Baslinux mailing list

[Sindi Keesan <keesan AT iamjlamb.com>]

Progresson finding a way to decipher pdf 1.5.

The pdftotext and pdftops and pdfimages which were compiled for BL2 by 
Doug Kaufmann 1.5 years ago won't handle pdf 1.5 but there is a new 
version of xpdf out since Jan 22 2004, version 3.00, which does pdf 1.5.  
I will attempt to compile at least the console parts of this, maybe also 
the X parts (this probably requires adding some library to compile with).

xpdf from SW71 won't do pdf 1.5 either.

Use of the pdftops part of this would produce a file that can probably be 
read with the older (Sw71, GS 5.10) ghostscript and with bmv, both of 
which will read postscript, making it unnecessary to use a newer 
ghostscript to tackle pdf 1.5.  Ghostscript keeps getting bigger.
So I will probably attempt to compile just the console parts of xpdf first 
before tacking gs 8.14.

Ghostscript 8.14 now does encrypted pdf files, the first version to do so. 
It is very large because it also does fancy things with colored ink.  I 
could not figure out which was the first gs version to do pdf 1.5.  Maybe 
none of them do it.  The changelogs are only to be found in the large 
downloaded file of source code, not online, as far as I could figure out.

I tried several times to get more info about pdftohtml via the link at the 
site to which James pointed us but gave up and it would probably have 
timed out.  But his link explained that you can generate pages of html 
with png files using -c switch.  It might be a much smaller way to view 
scanned files that were converted to pdf, instead of xpdf or ghostscript.  
The png files can be viewed with zgv or xv (which will also print them).  

xpdf was written in C++ - is that why it is slower than gs?

Another way to attack scanner output is to convince the sender to send not 
pdf files but WORD files if they are incapable of using their scanner to 
produce png or gif files. (Don't accept jpeg versions of scanned text!).  
I am told that WORD with embedded images as produced by the scanner 
program is likely to contain PNG files.  Antiword (latest version, anywya) 
can convert the WORD to ps.  (tiff > png > WORD/png > ps).  

Wvware converts WORD to html with links to images.  


On Tue, 29 Jun 2004, Sindi Keesan wrote:

> Someone sent me 2 pages that they scanned and converted to pdf.
> Xpdf displayed them as blank.
> pdf2ps produced a 0 length file.
> pdftops (DOS) told me that this is pdf 1.5 and I need 1.4.
> 
> gs 5.10 gave me a page of error messages (DOS or linux, Alladin)
> 
> I downloaded the SW81 gs 7.05 and installed it and discovered that it does 
> not support vgalib display, just X.  (To determine the display devices 
> available type gs or gs --help).  I uninstalled it as I prefer not to 
> waste time loading X every time I want to view a pdf file.
> 
> Delilinux has gs 7.05 so I downloaded that.  It is about 5M (downloaded at
> 1.3 to 3.1 k/sec) because it includes the fonts.  I did explodepkg and
> then tried typing ./gs and was told it needs libgimpprint.so.1
> 
> MANIFEST.z lists a /libgimp/ directory but not libgimpprint.
> gimp is in the gtk directory.
> Would i need to install gtk and gimp in order to run this gs?
> I have not used either of these.
> 
> Can I find just that one library for SW71?  Is there any chance that the 
> author of delilinux accidentally compiled with options such that it will 
> not work with SW71 files?  (Unlikely, his other programs have been 
> perfect - Xvesa, zgv, w3m, links2 recent version.)
> 
> The delilinux gs requires a lot of libraries not needed by the SW71 
> version (libping libjpeg and various X libraries).  It does not require 
> libvga and libvgagl, which the SW71 version does require, so probably will 
> not work with libvga anyway.  Or am I wrong?
> 
> There was also a gs 7.05 at linuxnet or linux.net but I cannot find it
> again.  Can someone please remind me of the URL.  I recall being unable to
> figure out how to download anything there so did not bookmark it.
> 
> The delilinux gs came with ps2pdf14, which converts ps to pdf 1.4, but I 
> did not see anything that will convert pdf 1.5 to pdf 1.4 - does anyone 
> know of such a program?  Perhaps the pdf2ps that comes with 7.05 will work 
> on pdf 1.5.  It is twice the size of the old one.  But it needs gs to run 
> and gives me the same error message about libbimpprint.
> 
> I will see if the SW81 gs 7.05 has a pdf2ps that works on pdf 1.5 (but I 
> need to download it again first).  
> 
> The people who sent me the pdf 1.5 eventually managed to produced a pdf
> 1.4.  They are clever, and also know how to produce gifs (I trained them
> to send mono gifs instead of 16 million color jpegs of a BW page) with
> their scanner, but most of the people who scan pages and send them as pdf
> have no idea how their 'pdf machine' software works (one of them produced
> pdf at fax resolution, which looked terrible), so I need some way to deal
> with this.  I hope not to have to compile my own gs (or use gs with X,
> because I am all set up to view with bmv in svgalib, and don't want to
> chase down ghostview and learn that).
> 
> 
> Sindi
> 
> _______________________________________________
> BasLinux mailing list
> BasLinux AT lists.ibiblio.org
> http://lists.ibiblio.org/mailman/listinfo/baslinux
>