xEm^Ui(y<SufIQk
zQ2sDLYV<D8Rdy9N)&-dm-eex}<a7v(pHFQFu~W#{U9bKoB(W{d0oN7%i=%{;md25H
ziQ%f%5VW_2Z4fK3$?7zczdH^~zg~8m`7l+ir?f)#N(C6@wQ1zsL~0CFNn|jvnqr6@
zX*E&7oIa<TlA&dPgKC}mkvel?2+crOb+;*DS`?R<u&n$-v|+N&sP!REnniCu*a3#h
zuZr|F?{poUE2hR);`wpTL&TW@$*aY%JRw5T^;L?7uFlfh(>><sh_5fSIDh!}hgg}=
zYDxCBpFNt?YL0;ge%Yv6N2X+VhS7{3#sv;j>!`Ntp_4Jj%9qY1^pBqOL=|8JXbO=N
z?~ySTca{m<>$WPjc5<%HUWs+8D+pLJ1gnPg4#k%v>Q9G)Hxm<4H!YnEf_UttdrP9C
zLHJ}c&`>5O*>NFtxD0}ot&Tu?18c#26i3T9wbRH2dYu_&lFk6p<Fty%8ZpGv7{A9*
zw+-$7w2KFRvMky2V`3Io)Tqzt*m+zt*Q3LJxb1h|P6K#2X@W(N>^#HOW)dvHQfw3d
z@l+pkW3x?bt-k#469`eni`Zeqx$6>=>ZQvvFt^N=XBG|8VdYy44HcXm=x>jTpN{(}
zI=vERXNbyRR8m}u<w*5-?itWnaz0rsR5QmfdrK+bzFaQtfVipO?Y9*V3ci#vsE@e?
zGw?B_7k(L^et9n!US#XZs>2>>+|0Plc~Ti7d0P$pDU|@F(=9?;dwRW6G^ROT8B54Z
zjItQZ?Pah07Arb?uI;#%?>n1-7&6b?RU{dJbT!YydVbcX(5R`Hnohg`Al00WUOSB5
zPvr(AtCze8!IY$?Q=BaEM-2`y`fzca)0O&x-cwwooa;d>5bP8Nd!h>M)*u3x*YE_c
z_5?1P-9+0rM7u@i@L8Co92ECxYu?GkU>m2q``qz-8VqYGf>Y57kh79zD%?=#-AA$I
zJ)q>1`BsMc^j*zKX)`38vNhL$4f`Z^?%g0Q8sE<eP;#d+HL+uEM3loyz0M@75Vq^<
zBu1!8)escc>vR!b`(4C}VEH>}VGnbe)iCPCFl*5XVk)B6mqc9GDCICRPIcjJ=&#p#
z=$qX!PUe@llF3C+r<ZbH(ys!Q_T|`Z^>^B_r|yls`MU-bVho^9#u0=2pJ=Sh@sqbZ
z4@AOcxZ!Psr}RJ9nN%;YEVP(By3DGTs1S?CXFLFolcp?-EDs7T_n`t~uE}rMh>MAG
z1k~=p5Af9o_ALQKg_=X(yUIg<)_mptg6@xvY1fHthk|kD$)B4~N1EkG(M0(sfbEyd
zAc>rp<aowtH!^+G+!w;_E;o#zLVH1-l%*}<`Jeh8aMao<at(9~S3P9)3FY$)F+T+!
zXG<?m7WvsPnW}j{o)3`l=ENJ2Jkje#ZllK%9w6q4x-|4Gc3&HDZ5~n%Kz&y*Ek>Fn
zz_!4GY$xwFvg)578~e2Doz>4F2ZuVmJ!0QcGVlW`b>k;FKeVZ84vASnqCXwY7hnXE
zbfB7cA0oF>815~|5gV0R#k*CA>>t#^?=IVG9vYw(<EZa&WiK?WAy7+~1#~AUbLQX3
z9w0ZtEsz{k6g7o*)O2zfi>ih;iQ?YnR9>uV48+@>4#5Z3yvMzm0E7xVv60zdPwKjW
zBmo`_5nb10sgWGCb2RY@2Sj${hu>>T1qkd9ptYJQk~#Fv8q6ArQQ)z#T?59oM4(~g
zpouYZ^^c^nxfN@Dg!CyX9HK$B8S)}~s(>uDL;9%ljjVYqDP`kV4+?yR?=-6H*eOgD
zuruhk)>AbZ*9|M0xgSI_)AXGOTSq6gV^rRSQ#lBlT9bCR)R1mekS(wPG4yw#j!SyY
z)V~apQ}*-UO;GU@c7N|;;V95^nGgC-M__sv52;i{0;pZ=@GgD9qG2LQ&;RP46(g6{
zoBX;=ceOljc!wK=DNr|d_6|8=kZ=XHxhcO_chU-LDJ@g^2Zm;J+?U%A=C8agkV$f+
zAOxCJ?qpn@_KGStxA}OfKZS5e`spq~J~<4yIRplMbO;Ws;ourK(<@n|mPDJ`r*b+2
f<|2Ni{{sZ9i9!lxZe(+Ga%Ev{3T19&Z(?c+QF+1F

literal 0
HcmV?d00001

diff --git a/services/clsi/test/unit/lib/pdfjsTests.js b/services/clsi/test/unit/lib/pdfjsTests.js
new file mode 100644
index 0000000000..64f062c0b8
--- /dev/null
+++ b/services/clsi/test/unit/lib/pdfjsTests.js
@@ -0,0 +1,78 @@
+const fs = require('fs')
+const Path = require('path')
+const { expect } = require('chai')
+const { parseXrefTable } = require('../../../app/lib/pdfjs/parseXrefTable')
+const PATH_EXAMPLES = 'test/acceptance/fixtures/examples/'
+const PATH_SNAPSHOTS = 'test/unit/lib/snapshots/'
+const EXAMPLES = fs.readdirSync(PATH_EXAMPLES)
+
+function snapshotPath(example) {
+  return Path.join(PATH_SNAPSHOTS, example, 'XrefTable.json')
+}
+
+function pdfPath(example) {
+  return Path.join(PATH_EXAMPLES, example, 'output.pdf')
+}
+
+async function loadContext(example) {
+  const size = (await fs.promises.stat(pdfPath(example))).size
+
+  let blob
+  try {
+    blob = await fs.promises.readFile(snapshotPath(example))
+  } catch (e) {
+    if (e.code !== 'ENOENT') {
+      throw e
+    }
+  }
+  const snapshot = blob ? JSON.parse(blob) : null
+  return {
+    size,
+    snapshot
+  }
+}
+
+async function backFillSnapshot(example, size) {
+  const table = await parseXrefTable(pdfPath(example), size)
+  await fs.promises.mkdir(Path.dirname(snapshotPath(example)), {
+    recursive: true
+  })
+  await fs.promises.writeFile(
+    snapshotPath(example),
+    JSON.stringify(table, null, 2)
+  )
+  return table
+}
+
+describe('pdfjs', function () {
+  describe('when the pdf is an empty file', function () {
+    it('should yield no entries', async function () {
+      const path = 'does/not/matter.pdf'
+      const table = await parseXrefTable(path, 0)
+      expect(table).to.deep.equal([])
+    })
+  })
+
+  for (const example of EXAMPLES) {
+    describe(example, function () {
+      let size, snapshot
+      before('load snapshot', async function () {
+        const ctx = await loadContext(example)
+        size = ctx.size
+        snapshot = ctx.snapshot
+      })
+
+      before('back fill new snapshot', async function () {
+        if (snapshot === null) {
+          console.error('back filling snapshot for', example)
+          snapshot = await backFillSnapshot(example, size)
+        }
+      })
+
+      it('should produce the expected xRef table', async function () {
+        const table = await parseXrefTable(pdfPath(example), size)
+        expect(table).to.deep.equal(snapshot)
+      })
+    })
+  }
+})
diff --git a/services/clsi/test/unit/lib/snapshots/asymptote/XrefTable.json b/services/clsi/test/unit/lib/snapshots/asymptote/XrefTable.json
new file mode 100644
index 0000000000..be9f29164d
--- /dev/null
+++ b/services/clsi/test/unit/lib/snapshots/asymptote/XrefTable.json
@@ -0,0 +1,356 @@
+[
+  {
+    "offset": 0,
+    "gen": 0,
+    "free": true
+  },
+  {
