On Thu 18 Jan 2018 at 09:55:28 -0600, David Wright wrote: > On Thu 18 Jan 2018 at 12:40:10 (+0000), Brian wrote: > > On Thu 18 Jan 2018 at 09:44:51 +0000, Curt wrote: > > > > > On 2018-01-17, Chris Ramsden <chris.rams...@gmail.com> wrote: > > > > On 17/01/18 21:42, Ben Caradoc-Davies wrote: > > > >> On 18/01/18 10:37, Joel Wirāmu Pauling wrote: > > > >>> Works fine for txt, although as it rasterizes things it's not going > > > >>> to be > > > >>> optimized for size. > > > >> > > > >> Yes, typically, but for large fonts and low resolution outputs with > > > >> few pages, rasterised pages may be smaller. > > > >> > > > >> Kind regards, > > > >> > > > > If I feed it a text file, it gives me an error: > > > > > > > > convert: improper image header `self_spam.txt' @ > > > > error/txt.c/ReadTXTImage/439. > > > > > > > > What's the trick to making it work with a text file as input? > > > > > > I ran into the same error (fixed in later versions, apparently, though I > > > am in the dark as to what version you are using and how late is later). > > > > > > I read an explicit > > > > > > convert text:mytext.txt mytext.pdf > > > > > > works. > > > > > > But it doesn't work here. > > > > > > Also: all roads lead to Rome, but some may be more suitable for wagons > > > than others. > > > > As is obvious, I hadn't realised imagemagick converted text to pdf. > > However, the command you give works for me on stable and unstable. > > It gives unsearchable PDFs. > > It serves more as a reminder of convert's way of specifying the > contents of the file than much else. Adding -font unicode also dredges > up the awful-looking ttf font. But given that the PDF produced is so > blurry, that font is probably no worse looking than any other. > > The PDF is unsearchable because it's really just an image wrapped in > a PDF container (like the output from some scanners). Being an image, > there's no concept of wrapping long lines either.
Two strikes. > I can't really be bothered to figure out what's missing on those > systems of mine that say: > > $ convert -font unifont text:/etc/default/grub /tmp/grubby.pdf > convert: not authorized `/etc/default/grub' @ > error/constitute.c/ReadImage/412. > convert: no images defined `/tmp/grubby.pdf' @ > error/convert.c/ConvertImageCommand/3210. I used FreeMono, with the full path to the font. Looks resonable until it is blown up and the bitmapped nature of the PDF shows up clearly, Strike three. I can accept that imagemagick is excellent for processing images (not that I have done much of that with it) but, for producing a quality PDF from text, it has been crossed off my list. -- Brian.
