Fix ZeroDivisionError on broken image files

[timobrembeck]

Short description

This error happened on images where the width could not be correctly identified:

Traceback (most recent call last):
  File "/home/xhtml2pdf/document.py", line 196, in pisaDocument
    doc.build(context.story)
  File "/home/.venv/lib/python3.11/site-packages/reportlab/platypus/doctemplate.py", line 1082, in build
    self.handle_flowable(flowables)
  File "/home/.venv/lib/python3.11/site-packages/reportlab/platypus/doctemplate.py", line 931, in handle_flowable
    if frame.add(f, canv, trySplit=self.allowSplitting):
       ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/.venv/lib/python3.11/site-packages/reportlab/platypus/frames.py", line 212, in _add
    flowable.drawOn(canv, self._x + self._leftExtraIndent, y, _sW=aW-w)
  File "/home/.venv/lib/python3.11/site-packages/reportlab/platypus/flowables.py", line 112, in drawOn
    self._drawOn(canvas)
  File "/home/.venv/lib/python3.11/site-packages/reportlab/platypus/flowables.py", line 93, in _drawOn
    self.draw()#this is the bit you overload
    ^^^^^^^^^^^
  File "/home/xhtml2pdf/xhtml2pdf_reportlab.py", line 716, in draw
    Paragraph.draw(self)
  File "/home/xhtml2pdf/reportlab_paragraph.py", line 1212, in draw
    self.drawPara(self.debug)
  File "/home/xhtml2pdf/reportlab_paragraph.py", line 1769, in drawPara
    dpl(tx, offset, lines[0], noJustifyLast and nLines == 1)
  File "/home/xhtml2pdf/reportlab_paragraph.py", line 379, in _leftDrawParaLineX
    _putFragLine(offset, tx, line)
  File "/home/xhtml2pdf/reportlab_paragraph.py", line 219, in _putFragLine
    drawing = cbDefn.image.getDrawing(w, h)
              ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/xhtml2pdf/xhtml2pdf_reportlab.py", line 538, in getDrawing
    scale_x = width / drawing.width
              ~~~~~~^~~~~~~~~~~~~~~
ZeroDivisionError: float division by zero

[stefan6419846]

Do you have an example for this to provide a corresponding test as well? Where is the width of zero coming from?

[timobrembeck]

@stefan6419846 if the image file is broken, in my case if was an incorrectly configured web server which returns html instead of an image until the cookies are accepted:
https://foresight4youth.com/wp-content/uploads/2018/07/Logo-Caritas-Dortmund-300x124.jpg

This error also happens when normal text files are embedded as images.

Of course, this is a content problem and not necessarily task of the library to prevent such mistakes, but in my opinion it would be more convenient if the PDF would be created nevertheless and only throw a warning in the logs, not cause an internal server error.

[timobrembeck]

@stefan6419846 Do you think this will do? Unfortunately we don't have access to the file name in that scope, and it would be a bit complicated to escalate the exception down to the scope that has access to the filename...

[stefan6419846]

We do not have the filename there yet, yes. But shouldn't we be able to extend PmlImage to hold the src value as well? (We might have to account for non-filenames in the src attribute for display, but this should be doable as well.)

[timobrembeck]

@stefan6419846 cool idea, maybe like this?

[stefan6419846]

LGTM, as it makes the error message more helpful/direct.