| 353 | | # (sigh) - the odd PDF file has a length that is too long, so we'd |
|---|
| 354 | | # need to read backwards to find the "endstream" ending. Really, |
|---|
| 355 | | # who cares - I am sure this code works properly given a correct |
|---|
| 356 | | # PDF file, so I'm removing this assertion. It's not necessary to |
|---|
| 357 | | # read to the end of the object because streams are always in |
|---|
| 358 | | # indirect objects - there's never an object after this one. |
|---|
| 359 | | #e = readNonWhitespace(stream) |
|---|
| 360 | | #ndstream = stream.read(8) |
|---|
| 361 | | #assert e == "e" and ndstream == "ndstream" |
|---|
| | 353 | e = readNonWhitespace(stream) |
|---|
| | 354 | ndstream = stream.read(8) |
|---|
| | 355 | if (e + ndstream) != "endstream": |
|---|
| | 356 | # (sigh) - the odd PDF file has a length that is too long, so |
|---|
| | 357 | # we need to read backwards to find the "endstream" ending. |
|---|
| | 358 | # ReportLab (unknown version) generates files with this bug, |
|---|
| | 359 | # and Python users into PDF files tend to be our audience. |
|---|
| | 360 | # we need to do this to correct the streamdata and chop off |
|---|
| | 361 | # an extra character. |
|---|
| | 362 | pos = stream.tell() |
|---|
| | 363 | stream.seek(-10, 1) |
|---|
| | 364 | end = stream.read(9) |
|---|
| | 365 | if end == "endstream": |
|---|
| | 366 | # we found it by looking back one character further. |
|---|
| | 367 | data["__streamdata__"] = data["__streamdata__"][:-1] |
|---|
| | 368 | else: |
|---|
| | 369 | stream.seek(pos, 0) |
|---|
| | 370 | raise "Unable to find 'endstream' marker after stream." |
|---|
