This change is based on @jelmersnoeck's initial PDF comparison code. In order to compare two PDFs, their internal resource catalogs must be similarly sorted and their internal creation timestamps must be the same. Methods were added to enable this. The catalog order needed to be addressed because gofpdf uses hash maps for internal resources and, when writing them to the PDF document, the non-deterministic iteration order resulted in inconsistent internal structure.

A helper function to compare PDFs was added. This is called by the example.Summary method if a reference file corresponding to an example PDF is found in the pdf/reference directory.

This change has not been tested on Mac or Windows yet, nor with versions of Go earlier than 1.5. Verification that it works in these different environments will be appreciated.

A desirable side effect of having the reference PDFs committed to the repository is that we can include links to them in the documentation.

Is this possible? I would use go to query the pdf in my database, then get additional data using a JSON api to generate a new pdf with gofpdf, then combine them with gofpdf into a single pdf file.

At the moment, the tests validate that the code runs. It also provides a lot of examples for the users of gofpdf.

However, it doesn't test if functionality is as expected. This is particularly important when changing functionality or adding new features.

My suggestion for this would be to have a set of pre-defined PDFs (the ones that are generated through the examples now) and compare the same PDF generation instructions to those PDFs. This assures that new/changed code wouldn't change old output (unintentional).

To do this, there needs to be a change to the way the file gets parsed. At the moment, time.Now() is used to put the creation date in place. This makes it tricky to compare the file as this date always changes and thus the output will always be different.

I'm suggesting a simple Clock implementation here to make it possible to mock this out during tests.

I have made a rough version of how this could work here: https://github.com/jung-kurt/gofpdf/compare/master...jelmersnoeck:test-comparison?expand=1

Would you be open to accept this kind of system for testing output? (Note, the version above is a POC, the actual implementation will probably be a bit different)

This is a wrapper around manually having to download the image and registering it. It registers an image based on it's URL, it will download that image and add it to the PDF. It will however not add it immediately to the page. This is done by calling Image() with the same URL.

As discussed in https://github.com/jung-kurt/gofpdf/pull/35, this contains the changes within a single commit.

Hi Jung Kurt

Please assist as i am recieving an index out range error on fpdf.go link 2578 where the font is set after deploying to openshift, yet the code is working locally.

Please advise what if the library would behave differently on different environments

This lets you register an image that is hosted somewhere else. This way you can include an image from another service and you don't have to store it locally.

I'm fairly new to Go and was just wondering what the best way to run the test was? At the moment I made a package in my go folder github.com/jung-kurt and added a symlink to my local repository to run the tests from there.

GetFontSize returns the total height of the text (from Desc to Asc in picture below), but I need to know the baseline. For example in the text "go" I'd like to know how to calculate the bottom of the "o", not of the "g", is this possible (base in picture below)? I have the feeling this info could be parsed somewhere in the getInfoFrom* functions in font.go.

I realized last night that I introduced a bug with the addition of Encoding and Decoding templates. When a template is created it is given an ID by the GenerateTemplateID method. This worked in the past because a template could not exist outside the context of it's original gofpdf package. However now that templates can be saved to files, or templates can be generated across a cluster of machines maintaining the original ID for a template will quickly create Template ID conflicts where many Templates claim to have an ID of 1. I'll start working on a PR to fix this, but figured I should bring it to your attention.

Thanks!

Implementation of templating in the style of FPDF_TPL. There are two parts to this: a Template interface and a specific implementation FpdfTpl. It diverges from the PHP original in that the method Fpdf.CreateTemplate uses a callback in which you write the template and then returns a read-only template, which should help ensure type safety and should avoid confusion.

The interface allows other code to insert templates by other mechanisms; the intention is to eventually use this to implement the document-reading functionality of FPDI. It may have other uses, such as a mechanism for reading SVGs?

Also, I added a few more util methods that I needed, and that may prove useful to others: PointType.Transform, SizeType.Orientation, SizeType.ScaleBy, SizeType.ScaleWidth, SizeType.ScaleHeight. The PHP version of these made heavy use of optional arguments (pass either an integer or null to get different behaviour), so I had to separate the use cases into different methods for the Go port.

Are the test cases sufficient? go test -cover reports 85.4% coverage (up of 0.2pp from 85.2%), and I test some edge cases like images within subtemplates. However there are cases the tests definitely don't cover, such as creating a template divorced from a document. I may add more tests.

Apologies that this took so long to fix up. It was mostly implemented months ago, but life has not been generous lately.

As the number of tutorials are growing, it gets hard to quickly look up some example code. Now you first have to find a relevant function, read the tutorial number and scroll back. It would maybe more efficient to have some keywords in the title instead of just a number.

