sai1919 / xml-streamer Goto Github PK

Trying to parse xml file that contains multi-byte chars sometime results in a character being replaced with two invalid characters in text nodes (and potentially other).

The cause is probably the boundary of chunks when reading the file, that split a two (or more) bytes char.

Please do note that there is security advisory for the version of lodash this package uses: https://www.npmjs.com/advisories/1065

Parsing xml nodes that contains text with new lines results in a text with no new lines.

The cause seems to be that the processText function ignores fully blank strings.

However, the expat parser doc states that : "A single block of contiguous text free of markup may still result in a sequence of calls to this handler."
The expat code seems to break for new lines and for entities. So I suppose "blank" entities will also be lost.

What about a new parsing option to preserve all text ?

Hey @Sai1919, thanks for taking care of the dependency updates. Would it be possible to release 2.0.0 on npm? This would help us getting rid of dependabot security issues.

Kind regards

When the root node is not matched, XmlStreamer will prematurely close. If using pipe, this leaves the input stream partially consumed and unclosed, requiring extra code for cleanup.

var file = fs.createInputStream('file_without_a.xml');
var streamer = new XmlStreamer({
    resourcePath: '/a'
});
file.stream(streamer);
// extra code required to close file in case the root tag did not match