pdf-file not being converted

[jwilk]

Issue reported by thanks_for_the_fish at Bitbucket:

Hey,

I tried to convert one large e-book but the conversion seemed to be endless. I ripped the file in single pages and converted each one. Once again, certain pages could not be converted - I didn't get error messages from pdf2djvu. I noticed that the memory usage during the conversion-process of these files rises high - sometimes up to 800-900mb. I've attached some of these files.

cheers

---

Attachment: Page156.pdf Attachment: Page163.pdf Attachment: Page171.pdf Attachment: Page172.pdf Attachment: Page195.pdf Attachment: Page229.pdf Attachment: Page231.pdf

[jwilk]

I gave your files a try, and these are the results:

• 172 and 231 were converted successfully and quickly (less than 5 seconds each);
• 163 and 195 were converted successfully, but more slowly (20-80s);
• 171 provoked warnings from Poppler (Out of memory, Bogus memory allocation size), but apparently the conversion was successful nevertheless, and the output file looked good;
• conversions of 156 and 229 were aborted by the OOM killer.

The test was done on a x86-64 VM with 600 MB of memory. I used pdf2djvu 0.7.21, Poppler 0.26.5. I didn't pass any special options to pdf2djvu.

I will investigate this further.

[jwilk]

At least 156 makes pdftoppm run out of memory, too:

$ pdftoppm -r 300 Page156.pdf > /dev/null
Killed

I guess it'll have to be fixed in Poppler.

[jwilk]

I reported the OOM to Poppler developers: https://bugs.freedesktop.org/show_bug.cgi?id=90596

Let's see how it goes.

[jwilk]

https://bugs.freedesktop.org/show_bug.cgi?id=90596

This was fixed in Poppler 0.42.

I can now successfully convert all the PDFs attached to this issue.

Could you check if you can still reproduce the bug?