+    "offset": 123086,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 123405,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 15,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 216,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 1084,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 1244,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 4001,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 4155,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 4297,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 4932,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 5307,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 5495,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 30246,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 31466,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 38398,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 39039,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 40158,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 40897,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 65550,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 74691,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 81693,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 97169,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 104103,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 111180,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 118555,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 6,
+    "gen": 0
+  },
+  {
+    "offset": 6,
+    "gen": 1
+  },
+  {
+    "offset": 6,
+    "gen": 2
+  },
+  {
+    "offset": 6,
+    "gen": 3
+  },
+  {
+    "offset": 6,
+    "gen": 4
+  },
+  {
+    "offset": 6,
+    "gen": 5
+  },
+  {
+    "offset": 6,
+    "gen": 6
+  },
+  {
+    "offset": 6,
+    "gen": 7
+  },
+  {
+    "offset": 6,
+    "gen": 8
+  },
+  {
+    "offset": 6,
+    "gen": 9
+  },
+  {
+    "offset": 6,
+    "gen": 10
+  },
+  {
+    "offset": 6,
+    "gen": 11
+  },
+  {
+    "offset": 6,
+    "gen": 12
+  },
+  {
+    "offset": 6,
+    "gen": 13
+  },
+  {
+    "offset": 6,
+    "gen": 14
+  },
+  {
+    "offset": 6,
+    "gen": 15
+  },
+  {
+    "offset": 6,
+    "gen": 16
+  },
+  {
+    "offset": 6,
+    "gen": 17
+  },
+  {
+    "offset": 6,
+    "gen": 18
+  },
+  {
+    "offset": 6,
+    "gen": 19
+  },
+  {
+    "offset": 6,
+    "gen": 20
+  },
+  {
+    "offset": 6,
+    "gen": 21
+  },
+  {
+    "offset": 6,
+    "gen": 22
+  },
+  {
+    "offset": 6,
+    "gen": 23
+  },
+  {
+    "offset": 6,
+    "gen": 24
+  },
+  {
+    "offset": 6,
+    "gen": 25
+  },
+  {
+    "offset": 6,
+    "gen": 26
+  },
+  {
+    "offset": 6,
+    "gen": 27
+  },
+  {
+    "offset": 6,
+    "gen": 28
+  },
+  {
+    "offset": 6,
+    "gen": 29
+  },
+  {
+    "offset": 6,
+    "gen": 30
+  },
+  {
+    "offset": 6,
+    "gen": 31
+  },
+  {
+    "offset": 6,
+    "gen": 32
+  },
+  {
+    "offset": 6,
+    "gen": 33
+  },
+  {
+    "offset": 6,
+    "gen": 34
+  },
+  {
+    "offset": 6,
+    "gen": 35
+  },
+  {
+    "offset": 6,
+    "gen": 36
+  },
+  {
+    "offset": 6,
+    "gen": 37
+  },
+  {
+    "offset": 6,
+    "gen": 38
+  },
+  {
+    "offset": 6,
+    "gen": 39
+  },
+  {
+    "offset": 6,
+    "gen": 40
+  },
+  {
+    "offset": 6,
+    "gen": 41
+  },
+  {
+    "offset": 6,
+    "gen": 42
+  },
+  {
+    "offset": 6,
+    "gen": 43
+  },
+  {
+    "offset": 6,
+    "gen": 44
+  },
+  {
+    "offset": 6,
+    "gen": 45
+  },
+  {
+    "offset": 6,
+    "gen": 46
+  },
+  {
+    "offset": 6,
+    "gen": 47
+  },
+  {
+    "offset": 6,
+    "gen": 48
+  },
+  {
+    "offset": 6,
+    "gen": 49
+  },
+  {
+    "offset": 6,
+    "gen": 50
+  },
+  {
+    "offset": 6,
+    "gen": 51
+  },
+  {
+    "offset": 6,
+    "gen": 52
+  },
+  {
+    "offset": 6,
+    "gen": 53
+  },
+  {
+    "offset": 6,
+    "gen": 54
+  },
+  {
+    "offset": 6,
+    "gen": 55
+  }
+]
\ No newline at end of file
diff --git a/services/clsi/test/unit/lib/snapshots/biber_bibliography/XrefTable.json b/services/clsi/test/unit/lib/snapshots/biber_bibliography/XrefTable.json
new file mode 100644
index 0000000000..94c0fddf42
--- /dev/null
+++ b/services/clsi/test/unit/lib/snapshots/biber_bibliography/XrefTable.json
@@ -0,0 +1,128 @@
+[
+  {
+    "offset": 0,
+    "gen": 255,
+    "free": true
+  },
+  {
+    "offset": 9,
+    "gen": 1
+  },
+  {
+    "offset": 9,
+    "gen": 0
+  },
+  {
+    "offset": 15,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 9,
+    "gen": 14
+  },
+  {
+    "offset": 9,
+    "gen": 12
+  },
+  {
+    "offset": 9,
+    "gen": 15
+  },
+  {
+    "offset": 9,
+    "gen": 13
+  },
+  {
+    "offset": 9,
+    "gen": 16
+  },
+  {
+    "offset": 57274,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 9,
+    "gen": 17
+  },
+  {
+    "offset": 9,
+    "gen": 2
+  },
+  {
+    "offset": 9,
+    "gen": 3
+  },
+  {
+    "offset": 9,
+    "gen": 4
+  },
+  {
+    "offset": 9,
+    "gen": 5
+  },
+  {
+    "offset": 9,
+    "gen": 6
+  },
+  {
+    "offset": 522,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 9,
+    "gen": 7
+  },
+  {
+    "offset": 8788,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 9,
+    "gen": 8
+  },
+  {
+    "offset": 17289,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 9,
+    "gen": 9
+  },
+  {
+    "offset": 32619,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 9,
+    "gen": 10
+  },
+  {
+    "offset": 44596,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 9,
+    "gen": 11
+  },
+  {
+    "offset": 9,
+    "gen": 18
+  },
+  {
+    "offset": 57027,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 58655,
+    "gen": 0,
+    "uncompressed": true
+  }
+]
\ No newline at end of file
diff --git a/services/clsi/test/unit/lib/snapshots/epstopdf/XrefTable.json b/services/clsi/test/unit/lib/snapshots/epstopdf/XrefTable.json
new file mode 100644
index 0000000000..fca6e5447d
--- /dev/null
+++ b/services/clsi/test/unit/lib/snapshots/epstopdf/XrefTable.json
@@ -0,0 +1,125 @@
+[
+  {
+    "offset": 0,
+    "gen": 255,
+    "free": true
+  },
+  {
+    "offset": 208,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 6,
+    "gen": 11
+  },
+  {
+    "offset": 6,
+    "gen": 0
+  },
+  {
