[pdftex] pdflatex failures on big-endian architectures (s390x / ppc64)

[Hilmar Preuße]

On 10/28/23 09:57, The Thanh Han wrote:

Hi Thanh,

> I tested the binary /var/tmp/pdftex_hille42_20231027 and the pdf
> output has the correct CMap entries. But I haven't tested text
> extraction yet. There might be more problems than the one Ross
> discovered.
> 
Maybe I tested the pdftex binary the wrong way: I copied it to cfarm203 
and then compiled the document, by explicitly calling that program:

./pdflatex 1054218_ppc64_uncomp_20231027.tex

I didn't copy any libs, as I assumed that the patch applies to files, 
which end up in the pdftex binary itself. Then I converted the pdf file 
using gs:

gs -o a.txt -sDEVICE=txtwrite 1054218_ppc64_uncomp_20231027.pdf

as there is no pdftotext on cfarm203. The resulting text file does not 
contain the expected text. Same happens, when copying the pdf to local 
computer and calling pdftotext. Further I did a "less" on the pdf file 
and I can't confirm that the CMap looks better. Still entries like this:

/CMapType 2 def
1 begincodespacerange
<00> <FF>
endcodespacerange
0 beginbfrange
endbfrange
100 beginbfchar
<00> <0000>
<01> <0000>
<02> <0000>
<03> <0000>
<04> <0000>

No, I don't speak PDF, I don't know what that means. I diff shows just 
that the Creation and Mod Date changed. Maybe for any reason still the 
old (unfixed) pdftex binary is called..

Hilmar
-- 
Testmail

-------------- next part --------------
A non-text attachment was scrubbed...
Name: OpenPGP_signature.asc
Type: application/pgp-signature
Size: 840 bytes
Desc: OpenPGP digital signature
URL: <https://tug.org/pipermail/pdftex/attachments/20231028/90cf3442/attachment.sig>