Corrupted pdf (inconsistent)

Hi @rkusa,

I’m having some intermittent problems that appear related to closing the pdf document. Here is a reduced snippet demonstrating te issue:

export const testPDF = async function () {

  const readFile = function (f) {
    return new Promise(function (resolve, reject) {
      fs.readFile(f, (err, data) => {
        err ? reject(err) : resolve(data);
      });
    });
  }

  const tempFile = `${Random.id()}.pdf`
  const tempPath = `${os.tmpdir()}${tempFile}`;

  const doc = new pdf.Document();
  doc.pipe(fs.createWriteStream(tempPath))

  const sampleCell = doc.cell();
  sampleCell.text({fontSize: 14 }).add('headline')
  sampleCell.text({fontSize: 12 }).add('Line one')
  sampleCell.text().add('Line two')
  
  await doc.end();

  const pdfData1 = await readFile(tempPath);
  // const pdfData2 = await readFile(tempPath);

  console.log(`pdfData1 ${tempFile} length: ${pdfData1.length}`);
  // console.log(`pdfData2 ${tempFile} length: ${pdfData2.length}`);

  return true;

};

invoking the above function like

    const res = await testPDF();

sometimes (maybe half the time) results in pdfData1 having corrupted (truncated?) data. example of invoking it six times:

    const res1 = await testPDF();
    const res2 = await testPDF();
    const res3 = await testPDF();
    const res4 = await testPDF();
    const res5 = await testPDF();
    const res6 = await testPDF();

results in following console log:

I20190104-09:29:53.937(11)? pdfData1 6EXKAv5E2nACnDrff.pdf length: 5258
I20190104-09:29:53.941(11)? pdfData1 jfzRdCmNpC7XDkjNT.pdf length: 16
I20190104-09:29:53.944(11)? pdfData1 fyEZg7XyEg9CanpoR.pdf length: 5258
I20190104-09:29:53.946(11)? pdfData1 m2atPJS4v7B7g9iBi.pdf length: 16
I20190104-09:29:53.949(11)? pdfData1 eBoKr5zHuXyfCTc2L.pdf length: 16
I20190104-09:29:53.953(11)? pdfData1 w2pwKQAJEQwtLHvGu.pdf length: 16

Wen reading the same file again (as when adding the commented out lines) will always result in pdfData2 having correct data. Sample output with commented out lines added:

I20190104-09:29:17.490(11)? pdfData1 JAzWC7KhfprnQqAs7.pdf length: 16
I20190104-09:29:17.491(11)? pdfData2 JAzWC7KhfprnQqAs7.pdf length: 5258
I20190104-09:29:17.496(11)? pdfData1 YT9Rt97BNwW7PuhuH.pdf length: 16
I20190104-09:29:17.497(11)? pdfData2 YT9Rt97BNwW7PuhuH.pdf length: 5258
I20190104-09:29:17.501(11)? pdfData1 Hrf9H5tjEt2cTv8qo.pdf length: 5258
I20190104-09:29:17.502(11)? pdfData2 Hrf9H5tjEt2cTv8qo.pdf length: 5258
I20190104-09:29:17.505(11)? pdfData1 cx9axS5guzmh8WeMn.pdf length: 16
I20190104-09:29:17.505(11)? pdfData2 cx9axS5guzmh8WeMn.pdf length: 5258
I20190104-09:29:17.508(11)? pdfData1 fi3FeGKdnNhCi5YJv.pdf length: 5258
I20190104-09:29:17.509(11)? pdfData2 fi3FeGKdnNhCi5YJv.pdf length: 5258
I20190104-09:29:17.511(11)? pdfData1 x5jsesFbKnLMSCJHC.pdf length: 5258
I20190104-09:29:17.512(11)? pdfData2 x5jsesFbKnLMSCJHC.pdf length: 5258

A second question: Is there a way of getting the generated data into a variable (eg, pdfData1 above) without having to write the file into disk first?

Thanks for looking into this.

[hluz]

Works great. Thanks!

[rkusa]

Updated snippet with the commit above:

const doc = new pdf.Document();

// start reading here
const buff = doc.asBuffer({end: false})

// MERGE documents here

doc.end();
const pdfData = await buff;