Token '-0.' treated as string rather than decimal

[jbarlow83]

The attached file produces these warnings:

checking negative_zero_decimal.pdf
PDF Version: 1.3
File is not encrypted
File is not linearized
WARNING: _this_one_warns.pdf (file position 70213): unknown token while reading object; treating as string
WARNING: _this_one_warns.pdf (file position 70988): unknown token while reading object; treating as string
WARNING: _this_one_warns.pdf (file position 71751): unknown token while reading object; treating as string
WARNING: _this_one_warns.pdf (file position 72369): unknown token while reading object; treating as string
WARNING: _this_one_warns.pdf (file position 73138): unknown token while reading object; treating as string

When QPDF rewrites the file, the rewritten version eliminates the warning.

At the file positions it complains about, the token is always -0., as one of the values in a transformation matrix inside the file.

As I understand the PDF spec (7.3.3) that token should be interpreted as a real number (sign, digit, trailing period).

A real value shall be written as one or more decimal digits with an optional sign and a leading, trailing, or embedded PERIOD (2Eh) (decimal point). The value shall be interpreted as a real number and shall be converted to a real object.

negative_zero_decimal.pdf

[jberkenbilt]

Thanks for reporting. If the qpdf tokenizer is failing to interpret a number with a trailing period as a floating point number, this will be an easy fix, and I can get it in next time I make some updates.