Maintainance take over timeline from v0.1.27 to 0.3.0

[karfau]

The purpose of this issue is to list how the package has been modified since the last version published by the original author.
The information is mostly derived from commits, github issues and is meant to be an easily readable timeline. As such it focuses only on the details of actual code changes, other changes are only mentioned.

28th Nov 2016

• v0.1.27 was published on npm by @jindw as xmldom and xmldom-alpha with the same content (except the package name). The commit was pushed by @jinjinyun on the same day. (Was tagged in xmldom/xmldom with that version by @brodybits on the 19th Dec 2019.)

29th Nov 2016

• license field in package.json was fixed in f62560d by @jindw by merging [LICENSE] Update license in package.json to be an SPDX compatible expression jindw/xmldom#178

6th Jan 2017

• Conditional converting of HTML entities was merged into master in 6b47d09 by @jindw
• BREAKING introduced regression regarding   in some (x)html documents see Regression regarding handling of   with "(x)html" mimeType #57

11th Jan 2017

56eb39f by @jinjinyun contains several changes

• dom-parser.js changed html mime type check to /\/x?html?$/, consistently in dom-parser.js
• dom.js fixed serialization issue for missing document element (example that failed on toString() before this change)
• extracted entities.js
• widened regression Regression regarding handling of   with "(x)html" mimeType #57 to all documents treated as html

8th May 2017

ffd8cac by @jinjinyun

• dom-parser.js requires entities.js initially upfront and no longer when invoking parseFromString

9th May 2017

• @jindw published ffd8cac to npm as xmldom-alpha@0.1.28 (that still has depentents)
• @jinjinyun pushed 0be2ae9 that updates the package.json to match the previously published version and removes the Object.create check from the _extends method of dom.js that added a __proto__ property (BREAKING?).
• @jinjinyun pushed 366159a that removed code that added a __proto__ property that seems to "only" be used by a "test"

19th Dec 2019

• all changes by @brodybits
• pushed tags for prev versions as stated above

on the v0.1.x branch

• was branched from f62560d created on 29th Nov 2016, so it only contained the license fix
• published and tagged v0.1.29, v0.1.30, v0.1.31 with only package.json metadata fixes.

on the master branch

• following the changes of 366159a from 9th of May 2017
• corrections in package.json

20th Dec 2019

• all changes by @brodybits

on the master branch

• formatting/corrections in package.json
• various changes in .travis.yml
• published 1bcd95e as version/tag 0.2.0 (without v prefix!), (first published version of xmldom containing the regression listed above)
• corrections in package.json
• published 1bcd95e as version/tag 0.2.1

2nd March 2020

• @kethinov announced the fork in the original repo

CHANGELOG

Starting from version 0.3.0 the changes are documented in https://github.com/xmldom/xmldom/blob/master/CHANGELOG.md

[karfau]

@brodybits I hope this is what you were thinking about when you asked for such an issue in #59 .
We could either link this issue in the CHANGELOG.md or even integrate it's content there.

Maybe we should even pin this issue, so it's easier to find and for people to ask related questions.

Feel free to edit the description as you like.

[brodybits]

Thanks for the research and updates!

My original idea in PR #59 was to track regressions and other suddenly changed behavior somewhere for discussion.

I would definitely favor moving most of this information into CHANGELOG.md, maybe part of PR #63, and just track the (possible) regressions and other suddenly changed behavior in an issue for discussion.

Here is my understanding of the possible and definite regressions found so far:

• nbsp issue raised in Regression regarding handling of   with "(x)html" mimeType #57
• from test update in 37d5fb9: "The assert.throw fails since the code no longer throws exceptions when parsing fails."
• changes related to __proto__ property of:
  • object from dom.js, now lib/dom.js - 0be2ae9
  • object from internal sax.js module, now lib/sax.js - 366159a

Some quick comments on the description:

• xmldom vs xmldom-alpha did cause me some confusion in the past, already raise discussion in xmldom vs xmldom-alpha #7
• I think the v0.1.28-not-published tag should be renamed to something like v0.1.28-xmldom-alpha or v0.1.28-xmldom-alpha-only
• the versions I published from the 0.1.x branch should be listed in CHANGELOG.md in PR docs: Update CHANGELOG.md through version 0.3.0 #63, I think
• I think the regression in Regression regarding handling of   with "(x)html" mimeType #57 was also part of the releases that I published in v0.1.29, v0.1.30, v0.1.31

[brodybits]

I pushed v0.1.28-xmldom-alpha tag to make it clear that v0.1.28 was only released in xmldom-alpha. Maybe we should delete v0.1.28-not-published tag now?

[karfau]

I pushed v0.1.28-xmldom-alpha tag to make it clear that v0.1.28 was only released in xmldom-alpha. Maybe we should delete v0.1.28-not-published tag now?

I have to tell you the both v0.1.28-... tags are acctualy on the wrong commit.
I compared the content of the package on npm with the code of the repo and the tags are one commit off, as described in the description they should be on ffd8cac.

I think it makes sense to also tag the other versions that have been published to xmldom-alpha with such a tag. So all the information is in one place.

And I agree that we should drop the v0.1.28-not-published tag, I don't think it currently helps anybody.

[karfau]

I think the regression in #57 was also part of the releases that I published in v0.1.29, v0.1.30, v0.1.31

No, since you branched it of from the license fix commit f62560d (:relieved: :+1: ), they are not part of it.

[karfau]

Question: Are you creating these changelog entries by hand? I know there are tools for that, some also work if there are silly commit messages (like in our repo in the past).
If you have any preference let me know, otherwise I will test some to add entries in #63

[brodybits]

Thanks, I may need a few days to look at this. And the existing CHANGELOG.md was made by someone else in PR #28, which I think you know is after the fact. So I would not consider the changelog to be really automated.

[karfau]

@brodybits in my comments above there are still some questions unanswered. For me to integrate this information into the changelog we need to resolve those.
I would prefer you reread my comments and if there is anything not clear we continue the discussion.
But if that doesn't work for you, I can also invest the time to list all open questions in a new comment.

[karfau]

I took the liberty to clean up the tags in the way we talked about:
I deleted the tags v0.1.28-xmldom-alpha and v0.1.28-not-published that were pointing to a wrong commit and created xmldom-alpha@v0.1.28 (to stick to a well known naming scheme and the v prefeix that was present in all versions back then) on the correct commit.