Baslinux mailing list

[sindi keesan <keesan AT sdf.lonestar.org>]

On Mon, 22 Jan 2007, Lee Forrest wrote:

> On Mon, Jan 22, 2007 at 06:40:33AM +0000, sindi keesan wrote:
> > On Sun, 21 Jan 2007, Lee Forrest wrote:
> >
> > htmldoc is available as open source 3.2MB tar.bz2 download.  It can
>
> [delete]
>
> > The dependencies for Slackware 4.0 ghostscript are the two svgalib
> > libraries:
> >
> > libvga.so.1
>
> 370K
>
> Big library.
>
> > libvgagl.so.1
>
> 58K
>
> I don't have those anywhere on BL.

You also don't have a video chip that supports svgalib.

> > (which I just packaged together with the newzgv from BL1 and uploaded to
> > http://keesan.freeshell.org/bl/zgv-bl3.tgz)
>
> That's good.

Only if you can use svgalib.   Otherwise add some other viewer such as xv 
which will view more formats than xli does (such as pbm).

gs will also work under X, and you can specify X as the default, 
in /etc/profile, export GS_DEVICE=x11 (?).  GS for SW4.0 and 7.1 use 
'vgalib' as the default device.

It might be possible to compile gs not to require these two libraries.

> > libc.so.5
>
> 570K
>
> > libm.so.5
>
> 32K
>
> Got those. Seem to be standard.

Almost nothing but busybox or other uclibc-compiled things will work 
without them in BL3.

> > Slackware 4.0 gs was compiled not to require X, and uses svgalib to
> > display files instead.  gs itself is 1.3MB, there are another 863k of
> > files in one directory and documentation.  And you probably need more than
> > the basic font selection that I packaged for use with the xpdf CLI
> > programs because I don't have it displaying anything with fonts yet, tho I
> > made a symlink to /usr/share/ghostscript/5.10/fonts from
> > /usr/local/share... (where xpdf programs were looking for the fonts).
> > It displayed a scanned image file and zeroline.ps.

> Could be that xpdf wants fonts that encoded differently than gs.

No, it uses glibc X not libc5 X.  I had the same problem trying to use a 
glibc gs, would not even see the older libraries, which it required.

> > ps2pdf may be working but I cannot view the output with gs (or any other
> > pdf file) yet.

It was not working, could not find enough fonts.

> > gs seems far smaller or at least simpler than HTMLDOC-cum-dependencies and
> > should be adequate to print the output of Abiword, either as user with
> > printcap and lpr, or as root after producing a ps file with Abiword and
> > then using the gs-based pstopdf|pdftoppm|pbmtolj > /dev/lp0 (pdftoppm -
> > xpdf, pbmtolj - netbbm)
>
> Looks like they may be about the same size, but htmldoc is _way_ simple
> to use:
>
> htmldoc -f foo.pdf foo.html --webpage
>
> That's it. And you have pdf doc from a webpage. The pdf doc could
> be many pages long, if the webpage was long. It has lots of other
> options, but they are easy to understand. You can make a whole book
> from webpages, ready for the printer and bindery, in a jiff.

Sounds good, but large and time-consuming to gather all the libraries and 
then also compile it, and it needs kernel 2.4.

Opera can apparently make a ps file from a webpage, also printable.
Or gs can convert it to pdf.  Can opera convert html to ps CLI?

> A bash script that would convert _simple_ webpages into ps wouldn't be that
> hard, I _think_. PS is all text (pretty sure), so images must be converted
> to ascii.

Ron might be interested.  I have now conquered 'user' and will progress to 
lpr as root.  Learned a lot on the way.

> It's just another markup language like html/tex/groff...One that
> printers understand. (I hope I'm right about that last.)
Expensive printers.

> I just made a PS3 doc from a simple webpage (htmldoc) with a few
> short paragraphs and one image. Total html file length 1K. Image
> 62K jpeg.
>
> The vast majority of the 807K PS3 doc is the image converted into
> ascii, like mimencode or uuencode, but vastly less efficient.
> The same image mimencoded is 83K.

pdf compresses images.  PS prints faster but takes up more space.

