696508 - "File, Save Page As" changes the source of the page

Status: RESOLVED DUPLICATE of bug 120556

(Core Graveyard :: File Handling, defect)

Description

[Peter Brown]

User Agent: Mozilla/5.0 (OS/2; Warp 4.5; rv:6.0.2) Gecko/20110905 Firefox/6.0.2 SeaMonkey/2.3.3
Build ID: 20110905145903

Steps to reproduce:

While viewing my hompage in my browser I selected File, Save Page As...


Actual results:

The Save Page progressed fine - but the Saved Page is *not* the same source as the webpage.

For reference my homepage is http://homepage.ntlworld.com/peter.brown35/

Using File, Save Page As... I get lines like this:-

    <meta name="keywords" content="os/2, ecomstation, ecs, system builds, systems preloaded with ecomstation">

    <meta name="description" content="Another website about OS/2 and eComStation operating systems">

    <meta name="author" content="Peter Brown">




Expected results:

The original source looks like this:-

    <meta name="keywords" content="os/2, ecomstation, ecs, system builds, systems preloaded with ecomstation" />

    <meta name="description" content="Another website about OS/2 and eComStation operating systems" />

    <meta name="author" content="Peter Brown" />


So Seamonkey - and, I'm told, Firefox - are changing " />" to ">" when using File, Save Page As...


Of possible interest and definitely related:-

1] If I open my webpage in composer by using File, Edit Page from the browser the source shown in Composer has the same errors as reported above.

2] If I use View, Page Source from the browser the source shown is *correct* - and if I use File, Save Page As... from the View Source window the file is saved correctly, the code is not modified as above.

Comment 1

[Steve Wendt]

Works fine with "Web page, HTML only" but "Web Page, complete" shows the described behavior.  Does the same with:
Mozilla/5.0 (Windows NT 6.1; WOW64; rv:7.0.1) Gecko/20110928 Firefox/7.0.1

Comment 2

[Steve Wendt]

Same behavior with FF8:
Mozilla/5.0 (Windows NT 6.1; WOW64; rv:8.0) Gecko/20100101 Firefox/8.0

Comment 3

[Boris Zbarsky [:bzbarsky]]

Are you saving as "HTML only" or "complete"?

Comment 4

[Tony Mechelynck [:tonymec]]

(In reply to Boris Zbarsky (:bz) from comment #3)
> Are you saving as "HTML only" or "complete"?

See comment #1

Comment 5

[Peter Brown]

I'm missing something here as I keep seeing the question: Are you saving as "HTML only" or "complete"?

How do I check whether Seamonkey is saving as "HTML only" or "complete"? - there does not seem to be any options In Preferences for this and Seamonkey does not offer a choice when I click "File, Save Page As".

Comment 6

[Tony Mechelynck [:tonymec]]

(In reply to Peter Brown from comment #5)
> I'm missing something here as I keep seeing the question: Are you saving as
> "HTML only" or "complete"?
> 
> How do I check whether Seamonkey is saving as "HTML only" or "complete"? -
> there does not seem to be any options In Preferences for this and Seamonkey
> does not offer a choice when I click "File, Save Page As".

After selecting "File → Save Page As…" you see a file chooser. There are two possible versions of the file chooser and I don't know which one you use: I use the XUL file chooser which is not the default (i.e. I have ui.allow_platform_file_picker set to false in about:config) and at the bottom of the file chooser I see a filetype rolldown:

Web Page, complete (*.htm; *.html)
Web Page, HTML only (*.htm; *.html) (default)
Text Files (*.txt; *.text)
All Files (*)

That rolldown is what allows me to select whether to save as "HTML only" or as "Web page, complete".

Comment 7

[Tony Mechelynck [:tonymec]]

Mozilla/5.0 (X11; Linux x86_64; rv:10.0a1) Gecko/20111017 Firefox/10.0a1 SeaMonkey/2.7a1 ID:20111017003001

When saving the same page as "HTML only" and as "complete" I see differences in the saved HTML; I suppose this is also what previous commenters saw:

- The "HTML only" version has /> at the end of unpaired tags, which is normal in XHTML (and this page has an XHTML DOCTYPE line); the other has only >
- The "HTML only" version has more linebreaks and more empty lines.

Comment 8

[Boris Zbarsky [:bzbarsky]]

The "complete" mode has to serialize the DOM, so it can't preserve distinctions that are not present in the DOM, like the trailing "/".

This is a duplicate, but just resolving wontfix.... if someone wants to look up the original please feel free.

Comment 9

[Peter Brown]

(In reply to Boris Zbarsky (:bz) from comment #8)
> The "complete" mode has to serialize the DOM, so it can't preserve
> distinctions that are not present in the DOM, like the trailing "/".
> 
> This is a duplicate, but just resolving wontfix.... if someone wants to look
> up the original please feel free.


I did search for this in bugzilla before opening this bug. I found several similar bugs but not this exact bug. Sorry if it exists and I've missed it - and I have not found any matches at all on my current search. Maybe Bugzilla is having some problem as I am using the same search term that previously returned some close matches?

As for "resolved, wontfix": The browser offers to save the page then changes the source during the save and that is acceptable to you? It is not to me.

Comment 10

[Boris Zbarsky [:bzbarsky]]

The browser offers two save modes:

1)  Exactly what I'm viewing right now (with all images, stylesheets, etc)
2)  The original HTML

You picked #1.