Points of Interests

The following is a list of common steps executed by pdftotext to extract text from a PDF file, and in which file the corresponding code is located. Note that the stated locations refer to commit 065dca3 and may have changed by now.

Opening and reading the PDF file
PDFDoc::PDFDoc(), line 144ff
Parsing the PDF version number from the PDF file header
PDFDoc::checkHeader(), line 350
Parsing startxref
PDFDoc::getStartXRef(), line 1999ff
Parsing the xref table and the trailer dictionary
XRef::readXRefTable(), line 535
Parsing the document catalog
Catalog::Catalog(), line 76
Extracting the PDF subtype
PDFDoc::extractPDFSubtype(), line 516
Parsing a page (i.e., a page object node from the page tree)
Catalog::cachePageTree(), line 315
Parsing the commands of a page's content stream
Gfx::display(), line 644
Executing the commands of a page's content stream
Gfx::go(), line 650
- Executing a BT command (begin text)
  Gfx::opBeginText(), line 3585
- Executing a Tj command (showing text)
  Gfx::opShowText(), line 3709
- Executing a TJ command (showing text with individual glyph positioning)
  Gfx::opShowSpaceText(), line 3779
- Executing a Tf command (setting font)
  Gfx::opSetFont(), line 3610
- Executing a ET command (end text)
  Gfx::opEndText(), line 3595

HOWTOs

Create a PDF with human-readable objects + content streams

Put the following in the preamble of your TeX file (between \documentclass{} and \begin{document}):

   1 \pdfobjcompresslevel=0 
   2 \pdfcompresslevel=0

Create a PDF with specified crop box

Put the following in the preamble of your TeX file (between \documentclass{} and \begin{document}):

   1 \pdfpageattr{
   2   /CropBox [50 50 100 100]
   3 }

Create A PDF without page numbering

Put the following in the preamble of your TeX file:

   1 \thispagestyle{empty}