> I don't know what kind of ascii-encoded-image that is, but a
> utility that converted images into that format would be needed
> for the script. If nothing else, it could be pulled out of the
> htmldoc code, but that shouldn't be necessary.
>
> (The same webpage converted to PDF is 457K. But it's binary.)

That is also a lot less efficient.  People keep scanning images on their 
'pdf machines' which make pdfs out of them, instead of just making 
unwrapped images.  200K per page instead of 50K.  pdfimages unwraps the 
images to ppm form and you can then turn them into small gifs.  The pdf is 
the default for some scanner program, unfortunately.

> I'm talking about a script that deals with webpages that could be
> printed just as they are displayed on the screen, one page of ps
> per screenful of processed html. What you see is what you get. It
> will _all_ print. And could easily be written manually. HTML1 or
> 2.
>
> You could make tasteful books/magazines/newspapers with it.

Personally I prefer plain text, which is why I print webpages that way.


> Writing such a script for _all_ webpages would be a formidable
> task, and that's the least of the reasons I have no interest in
> such a project.
>
> (A script to convert modern-dreadful HTML to basic HTML would be
> great. A filter to run every webpage throught _before_ it got
> to your browser.

Might slow things down on a slow computer.

> Here are a couple of interesting PS links:
>
> http://www.linuxjournal.com/article/2386
>
> http://www.adobe.com/cfusion/search/index.cfm?loc=en_us&term=postscript+language
>
> Docs are not a problem, but you need a PDF viewer to display them! :-)


To avoid dependencies you can convert online:

pdf to html or text:

Google converts pdf to html or text for you (but very poor quality text).

Adobe does free online conversion of pdf to html (for people with 
disabilities).

http:// www.pdfonline.com  Upload a document you want converted from html 
to pdf.

Other formats to pdf:

http://www.k2pdf.com/convert.html  converts just about everything except 
html to pdf, free online:  rtf txt WORD Excel Powerpoint ps gif jpg png 
tif zip.....  Upload your file to their site.  Works with Opera 8.52.

If you don't give them your email address they only do .doc .xls .ppt .txt 
and .rtf.

It took about 30 sec to upload and convert 1 line of text, which I can 
retrieve from their URL within 24 hours.

No dependencies but upload and download required.

I could not find any CLI linux program that converted ppt to anything so 
someone teaching a course I audited printed it for me.


Gmail converts various formats to HTML including ppt, xls, doc, pdf, and 
rtf.

Xpdf's pdftops or gs's pdf2ps for ps from pdf.

> Writing basic PS by hand would not be practical. But writing
> basic HTML by hand, _is_. It would be possible to write an
> interactive script that called basic PS formatting functions,
> but we all know and use basic HTML anyway.

What basic functions are you thinking of?
I am told HTML needs css to do hard page breaks.


> Any such script is now at the bottom of a list of projects I need
> to tackle in order. First up is getting an Xserver that will get
> good images out of my i810 chip.

> Am working on building 3.3.6 on my uclibc OS. Has an i810 server
> with all the functionality _I_ need.

What progress?

> Instead of trading libc5 for libc6, I plan on trading libc5 for
> uclibc (Which is half the size of libc5 but does everything I
> need (keep your fingers crossed) that libc6 offers. I built ppp
> (pppd,chat...) and setserial on it today, and it already has most
> of what the stock BL3 has, including busybox.

Are you compiling dynamically against uclibc?  I think Steven was 
considering that eventually, maybe after more bugs are worked out.

> When I get everything I have on BL3 on uclibc, I'll move the libs
> and executables over to BL3. They both use the same kernel.
>
> Hopefully it will be less than 20M without a compiled-apps
> development environment.


> [delete]
>
> Lee
>
> --
> BasicLinux: Small is Beautiful
> http://www.basiclinux.com.ru
>
>
> -----------------------
> BasicLinux mailing list
> -----------------------
> http://www.basiclinux.com.ru
> http://www.ibiblio.org/pub/linux/distributions/baslinux/
> ------------------------------------
> To exit, send subject=unsubscribe to
> baslinux-request AT lists.ibiblio.org

keesan AT sdf.lonestar.org
SDF Public Access UNIX System - http://sdf.lonestar.org