jbowtie / ratago

Go implementation of an XSLT 1.0 processor

ratago

Ratago is a (mostly-compliant) implementation of an XSLT 1.0 processor written in Go and released under an MIT license.

Currently it should be seen as experimental - it lacks full compliance with the spec. It has been run successfully on a number of scripts of moderate complexity as of the 0.4-pre release.

The test suite is derived from the test suite used by the libxslt library written by Daniel Veillard. See http://xmlsoft.org/XSLT/ for details on libxslt.

Installation

For MacOS:

# Need pkg-config, see https://stackoverflow.com/a/36794452/700471
brew install pkg-config 
# Need libxml2 source, see https://github.com/mitmproxy/mitmproxy/issues/68#issuecomment-120301708
brew install libxml2
sudo ln -s /usr/local/opt/libxml2/include/libxml2/libxml /usr/local/include/libxml 
# Install with Go Modules
GO111MODULE=on go get github.com/jbowtie/ratago

Example Usage

ratago -indent transform.xslt data.xml > result.xml

TODO

There are several tasks remaining to reach full compliance. Until these tasks are complete the API is subject to change.

• Implement xsl:decimal-format and format-number.
• Ensure that errors are properly progogated in Go fashion.

ratago's Issues

Support for EXSLT math core functions

As we've started adding support for other functions defined in EXSLT math we should support the core functions (math:min, math:max, math:highest, math:lowest)

See http://exslt.org/math/ for documentation.

copy-of should explicitly set the empty namspace when needed

There is a test for this in general/bug-103.xsl

I'm unclear on how well we support the empty namespace is general, worth review.

Not all format-number tests pass

From the libxslt suite tests 61, 75, 95 fail when they are enabled. They need deeper investigation.

Implement unparsed-entity-uri

Tested by general/bug-21-.xsl

This will need to get an NDATA ENTITY declared in the DTD from Gokogiri somehow; most likely we'll need an upstream patch to expose xmlGetDocEntity().

copy-of should preserve an attribute prefix when plausible

I think this only comes up when there are multiple prefixes mapped to the same URI. Tested by general/bug-104.xsl

Error with XSLT contain URL on include element

Error with XSLT contain URL on include element

This error:

Fixed:

travis-ci build panics running xsltREC test

Need to see if we can replicate and determine cause.

Support for EXSLT func

The general test suite from libxslt exercises this in (currently not run) tests 137, 174, 178

See http://exslt.org/func/ for documentation.

ThomsonReutersEikon/gokogiri doesn't work with go1.6

Panics when trying to use it with go1.6: golang/go#12416

Theres' an issue on the moovweb repo: moovweb/gokogiri#92

Issues with some xsl:key usages

In the libxslt suite see tests 70, 146, and 134.

StylesheetOptions.Parameters is not used

StylesheetOptions.Parameters is not used within the codebase during stylesheet processing, so passing named parameters seems to be broken.

namespace declarations do not appear in deterministic order in output

Technically this isn't really an issue - the output is correct according to the XML spec - but it causes several tests to fail seemingly at random.

For example, see REC/test-7.1.1 (which is currently disabled pending investigation into this issue).

Making the test harness smart enough to recognise XML output and do canonicalization first would be a reasonable fix. Alternatively, sorting by prefix before writing (and fixing up the associated expected output) would be another way to fix it.

Fatal errors when running

More of a question than a bug. I'm trying to use this package in a project and get the following error when running go run *.go:

toby:go-xml-to-csv robert$ go run *.go
^[^[# github.com/jbowtie/gokogiri/help
../../go/src/github.com/jbowtie/gokogiri/help/help.go:6:10: fatal error: 'libxml/tree.h' file not found
#include <libxml/tree.h>
         ^~~~~~~~~~~~~~~
1 error generated.
# github.com/jbowtie/gokogiri/xpath
../../go/src/github.com/jbowtie/gokogiri/xpath/expression.go:4:10: fatal error: 'libxml/xpath.h' file not found
#include <libxml/xpath.h>
         ^~~~~~~~~~~~~~~~
1 error generated.

Do I need to do some kind of configuration with cgo to get things building/running? The gokogiri doc says to apt-get libxml2-dev; I've installed libxml2 via brew since I'm on os x (no libxml2-dev package was known to brew).

document('href') should return the root node of a newly parsed document

Need to handle this; we'll want to store a map of input documents to avoid repeated parsing.

Need tests for EXSLT math functions

The newly added EXSLT math functions need some test coverage.

Small test cases patterned after the other tests in stylesheet_test.go are best (see TestNaive for a template).

document('') should resolve to the stylesheet

This should already be implemented, however it seems to fail. This may require additional manipulation of the XPathContext. Tests general/array.xsl and general/bug-38-.xsl rely on this function.

Some tests fail due to attribute order

Technically we should only compare documents after canonicalization, which renders attribute order moot. Practically we need to determine why only a tiny subset of tests are affected by attribute order, none of which used to be.

decimal-format and format-number are not implemented

These are the last bits that need to be implemented for spec compliance.

Implement EXSLT node-set

The most widely-used extension function in the EXSLT namespace.

Support for EXSLT date

The general test suite from libxslt exercises this in (currently not run) test date_add
See http://exslt.org/date/ for documentation

How does StripSpace work

Hey @jbowtie this XSLT library looks good. I have been using it today :)

Can you please show me an example of how StripSpace and indentation works ?

Thanks 👍