[poppler] Poppler 0.21.0 (0.22 Alpha) released

[Albert Astals Cid]

El Dimecres, 3 d'octubre de 2012, a les 21:11:05, Hib Eris va escriure:
> Hi Albert,
> 
> On Wed, Oct 3, 2012 at 8:29 PM, Albert Astals Cid <aacid at kde.org> wrote:
> > El Dissabte, 22 de setembre de 2012, a les 15:55:21, Hib Eris va escriure:
> >> I compiled 0.21 and got one new warning, which I think might cause a
> >> buffer overflow.
> >> I am not sure though.
> > 
> > Is this really correct? If you have a char that is signed with -1, casting
> > it to unsigned will give you what? and will that be the array we want?
> Current code does a pdfDocEncoding[-1] which points you to an address
> before the start of the array, that seems incorrect to me.

Obvious :-)

> 
> Casting -1 to unsigned gives you 255. pdfDocEncoding is an array of
> length 256, so pdfDocEncoding[(unsigned char)-1] = pdfDocEncoding[255]
> will point to the last element of that array. Is this correct? I do no
> know for sure, but a grep on the source:
> 
> $ find -name "*.cc" | xargs grep 'pdfDocEncoding\['
> ./utils/pdfdetach.cc:          u = pdfDocEncoding[s1->getChar(j) & 0xff];
> ./utils/pdfdetach.cc:          u = pdfDocEncoding[s1->getChar(j) & 0xff];
> ./utils/pdfdetach.cc:          u = pdfDocEncoding[s1->getChar(j) & 0xff];
> ./utils/pdftohtml.cc:        unicodeString[i] =
> pdfDocEncoding[rawString->getChar(i) & 0xff];
> ./utils/pdftotext.cc:	u = pdfDocEncoding[s1->getChar(i) & 0xff];
> ./glib/poppler-document.cc:      ucs4_temp[i] =
> pdfDocEncoding[(unsigned char)s->getChar(i)];
> ./poppler/UTF.cc:      u[i] = pdfDocEncoding[s[i]];
> ./poppler/PDFDocEncoding.cc:Unicode pdfDocEncoding[256] = {
> ./poppler/Outline.cc:	title[i] = pdfDocEncoding[s->getChar(i) & 0xff];
> ./poppler/Annot.cc:        uChar = pdfDocEncoding[text->getChar(*i) & 0xff];
> ./poppler/Form.cc:    Unicode u = pdfDocEncoding[(unsigned
> int)((unsigned char)cstring[j])]&0xffff;
> 
> suggests that we do either '& 0xff' or cast to (unsigned char), except
> for in UTF.cc.
> Casting with (unsigned char) and '& 0xff' give the same result as far
> as I can see.

I do prefer the 0xff one to be honest

Pushed.

Cheers,
  Albert

> 
> Cheers,
> Hib Eris
> _______________________________________________
> poppler mailing list
> poppler at lists.freedesktop.org
> http://lists.freedesktop.org/mailman/listinfo/poppler