So instead of:

• ExampleFpdf_tutorial31

it could be:

• ExampleFpdf_tutorial31_line_caps_join_style (does not give golint error)
• ExampleFpdf_tutorial31LineCapsJoinStyle
• ExampleFpdf_31_line_caps_join_style (does not give golint error)
• ExampleFpdf_31LineCapsJoinStyle

When in a template and using AutoSetPageBreak(true, 0) the template will only render the last page within the template. I am fully aware that the Tpl isn't intended to add new pages, so this isn't really a bug, however I would like to know if you put any thought into making it allow pages and any ideas you may have before I try tackling it.

Thanks!

When I use a font file to write PDF, when the file contains unsupported characters. Expect to return an obvious error instead of an error "jung-kurt/gofpdf/fpdf.go:2789". Or you can skip unsupported data，Continue to write。

location: jung-kurt/gofpdf/fpdf.go:2789 ==》 l += float64(cw[int(c)])

error: panic: runtime error: index out of range [127775] with length 65536 [recovered] panic: runtime error: index out of range [127775] with length 65536

Hello, First of all, thanks so much for the great library, it's awesome!

Currently I am using this font (https://fonts.google.com/specimen/Tajawal?selection.family=Tajawal). I downloaded the ".ttf" file, and use it as follows:

pdf.RTL() pdf.AddUTF8Font("tajawal", "", "Tajawal-Regular.ttf")

The result is: ب س م ا ل ل ه ا ل ر ح م ن ا ل رح ي م While it should be: بسم الله الرحمن الرحيم

Is there anything else I can do to be able to generate Arabic text? By the way, the Arabic text in the test-case ExampleFpdf_AddUTF8Font are generated in the same way (not connected), and I also tried different fonts, all lead to the same result.

Hello everyone! This is my first question at Github! My task is to check how many pages contain a pdf document and if it has less than 8 pages print it silently. I'm using this imports

	"github.com/jung-kurt/gofpdf"
	"github.com/jung-kurt/gofpdf/contrib/gofpdi"

and code below ```

                       var err error

			pdf := gofpdf.New("P", "mm", "A4", "")

			pdfBytes, err := ioutil.ReadFile(DocName)
			if err != nil {
				panic(err)
			}
			rs := io.ReadSeeker(bytes.NewReader(pdfBytes))
			// Import in-memory PDF stream with gofpdi free pdf document importer
			imp := gofpdi.NewImporter()

			// import first page and determine page sizes
			tpl := imp.ImportPageFromStream(pdf, &rs, 1, "/MediaBox")
			pageSizes := imp.GetPageSizes()
			nrPages := len(imp.GetPageSizes())

			// add all pages from template pdf
			for i := 1; i <= nrPages; i++ {
				pdf.AddPage()
				if i > 1 {
					tpl = imp.ImportPageFromStream(pdf, &rs, i, "/MediaBox")
				}
				imp.UseImportedTemplate(pdf, tpl, 0, 0, pageSizes[i]["/MediaBox"]["w"], pageSizes[i]["/MediaBox"]["h"])
			}
			pdf.OutputFileAndClose(DocName + "test")

			pages := pdf.PageCount()
			fmt.Println(pages)

nrPages and pages is a places where I'm trying to get number of pages. And it's not always gives me proper number, but I really need it, sometimes I'm receiving files that contain up to 800 pages and I don't want to print them. I'm new at go, using it only for a week and maybe I just did something wrong. Here is a two examples of files, one returns proper number of pages the other one is giving a wrong number. The result files is those that I'm saving by trhis: pdf.OutputFileAndClose(DocName + "test") Incorrect number of pages.pdf Incorrect result.pdf Proper number of pages.pdf Proper result (2).pdf

Next step is to print it. So I've found solution with cmd and Acrobat Reader DC. It's here "C:\Program Files (x86)\Adobe\Acrobat Reader DC\Reader\AcroRd32.exe" /t D:\reports\print.pdf" "HP LaserJet 1018" "oem34.inf" "USB001".

Or maybe I will use a browser in kiosk-printing regime and browser extension that will do window.print().

So far I haven't found any better solution, but I'm going to dig it somehow. Thank you for your attencion.

Wondering if there is any information about how to managing printing on desktops in golang.

This is the sequence I am trying to achieve

After the user has selected a pdf template and my code has merged data into the pdf I want to let the user print it.

Opens the print dialogue for the OS. User selects printer, orientation, paper side. The orientation and paper size is used to defender the pdf correctly. Then user is show new pdf Then user confirms and prints Then system send to print spooler.

Am looking for tips about how to access these OS specific APIs. It's not PDF related but I would think others may find this useful. Hoping I can work this out and put into contribution repo

Based on this page https://github.com/jung-kurt/gofpdf/blob/master/contrib/gofpdi/gofpdi_test.go currently I have document with multiple orientation page. Legal page with portrait orientation and legal page with landscape orientation. After run code, unfortunately all of pages become legal portrait orientation. Some of content inside pages become stretch too. Thanks!