+    "offset": 15,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 6,
+    "gen": 14
+  },
+  {
+    "offset": 29580,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 6,
+    "gen": 15
+  },
+  {
+    "offset": 6,
+    "gen": 1
+  },
+  {
+    "offset": 6,
+    "gen": 2
+  },
+  {
+    "offset": 6,
+    "gen": 3
+  },
+  {
+    "offset": 6,
+    "gen": 4
+  },
+  {
+    "offset": 6,
+    "gen": 5
+  },
+  {
+    "offset": 6,
+    "gen": 6
+  },
+  {
+    "offset": 6,
+    "gen": 7
+  },
+  {
+    "offset": 6,
+    "gen": 8
+  },
+  {
+    "offset": 18310,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 6,
+    "gen": 9
+  },
+  {
+    "offset": 6,
+    "gen": 10
+  },
+  {
+    "offset": 18554,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 18945,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 19674,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 6,
+    "gen": 12
+  },
+  {
+    "offset": 22318,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 6,
+    "gen": 13
+  },
+  {
+    "offset": 6,
+    "gen": 16
+  },
+  {
+    "offset": 29321,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 30697,
+    "gen": 0,
+    "uncompressed": true
+  }
+]
\ No newline at end of file
diff --git a/services/clsi/test/unit/lib/snapshots/feynmf/XrefTable.json b/services/clsi/test/unit/lib/snapshots/feynmf/XrefTable.json
new file mode 100644
index 0000000000..0da87fc092
--- /dev/null
+++ b/services/clsi/test/unit/lib/snapshots/feynmf/XrefTable.json
@@ -0,0 +1,115 @@
+[
+  {
+    "offset": 0,
+    "gen": 255,
+    "free": true
+  },
+  {
+    "offset": 8,
+    "gen": 1
+  },
+  {
+    "offset": 8,
+    "gen": 0
+  },
+  {
+    "offset": 15,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 8,
+    "gen": 5
+  },
+  {
+    "offset": 8,
+    "gen": 12
+  },
+  {
+    "offset": 8,
+    "gen": 14
+  },
+  {
+    "offset": 8,
+    "gen": 13
+  },
+  {
+    "offset": 25628,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 8,
+    "gen": 15
+  },
+  {
+    "offset": 8,
+    "gen": 2
+  },
+  {
+    "offset": 8,
+    "gen": 3
+  },
+  {
+    "offset": 8,
+    "gen": 4
+  },
+  {
+    "offset": 250,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 8,
+    "gen": 6
+  },
+  {
+    "offset": 8,
+    "gen": 7
+  },
+  {
+    "offset": 8,
+    "gen": 8
+  },
+  {
+    "offset": 3854,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 8,
+    "gen": 9
+  },
+  {
+    "offset": 11227,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 8,
+    "gen": 10
+  },
+  {
+    "offset": 18230,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 8,
+    "gen": 11
+  },
+  {
+    "offset": 8,
+    "gen": 16
+  },
+  {
+    "offset": 25381,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 26423,
+    "gen": 0,
+    "uncompressed": true
+  }
+]
\ No newline at end of file
diff --git a/services/clsi/test/unit/lib/snapshots/feynmp/XrefTable.json b/services/clsi/test/unit/lib/snapshots/feynmp/XrefTable.json
new file mode 100644
index 0000000000..34048b037d
--- /dev/null
+++ b/services/clsi/test/unit/lib/snapshots/feynmp/XrefTable.json
@@ -0,0 +1,102 @@
+[
+  {
+    "offset": 0,
+    "gen": 65535,
+    "free": true
+  },
+  {
+    "offset": 1094,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 5512,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 1035,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 875,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 15,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 856,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 1159,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 1593,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 3149,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 1436,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 2412,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 1282,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 1768,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 1200,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 1230,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 2006,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 2625,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 3381,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 4069,
+    "gen": 0,
+    "uncompressed": true
+  }
+]
\ No newline at end of file
diff --git a/services/clsi/test/unit/lib/snapshots/fontawesome/XrefTable.json b/services/clsi/test/unit/lib/snapshots/fontawesome/XrefTable.json
new file mode 100644
index 0000000000..1951e17f7f
--- /dev/null
+++ b/services/clsi/test/unit/lib/snapshots/fontawesome/XrefTable.json
@@ -0,0 +1,93 @@
+[
+  {
+    "offset": 0,
+    "gen": 255,
+    "free": true
+  },
+  {
+    "offset": 7,
+    "gen": 1
+  },
+  {
+    "offset": 7,
+    "gen": 0
+  },
+  {
+    "offset": 15,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 7,
+    "gen": 11
+  },
+  {
+    "offset": 7,
+    "gen": 10
+  },
+  {
+    "offset": 7,
+    "gen": 9
+  },
+  {
+    "offset": 29476,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 7,
+    "gen": 12
+  },
+  {
+    "offset": 7,
+    "gen": 7
+  },
+  {
+    "offset": 7,
+    "gen": 2
+  },
+  {
+    "offset": 7,
+    "gen": 8
+  },
+  {
+    "offset": 7,
+    "gen": 3
+  },
+  {
+    "offset": 7,
+    "gen": 4
+  },
+  {
+    "offset": 255,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 7,
+    "gen": 5
+  },
+  {
+    "offset": 17910,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 7,
+    "gen": 6
+  },
+  {
+    "offset": 7,
+    "gen": 13
+  },
+  {
+    "offset": 29228,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 30448,
