Open GoogleCodeExporter opened 8 years ago
That's indeed a bug. Looks like a regression, but I'm not sure.
Original comment by guy...@gmail.com
on 9 Sep 2009 at 2:50
Look at the comment # 2 for issue #11:
http://code.google.com/p/daisydiff/issues/detail?id=11&can=1#c2
If you re-word this answer to this current example:
The reason for "another" being outside of the table is that:
"another" never was a part of a table without attribute cellpadding="5",
so it is not honest to render it inside such a table.
The fix with side effect for the issue 11 though would make your example much
pretier.
You will retain your table structure, but the removed word will be placed
outside of
the table for the reasons above.
The fix for the bug 11 is not in the SVN head, but you could take it from issue
11
attachments here:
http://code.google.com/p/daisydiff/issues/detail?id=11&can=1#c4
Also, the side effect is discussed in more details here:
http://groups.google.com/group/daisydiff/browse_thread/thread/21fb89ead6feda01
Attached is a zip of version 1.1.2 with the fixes that were done after the 1.0
is
released. Look in the readme.txt in the "Update Notes" for detailed info.
It does contain the issue 11 fix with its side effect.
Original comment by anastass...@businesswire.com
on 30 Oct 2009 at 3:34
Attachments:
Do you think it's best to commit this to trunk?
Original comment by guy...@gmail.com
on 9 Nov 2009 at 2:19
Anastassia can you post also the source for 1.1.2?
Or it just contains the same files mentioned in Issue 11 (Only changed files
are
TextNodeComparator.java, TagNode.java, TableStatics.java)?
Thank you.
Original comment by kkape...@gmail.com
on 9 Nov 2009 at 3:57
Kkapelon:
The difference between the source for 1.1.2 and the current trunk head is the 3
files from the issue 11 attachment.
Those files were not posted in the trunk head because of the side effect
(reversed
order of removed content that has to be placed outside of the structure). But
you do
have all the source available.
The current trunk head directories marked as the revision 134, however you
won't
find a file older than 127 - because revisions 128-134 were on the branch. The
branch is not a working version atm.
The difference between the source for 1.1.2 and 1.0 is more than the 3 files.
I don't know what revision 1.0 was build on. So I can't tell you.
Guy:
We probably should tag revisions with the corresponding build number if the
build
occurs - that way people can find answers on the questions like that.
About whether to commit it to the trunk: for me either way is fine, and it is a
question what is more important for you and the majority of people:
the order of differences
or the HTML structure of the document.
If you want to make this temporary fix a middle step (because at least it does
fix
the invalid HTML output) - let's post it to the trunk.
If you want to offer people a choice of whether to be a victim of the side
effect -
it's better to keep it separate for now as the issue attachment.
Thank you.
Original comment by anastass...@businesswire.com
on 9 Nov 2009 at 4:32
True.
Either way is fine by me aswell, as I'm not using DaisyDiff myself.
I suggest we keep it out of trunk until we had some decent QA, which I'm unable
to
do. Please let me know what I can do to facilitate work on this.
Original comment by guy...@gmail.com
on 9 Nov 2009 at 4:47
Original issue reported on code.google.com by
kkape...@gmail.com
on 21 Aug 2009 at 2:57Attachments: