HTML elements are reordered.

[Bogdan740]

I've uploaded an example.txt file that contains the html example. Unfortunately I couldn't get the example to be any smaller, so I apologise for the largeish html example.

When rendering the original html file in a HTML renderer you get:

TEXT 1

DKLASJD:  DASDSA ABT  31321 MT /  LSMGO  ABT dsada MT
TEXT 2
TEXT 3

Once you parse the HTML with Jsoup.parse(html) and render the new html with a renderer once again you get:

TEXT 1

TEXT 3
DKLASJD:  DASDSA ABT  31321 MT /  LSMGO  ABT dsada MT
TEXT 2

Notice how the text has been re-ordered. One thing that may help in investigation is that when you remove all of the inline style attributes, this issue no longer occurs and the text is in the correct order after parsing with Jsoup.parse().

Is this expected behaviour?

example.txt

[jhy]

Hi there,

I can't seem to replicate it. Can you show your code?

With your example HTML in try.jsoup: https://try.jsoup.org/~KDhcX48QN1pnbYhCj-WlusIyAWo

The result of body.text() is:

TEXT 1 DKLASJD: DASDSA ABT 31321 MT / LSMGO ABT dsada MT TEXT 2 TEXT 3

Which is the same as Chrome's rendered output:

TEXT 1

DKLASJD:  DASDSA ABT  31321 MT /  LSMGO  ABT dsada MT
TEXT 2
TEXT 3

[Bogdan740]

I looked into the example you shared here. I copied both the "input HTML" and the "renderer output" into an HTML renderer, and it seems like the HTML structure is being reordered, as shown in the attached image. I should also mention I'm using version 1.18.3 (which I believe is the latest version) as per https://github.com/jhy/jsoup/blob/master/CHANGES.md.

Interestingly, in the example you sent, the text appears in the correct order, but the HTML itself is still being rearranged. Do you know why this might be happening?

[jhy]

Sorry, but I can't make out what you're trying to show me in that image.

Given how badly formed the HTML is (incorrect closers for formatters, etc) there is a chance that the adoption agency algorithm is being executed differently by jsoup and (what browser are you testing? You haven't disclosed that). Or something akin.

To get to the bottom of it, my suggestion would be to write some debug code that traces the specific order of the DOM for both jsoup and browser. Add IDs to each element so we can see which one is which. Then traverse the doc and serialize the tag + ID. For the browser you can do this in JS. Then the difference is likely to be apparent.

But I am confused as to why my original output is different to yours. Can you show your code and confirm which version you're on.