+    "gen": 0,
+    "uncompressed": true
+  }
+]
\ No newline at end of file
diff --git a/services/clsi/test/unit/lib/snapshots/fontawesome_xelatex/XrefTable.json b/services/clsi/test/unit/lib/snapshots/fontawesome_xelatex/XrefTable.json
new file mode 100644
index 0000000000..e6315a28e7
--- /dev/null
+++ b/services/clsi/test/unit/lib/snapshots/fontawesome_xelatex/XrefTable.json
@@ -0,0 +1,126 @@
+[
+  {
+    "offset": 0,
+    "gen": 0,
+    "free": true
+  },
+  {
+    "offset": 6338,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 15,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 216,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 707,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 757,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 888,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 991,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 1257,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 1678,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 2050,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 4246,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 4339,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 5382,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 5475,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 5513,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 15,
+    "gen": 0
+  },
+  {
+    "offset": 15,
+    "gen": 1
+  },
+  {
+    "offset": 15,
+    "gen": 2
+  },
+  {
+    "offset": 15,
+    "gen": 3
+  },
+  {
+    "offset": 15,
+    "gen": 4
+  },
+  {
+    "offset": 15,
+    "gen": 5
+  },
+  {
+    "offset": 15,
+    "gen": 6
+  },
+  {
+    "offset": 15,
+    "gen": 7
+  },
+  {
+    "offset": 15,
+    "gen": 8
+  },
+  {
+    "offset": 15,
+    "gen": 9
+  },
+  {
+    "offset": 15,
+    "gen": 10
+  }
+]
\ No newline at end of file
diff --git a/services/clsi/test/unit/lib/snapshots/glossaries/XrefTable.json b/services/clsi/test/unit/lib/snapshots/glossaries/XrefTable.json
new file mode 100644
index 0000000000..51f7dca8bb
--- /dev/null
+++ b/services/clsi/test/unit/lib/snapshots/glossaries/XrefTable.json
@@ -0,0 +1,94 @@
+[
+  {
+    "offset": 0,
+    "gen": 255,
+    "free": true
+  },
+  {
+    "offset": 7,
+    "gen": 1
+  },
+  {
+    "offset": 7,
+    "gen": 0
+  },
+  {
+    "offset": 15,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 7,
+    "gen": 10
+  },
+  {
+    "offset": 7,
+    "gen": 9
+  },
+  {
+    "offset": 7,
+    "gen": 8
+  },
+  {
+    "offset": 33085,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 7,
+    "gen": 11
+  },
+  {
+    "offset": 7,
+    "gen": 2
+  },
+  {
+    "offset": 7,
+    "gen": 3
+  },
+  {
+    "offset": 7,
+    "gen": 4
+  },
+  {
+    "offset": 445,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 7,
+    "gen": 5
+  },
+  {
+    "offset": 10048,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 7,
+    "gen": 6
+  },
+  {
+    "offset": 18151,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 7,
+    "gen": 7
+  },
+  {
+    "offset": 7,
+    "gen": 12
+  },
+  {
+    "offset": 32838,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 34157,
+    "gen": 0,
+    "uncompressed": true
+  }
+]
\ No newline at end of file
diff --git a/services/clsi/test/unit/lib/snapshots/gnuplot/XrefTable.json b/services/clsi/test/unit/lib/snapshots/gnuplot/XrefTable.json
new file mode 100644
index 0000000000..5148059666
--- /dev/null
+++ b/services/clsi/test/unit/lib/snapshots/gnuplot/XrefTable.json
@@ -0,0 +1,89 @@
+[
+  {
+    "offset": 0,
+    "gen": 255,
+    "free": true
+  },
+  {
+    "offset": 4,
+    "gen": 0
+  },
+  {
+    "offset": 4,
+    "gen": 1
+  },
+  {
+    "offset": 4,
+    "gen": 2
+  },
+  {
+    "offset": 22047,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 4,
+    "gen": 4
+  },
+  {
+    "offset": 4,
+    "gen": 3
+  },
+  {
+    "offset": 15,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 4,
+    "gen": 10
+  },
+  {
+    "offset": 4,
+    "gen": 9
+  },
+  {
+    "offset": 4,
+    "gen": 11
+  },
+  {
+    "offset": 4,
+    "gen": 5
+  },
+  {
+    "offset": 4,
+    "gen": 6
+  },
+  {
+    "offset": 6451,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 4,
+    "gen": 7
+  },
+  {
+    "offset": 14825,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 4,
+    "gen": 8
+  },
+  {
+    "offset": 4,
+    "gen": 12
+  },
+  {
+    "offset": 21800,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 22696,
+    "gen": 0,
+    "uncompressed": true
+  }
+]
\ No newline at end of file
diff --git a/services/clsi/test/unit/lib/snapshots/hebrew/XrefTable.json b/services/clsi/test/unit/lib/snapshots/hebrew/XrefTable.json
new file mode 100644
index 0000000000..d3e8221a2e
--- /dev/null
+++ b/services/clsi/test/unit/lib/snapshots/hebrew/XrefTable.json
@@ -0,0 +1,81 @@
+[
+  {
+    "offset": 0,
+    "gen": 255,
+    "free": true
+  },
+  {
+    "offset": 6,
+    "gen": 1
+  },
+  {
+    "offset": 6,
+    "gen": 0
+  },
+  {
+    "offset": 15,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 6,
+    "gen": 8
+  },
+  {
+    "offset": 6,
+    "gen": 7
+  },
+  {
+    "offset": 22744,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 6,
+    "gen": 9
+  },
+  {
+    "offset": 6,
+    "gen": 2
+  },
+  {
+    "offset": 6,
+    "gen": 6
+  },
+  {
+    "offset": 6,
+    "gen": 3
+  },
+  {
+    "offset": 364,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 6,
+    "gen": 4
+  },
+  {
+    "offset": 12163,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 6,
+    "gen": 5
+  },
+  {
