balacode / one-file-pdf Goto Github PK

The generated png_images.pdf is about 6MB in size,
but based on my estimates it should be under 3MB:

dice.png is 420 x 315 pixels = 132,300 pixels
132,300 pixels x 3 colors = 396,900 bytes for RGB data
396,900 x 5 background varieties = 1,984,500 bytes

gophers.png is 600 x 476 pixels = 285,600 pixels
285,600 x 3 colors = 856,800 bytes for RGB data

Additional data for PDF commands, let's say 200K

The file size should be about 3MB,
but the generated demo is 6MB

** What to consider: **

• Is the library reusing all repeated images?
• Can the image byte stream be compressed to save space?

Release 1.0.1 seems to be restricted to pure ASCII.
Would it be possible to extend it to PDF encoding as described on https://gist.github.com/zmughal/5130211 ?
Unless it diverts work away from full unicode support of course.

I tried to use one-file-pdf for Swedish in https://github.com/jonasgit/pdfhw but it seems to fail. Unless I did something wrong?

Unicode support is essential in modern software, so we are going to implement some kind of Unicode support in this PDF writer. If you know or can link to any information about font embedding (to enable Unicode) or just referencing external fonts installed on the system in order to use Unicode, please let me know. Thanks for your attention.

I have been experimenting with using color PNG images with DrawImage() method, and so far found this to work with just a two line change. Is there much more than this needed in order to officially support color images instead of grayscale?

I tried to use nextline() and the x was zeroed but the Y was set o 2019.987654.
Something is not correct in the compute or the various devfaults.

Hello, first of all, I like this library, minimal yet powerful.

I have one question; does the API support streaming the pdf to disk, that way one could write lots of data to the PDF without being held in memory all at once?

Noticeable time (about 1 second) is spent when generating png_images.pdf

• Is the library regenerating every PNG image, even if it is repeated?
• If decoding and re-encoding the image data in a PDF stream is a slow operation, then goroutines could be used to speed up the process.

How to know that SaveFile failed except by reading log?
Please return errors instead of logging them.

Application code should log, libraries should not log but return errors.

Other than that good library looking forward using it. Thanks.

The metadata fields like Title, Subject, Author, etc. are not shown in the rendered PDF because the /Type/Info object isn't properly delimited by the code.

To fix this, add call to pdf.writeObj near line 1081, like so:

The flower icon in the hello.pdf file is not rendered on Debian while using the default PDF viewer. But ZapfDingbats is a standard PDF font. Probably, this may be a Debian issue, but I would appreciate any info if you experienced something similar on your OS, or PDF reader app. #

In SaveFile, ioutil.WriteFile is used to save the file without any file attributes. Therefore, the file cannot be read until one performs a chmod to set reasonable attributes.

var err = ioutil.WriteFile(filename, pdf.Bytes(), 0)

should be

var err = ioutil.WriteFile(filename, pdf.Bytes(), 0644)