302597 - Unable to paste HTML-formatted text after copying part of a table contained in <span>

Status: NEW

(Core :: DOM: Editor, defect, P5)

Description

[Stephen Donner [:stephend] Not actively reading bugmail]

Build ID: 2005-07-28-06, Windows XP SeaMonkey trunk.

Steps to Reproduce:

1. http://www.fixedorbit.com/AS/10/AS10565.htm
2. Select from the A in ASN Name down to the O in Verio
3. Attempt to paste into the HTML compose window in mail.

Expected Results:

Able to paste

Actual Results:

Silent failure to paste (no JS console messages logged)

Raw HTML source:

<b>ASN</b></td><td width="80%"><b>Name</b></td></tr><tr><td>2914</td><td><a
href="/AS/2/AS2914.htm">VERIO</a>

Plain-text output

ASN	Name
2914	VERIO

Comment 1

[Mike Schroepfer]

Can anyone confirm that this occurs in FF or TBird?   Until then minusing for beta.

Comment 2

[Stephen Donner [:stephend] Not actively reading bugmail]

(In reply to comment #1)
> Can anyone confirm that this occurs in FF or TBird?   Until then minusing for
beta.

Yes, I can confirm that.  Tried with today's Firefox trunk and yesterday's
Thunderbird trunk; can't paste.

Comment 3

[Asa Dotzler [:asa]]

When did this regress and can we get a reduced testcase? It appears that
selecting and copying only part of the table causes the failure but the full
table works. 

Comment 4

[Stephen Donner [:stephend] Not actively reading bugmail]

ASN	Name
2

That's the reduced testcase.  If I select just ASN    Name, it works, and pastes
a table with a row.  Sorry, no idea when this regressed.

Comment 5

[Mike Schroepfer]

Blake - can you take a look and figure out where to redirect? Unless we can get
someone to own this looks like it is not going to make it for 1.5b

Comment 6

[Stephen Donner [:stephend] Not actively reading bugmail]

Better steps:

1. Load http://www.fixedorbit.com/AS/10/AS10565.htm
2. Select from the "e" in Name _to_ the "2" in 2914.

That gives you the following HTML:

e</b></td></tr><tr><td>2

Comment 7

[Martijn Wargers (dead)]

This also doesn't work for me with Mozilla1.7

Comment 8

[Stephen Donner [:stephend] Not actively reading bugmail]

Attachment: Much-needed local testcase.

Select from B down to 2, then try to paste into Thunderbird or SeaMonkey.

Comment 9

[Stephen Donner [:stephend] Not actively reading bugmail]

Mats--think you might be able to take a look?

Comment 10

[Uri Bernstein]

Attachment: minimal testcase

The key here is the inline element (<span> in this testcase, <font> in the
original) which wraps the table. Without it everything is OK.

Comment 11

[Uri Bernstein]

What's being copied into the clipboard in this case is:

<span><tr><td>B</td></tr><tr><td>2</td></tr></span>

Notice that while the enclosing <span> is included, the <table> is not. This
probably later confuses the paste operation.

The reason that the <span> is included, but the <table> is not (although they
are both common ancestors of the selection start and end points) is that
nsDocumentEncoder::IncludeInContext(), called by
nsDocumentEncoder::SerializeRangeContextStart() for each of the common
ancestors, returns TRUE for <span>, <font>, etc., but FALSE for <table>.
(see
http://lxr.mozilla.org/mozilla/source/content/base/src/nsDocumentEncoder.cpp#1252 )

So that's (I think) the problem. I'll leave it to someone with better
understanding than mine to come up with a solution.

Comment 12

[Uri Bernstein]

Attachment: one option

This is kind of hackish: 
Exclude from the "context" anything which is outside the (innermost) <table>
element in the common ancestors list.

If anybody thinks this makes sense let me know and I'll find out who I should
ask for a review. I'm off to sleep now.

Comment 13

[Uri Bernstein]

more descriptive summary, to help searching.

Comment 14

[Robert O'Callahan (:roc) (email my personal email if necessary)]

I don't really have a clue what's going on here. How about Blake?

Comment 15

[Blake Kaplan (:mrbkap) (inactive)]

Comment on attachment 199863 [details] [diff] [review]
one option

I'd really rather not do this if possible. It seems that the hacks to paste tables would be lining up one after the other (see also bug 228920). I'll look into this and see if I can see what's going on to fix it in a cleaner way. So I'll minus this for now, but if I can't figure out a cleaner way, I'll plus it.

Comment 16

[Al]

Is this the same as bug 428096?

Comment 17

[Mats Palmgren (inactive)]

No, bug 428096 is OSX-only.

Comment 18

[Jens Stutte [:jstutte]]

Bulk-downgrade of unassigned, >=5 years untouched DOM/Storage bugs' priority.

If you have reason to believe this is wrong (especially for the severity), please write a comment and ni :jstutte.