+    "offset": 6,
+    "gen": 10
+  },
+  {
+    "offset": 22496,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 23856,
+    "gen": 0,
+    "uncompressed": true
+  }
+]
\ No newline at end of file
diff --git a/services/clsi/test/unit/lib/snapshots/knitr/XrefTable.json b/services/clsi/test/unit/lib/snapshots/knitr/XrefTable.json
new file mode 100644
index 0000000000..d57e124754
--- /dev/null
+++ b/services/clsi/test/unit/lib/snapshots/knitr/XrefTable.json
@@ -0,0 +1,145 @@
+[
+  {
+    "offset": 0,
+    "gen": 0,
+    "free": true
+  },
+  {
+    "offset": 43543,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 43792,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 15,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 216,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 734,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 784,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 912,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 1020,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 1544,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 5791,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 12911,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 23655,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 30651,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 42597,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 14,
+    "gen": 0
+  },
+  {
+    "offset": 14,
+    "gen": 1
+  },
+  {
+    "offset": 14,
+    "gen": 2
+  },
+  {
+    "offset": 14,
+    "gen": 3
+  },
+  {
+    "offset": 14,
+    "gen": 4
+  },
+  {
+    "offset": 14,
+    "gen": 5
+  },
+  {
+    "offset": 14,
+    "gen": 6
+  },
+  {
+    "offset": 14,
+    "gen": 7
+  },
+  {
+    "offset": 14,
+    "gen": 8
+  },
+  {
+    "offset": 14,
+    "gen": 9
+  },
+  {
+    "offset": 14,
+    "gen": 10
+  },
+  {
+    "offset": 14,
+    "gen": 11
+  },
+  {
+    "offset": 14,
+    "gen": 12
+  },
+  {
+    "offset": 14,
+    "gen": 13
+  },
+  {
+    "offset": 14,
+    "gen": 14
+  },
+  {
+    "offset": 14,
+    "gen": 15
+  },
+  {
+    "offset": 14,
+    "gen": 16
+  }
+]
\ No newline at end of file
diff --git a/services/clsi/test/unit/lib/snapshots/knitr_utf8/XrefTable.json b/services/clsi/test/unit/lib/snapshots/knitr_utf8/XrefTable.json
new file mode 100644
index 0000000000..14982555fc
--- /dev/null
+++ b/services/clsi/test/unit/lib/snapshots/knitr_utf8/XrefTable.json
@@ -0,0 +1,179 @@
+[
+  {
+    "offset": 0,
+    "gen": 0,
+    "free": true
+  },
+  {
+    "offset": 75291,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 75540,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 15,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 216,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 790,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 840,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 975,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 1083,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 2127,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 13797,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 23679,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 31863,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 36111,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 50346,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 61562,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 73508,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 16,
+    "gen": 0
+  },
+  {
+    "offset": 16,
+    "gen": 1
+  },
+  {
+    "offset": 16,
+    "gen": 2
+  },
+  {
+    "offset": 16,
+    "gen": 3
+  },
+  {
+    "offset": 16,
+    "gen": 4
+  },
+  {
+    "offset": 16,
+    "gen": 5
+  },
+  {
+    "offset": 16,
+    "gen": 6
+  },
+  {
+    "offset": 16,
+    "gen": 7
+  },
+  {
+    "offset": 16,
+    "gen": 8
+  },
+  {
+    "offset": 16,
+    "gen": 9
+  },
+  {
+    "offset": 16,
+    "gen": 10
+  },
+  {
+    "offset": 16,
+    "gen": 11
+  },
+  {
+    "offset": 16,
+    "gen": 12
+  },
+  {
+    "offset": 16,
+    "gen": 13
+  },
+  {
+    "offset": 16,
+    "gen": 14
+  },
+  {
+    "offset": 16,
+    "gen": 15
+  },
+  {
+    "offset": 16,
+    "gen": 16
+  },
+  {
+    "offset": 16,
+    "gen": 17
+  },
+  {
+    "offset": 16,
+    "gen": 18
+  },
+  {
+    "offset": 16,
+    "gen": 19
+  },
+  {
+    "offset": 16,
+    "gen": 20
+  },
+  {
+    "offset": 16,
+    "gen": 21
+  },
+  {
+    "offset": 16,
+    "gen": 22
+  }
+]
\ No newline at end of file
diff --git a/services/clsi/test/unit/lib/snapshots/latex_compiler/XrefTable.json b/services/clsi/test/unit/lib/snapshots/latex_compiler/XrefTable.json
new file mode 100644
index 0000000000..91e596e54f
--- /dev/null
+++ b/services/clsi/test/unit/lib/snapshots/latex_compiler/XrefTable.json
@@ -0,0 +1,132 @@
+[
+  {
+    "offset": 0,
+    "gen": 65535,
+    "free": true
+  },
+  {
+    "offset": 17349,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 25448,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 17290,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 17130,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 15,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 17109,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 17414,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 17548,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 18912,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 18381,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 22406,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 17897,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 21796,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 18758,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 23361,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 17455,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 17485,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 19218,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 22021,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 22638,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 23599,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 18051,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 18142,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 18657,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 24005,
+    "gen": 0,
+    "uncompressed": true
+  }
+]
\ No newline at end of file
diff --git a/services/clsi/test/unit/lib/snapshots/lualatex_compiler/XrefTable.json b/services/clsi/test/unit/lib/snapshots/lualatex_compiler/XrefTable.json
new file mode 100644
index 0000000000..a403bb4f6a
--- /dev/null
+++ b/services/clsi/test/unit/lib/snapshots/lualatex_compiler/XrefTable.json
@@ -0,0 +1,74 @@
+[
+  {
+    "offset": 0,
+    "gen": 255,
+    "free": true
+  },
+  {
+    "offset": 6,
+    "gen": 1
+  },
+  {
+    "offset": 6,
+    "gen": 0
+  },
+  {
+    "offset": 15,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 6,
+    "gen": 4
+  },
+  {
+    "offset": 6,
+    "gen": 6
+  },
+  {
+    "offset": 2320,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 6,
+    "gen": 2
+  },
+  {
+    "offset": 202,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 298,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 6,
+    "gen": 3
+  },
+  {
+    "offset": 1642,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 6,
+    "gen": 5
+  },
+  {
+    "offset": 6,
+    "gen": 7
+  },
+  {
+    "offset": 2120,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 2892,
+    "gen": 0,
+    "uncompressed": true
+  }
+]
\ No newline at end of file
diff --git a/services/clsi/test/unit/lib/snapshots/makeindex-custom-style/XrefTable.json b/services/clsi/test/unit/lib/snapshots/makeindex-custom-style/XrefTable.json
new file mode 100644
index 0000000000..b2b9dcef35
--- /dev/null
+++ b/services/clsi/test/unit/lib/snapshots/makeindex-custom-style/XrefTable.json
@@ -0,0 +1,107 @@
+[
+  {
+    "offset": 0,
+    "gen": 255,
+    "free": true
+  },
+  {
+    "offset": 5,
+    "gen": 1
+  },
+  {
+    "offset": 5,
+    "gen": 0
+  },
+  {
+    "offset": 15,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 5,
+    "gen": 12
+  },
+  {
+    "offset": 30470,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 5,
+    "gen": 13
+  },
+  {
+    "offset": 5,
+    "gen": 3
+  },
+  {
+    "offset": 5,
+    "gen": 2
+  },
+  {
+    "offset": 366,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 5,
+    "gen": 11
+  },
+  {
+    "offset": 5,
+    "gen": 10
+  },
+  {
+    "offset": 5,
+    "gen": 4
+  },
+  {
+    "offset": 5,
+    "gen": 5
+  },
+  {
+    "offset": 5,
+    "gen": 6
+  },
+  {
+    "offset": 638,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 5,
+    "gen": 7
+  },
+  {
+    "offset": 7838,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 5,
+    "gen": 8
+  },
+  {
+    "offset": 15674,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 5,
+    "gen": 9
+  },
+  {
+    "offset": 5,
+    "gen": 14
+  },
+  {
+    "offset": 30223,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 31457,
+    "gen": 0,
+    "uncompressed": true
+  }
+]
\ No newline at end of file
diff --git a/services/clsi/test/unit/lib/snapshots/makeindex/XrefTable.json b/services/clsi/test/unit/lib/snapshots/makeindex/XrefTable.json
new file mode 100644
index 0000000000..9a58bed4cb
--- /dev/null
+++ b/services/clsi/test/unit/lib/snapshots/makeindex/XrefTable.json
@@ -0,0 +1,90 @@
+[
+  {
+    "offset": 0,
+    "gen": 255,
+    "free": true
+  },
+  {
+    "offset": 5,
+    "gen": 1
+  },
+  {
+    "offset": 5,
+    "gen": 0
+  },
+  {
+    "offset": 15,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 5,
+    "gen": 9
+  },
+  {
+    "offset": 23352,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 5,
+    "gen": 10
+  },
+  {
+    "offset": 5,
+    "gen": 3
+  },
+  {
+    "offset": 5,
+    "gen": 2
+  },
+  {
+    "offset": 366,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 5,
+    "gen": 8
+  },
+  {
+    "offset": 5,
+    "gen": 4
+  },
+  {
+    "offset": 5,
+    "gen": 5
+  },
+  {
+    "offset": 589,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 5,
+    "gen": 6
+  },
+  {
+    "offset": 8425,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 5,
+    "gen": 7
+  },
+  {
+    "offset": 5,
+    "gen": 11
+  },
+  {
+    "offset": 23105,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 24245,
+    "gen": 0,
+    "uncompressed": true
+  }
+]
\ No newline at end of file
diff --git a/services/clsi/test/unit/lib/snapshots/minted/XrefTable.json b/services/clsi/test/unit/lib/snapshots/minted/XrefTable.json
new file mode 100644
index 0000000000..f0d2f2a195
--- /dev/null
+++ b/services/clsi/test/unit/lib/snapshots/minted/XrefTable.json
@@ -0,0 +1,77 @@
+[
+  {
+    "offset": 0,
+    "gen": 255,
+    "free": true
+  },
+  {
+    "offset": 6,
+    "gen": 1
+  },
+  {
+    "offset": 6,
+    "gen": 0
+  },
+  {
+    "offset": 15,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 6,
+    "gen": 7
+  },
+  {
+    "offset": 6,
+    "gen": 6
+  },
+  {
+    "offset": 19462,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 6,
+    "gen": 8
+  },
+  {
+    "offset": 6,
+    "gen": 2
+  },
+  {
+    "offset": 6,
+    "gen": 3
+  },
+  {
+    "offset": 340,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 6,
+    "gen": 4
+  },
+  {
+    "offset": 7343,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 6,
+    "gen": 5
+  },
+  {
+    "offset": 6,
+    "gen": 9
+  },
+  {
+    "offset": 19215,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 20089,
+    "gen": 0,
+    "uncompressed": true
+  }
+]
\ No newline at end of file
diff --git a/services/clsi/test/unit/lib/snapshots/multibib_bibliography/XrefTable.json b/services/clsi/test/unit/lib/snapshots/multibib_bibliography/XrefTable.json
new file mode 100644
index 0000000000..ffd44df50b
--- /dev/null
+++ b/services/clsi/test/unit/lib/snapshots/multibib_bibliography/XrefTable.json
@@ -0,0 +1,133 @@
+[
+  {
+    "offset": 0,
+    "gen": 255,
+    "free": true
+  },
+  {
+    "offset": 6,
+    "gen": 1
+  },
+  {
+    "offset": 6,
+    "gen": 0
+  },
+  {
+    "offset": 15,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 6,
+    "gen": 14
+  },
+  {
+    "offset": 6,
+    "gen": 15
+  },
+  {
+    "offset": 40591,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 6,
+    "gen": 17
+  },
+  {
+    "offset": 6,
+    "gen": 3
+  },
+  {
+    "offset": 6,
+    "gen": 2
+  },
+  {
+    "offset": 290,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 6,
+    "gen": 16
+  },
+  {
+    "offset": 6,
+    "gen": 5
+  },
+  {
+    "offset": 6,
+    "gen": 4
+  },
+  {
+    "offset": 595,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 6,
+    "gen": 7
+  },
+  {
+    "offset": 6,
+    "gen": 6
+  },
+  {
+    "offset": 844,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 6,
+    "gen": 8
+  },
+  {
+    "offset": 6,
+    "gen": 9
+  },
+  {
+    "offset": 6,
+    "gen": 10
+  },
+  {
+    "offset": 1107,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 6,
+    "gen": 11
+  },
+  {
+    "offset": 11816,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 6,
+    "gen": 12
+  },
+  {
+    "offset": 28180,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 6,
+    "gen": 13
+  },
+  {
+    "offset": 6,
+    "gen": 18
+  },
+  {
+    "offset": 40344,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 41851,
+    "gen": 0,
+    "uncompressed": true
+  }
+]
\ No newline at end of file
diff --git a/services/clsi/test/unit/lib/snapshots/nomenclature/XrefTable.json b/services/clsi/test/unit/lib/snapshots/nomenclature/XrefTable.json
new file mode 100644
index 0000000000..51286af2bc
--- /dev/null
+++ b/services/clsi/test/unit/lib/snapshots/nomenclature/XrefTable.json
@@ -0,0 +1,94 @@
+[
+  {
+    "offset": 0,
+    "gen": 255,
+    "free": true
+  },
+  {
+    "offset": 7,
+    "gen": 1
+  },
+  {
+    "offset": 7,
+    "gen": 0
+  },
+  {
+    "offset": 15,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 7,
+    "gen": 8
+  },
+  {
+    "offset": 7,
+    "gen": 9
+  },
+  {
+    "offset": 7,
+    "gen": 10
+  },
+  {
+    "offset": 32363,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 7,
+    "gen": 11
+  },
+  {
+    "offset": 7,
+    "gen": 2
+  },
+  {
+    "offset": 7,
+    "gen": 3
+  },
+  {
+    "offset": 7,
+    "gen": 4
+  },
+  {
+    "offset": 565,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 7,
+    "gen": 5
+  },
+  {
+    "offset": 10031,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 7,
+    "gen": 6
+  },
+  {
+    "offset": 18203,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 7,
+    "gen": 7
+  },
+  {
+    "offset": 7,
+    "gen": 12
+  },
+  {
+    "offset": 32116,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 33492,
+    "gen": 0,
+    "uncompressed": true
+  }
+]
\ No newline at end of file
diff --git a/services/clsi/test/unit/lib/snapshots/references_in_include/XrefTable.json b/services/clsi/test/unit/lib/snapshots/references_in_include/XrefTable.json
new file mode 100644
index 0000000000..b0c6731240
--- /dev/null
+++ b/services/clsi/test/unit/lib/snapshots/references_in_include/XrefTable.json
@@ -0,0 +1,90 @@
+[
+  {
+    "offset": 0,
+    "gen": 255,
+    "free": true
+  },
+  {
+    "offset": 5,
+    "gen": 1
+  },
+  {
+    "offset": 5,
+    "gen": 0
+  },
+  {
+    "offset": 15,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 5,
+    "gen": 9
+  },
+  {
+    "offset": 15855,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 5,
+    "gen": 10
+  },
+  {
+    "offset": 5,
+    "gen": 3
+  },
+  {
+    "offset": 5,
+    "gen": 2
+  },
+  {
+    "offset": 167,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 5,
+    "gen": 8
+  },
+  {
+    "offset": 5,
+    "gen": 4
+  },
+  {
+    "offset": 5,
+    "gen": 5
+  },
+  {
+    "offset": 341,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 5,
+    "gen": 6
+  },
+  {
+    "offset": 7911,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 5,
+    "gen": 7
+  },
+  {
+    "offset": 5,
+    "gen": 11
+  },
+  {
+    "offset": 15608,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 16597,
+    "gen": 0,
+    "uncompressed": true
+  }
+]
\ No newline at end of file
diff --git a/services/clsi/test/unit/lib/snapshots/simple_bibliography/XrefTable.json b/services/clsi/test/unit/lib/snapshots/simple_bibliography/XrefTable.json
new file mode 100644
index 0000000000..d633b7ba6f
--- /dev/null
+++ b/services/clsi/test/unit/lib/snapshots/simple_bibliography/XrefTable.json
@@ -0,0 +1,94 @@
+[
+  {
+    "offset": 0,
+    "gen": 255,
+    "free": true
+  },
+  {
+    "offset": 7,
+    "gen": 1
+  },
+  {
+    "offset": 7,
+    "gen": 0
+  },
+  {
+    "offset": 15,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 7,
+    "gen": 9
+  },
+  {
+    "offset": 7,
+    "gen": 8
+  },
+  {
+    "offset": 7,
+    "gen": 10
+  },
+  {
+    "offset": 35573,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 7,
+    "gen": 11
+  },
+  {
+    "offset": 7,
+    "gen": 2
+  },
+  {
+    "offset": 7,
+    "gen": 3
+  },
+  {
+    "offset": 7,
+    "gen": 4
+  },
+  {
+    "offset": 373,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 7,
+    "gen": 5
+  },
+  {
+    "offset": 8639,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 7,
+    "gen": 6
+  },
+  {
+    "offset": 23349,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 7,
+    "gen": 7
+  },
+  {
+    "offset": 7,
+    "gen": 12
+  },
+  {
+    "offset": 35326,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 36668,
+    "gen": 0,
+    "uncompressed": true
+  }
+]
\ No newline at end of file
diff --git a/services/clsi/test/unit/lib/snapshots/subdirectories/XrefTable.json b/services/clsi/test/unit/lib/snapshots/subdirectories/XrefTable.json
new file mode 100644
index 0000000000..06ed0fdf7d
--- /dev/null
+++ b/services/clsi/test/unit/lib/snapshots/subdirectories/XrefTable.json
@@ -0,0 +1,108 @@
+[
+  {
+    "offset": 0,
+    "gen": 255,
+    "free": true
+  },
+  {
+    "offset": 548,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 9,
+    "gen": 1
+  },
+  {
+    "offset": 9,
+    "gen": 2
+  },
+  {
+    "offset": 9,
+    "gen": 0
+  },
+  {
+    "offset": 15,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 9,
+    "gen": 10
+  },
+  {
+    "offset": 9,
+    "gen": 9
+  },
+  {
+    "offset": 9,
+    "gen": 11
+  },
+  {
+    "offset": 46398,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 9,
+    "gen": 12
+  },
+  {
+    "offset": 8100,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 9,
+    "gen": 3
+  },
+  {
+    "offset": 9,
+    "gen": 4
+  },
+  {
+    "offset": 9,
+    "gen": 5
+  },
+  {
+    "offset": 9693,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 9,
+    "gen": 6
+  },
+  {
+    "offset": 17959,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 9,
+    "gen": 7
+  },
+  {
+    "offset": 34174,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 9,
+    "gen": 8
+  },
+  {
+    "offset": 9,
+    "gen": 13
+  },
+  {
+    "offset": 46151,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 47562,
+    "gen": 0,
+    "uncompressed": true
+  }
+]
\ No newline at end of file
diff --git a/services/clsi/test/unit/lib/snapshots/tikz_feynman/XrefTable.json b/services/clsi/test/unit/lib/snapshots/tikz_feynman/XrefTable.json
new file mode 100644
index 0000000000..afeaa84536
--- /dev/null
+++ b/services/clsi/test/unit/lib/snapshots/tikz_feynman/XrefTable.json
@@ -0,0 +1,145 @@
+[
+  {
+    "offset": 0,
+    "gen": 255,
+    "free": true
+  },
+  {
+    "offset": 8,
+    "gen": 8
+  },
+  {
+    "offset": 8,
+    "gen": 9
+  },
+  {
+    "offset": 8,
+    "gen": 10
+  },
+  {
+    "offset": 8,
+    "gen": 1
+  },
+  {
+    "offset": 8,
+    "gen": 0
+  },
+  {
+    "offset": 15,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 8,
+    "gen": 20
+  },
+  {
+    "offset": 33577,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 8,
+    "gen": 3
+  },
+  {
+    "offset": 8,
+    "gen": 2
+  },
+  {
+    "offset": 1151,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 8,
+    "gen": 17
+  },
+  {
+    "offset": 8,
+    "gen": 19
+  },
+  {
+    "offset": 8,
+    "gen": 18
+  },
+  {
+    "offset": 8,
+    "gen": 5
+  },
+  {
+    "offset": 8,
+    "gen": 4
+  },
+  {
+    "offset": 2721,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 8,
+    "gen": 7
+  },
+  {
+    "offset": 8,
+    "gen": 6
+  },
+  {
+    "offset": 5757,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 8,
+    "gen": 11
+  },
+  {
+    "offset": 8,
+    "gen": 12
+  },
+  {
+    "offset": 8,
+    "gen": 13
+  },
+  {
+    "offset": 9558,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 8,
+    "gen": 14
+  },
+  {
+    "offset": 18967,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 8,
+    "gen": 15
+  },
+  {
+    "offset": 26388,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 8,
+    "gen": 16
+  },
+  {
+    "offset": 8,
+    "gen": 21
+  },
+  {
+    "offset": 33354,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 34451,
+    "gen": 0,
+    "uncompressed": true
+  }
+]
\ No newline at end of file
diff --git a/services/clsi/test/unit/lib/snapshots/xelatex_compiler/XrefTable.json b/services/clsi/test/unit/lib/snapshots/xelatex_compiler/XrefTable.json
new file mode 100644
index 0000000000..31080439b5
--- /dev/null
+++ b/services/clsi/test/unit/lib/snapshots/xelatex_compiler/XrefTable.json
@@ -0,0 +1,73 @@
+[
+  {
+    "offset": 0,
+    "gen": 65535,
+    "free": true
+  },
+  {
+    "offset": 8,
+    "gen": 4
+  },
+  {
+    "offset": 8,
+    "gen": 3
+  },
+  {
+    "offset": 8,
+    "gen": 1
+  },
+  {
+    "offset": 8,
+    "gen": 6
+  },
+  {
+    "offset": 8,
+    "gen": 8
+  },
+  {
+    "offset": 15,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 8,
+    "gen": 0
+  },
+  {
+    "offset": 6857,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 8,
+    "gen": 2
+  },
+  {
+    "offset": 265,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 8,
+    "gen": 7
+  },
+  {
+    "offset": 8,
+    "gen": 5
+  },
+  {
+    "offset": 701,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 6750,
+    "gen": 0,
+    "uncompressed": true
+  },
+  {
+    "offset": 7655,
+    "gen": 0,
+    "uncompressed": true
+  }
+]
\ No newline at end of file