58189 - file-upload should send Content-Transfer-Encoding header

Status: VERIFIED FIXED

(Core :: DOM: Core & HTML, defect, P3)

Description

[Adrian Havill]

As per the example in
<URL:http://www.w3.org/TR/html4/interact/forms.html#h-17.13.4.2>, file uploads
really should
set the Content-Transfer-Type when sending a file, as raw binary data can be
spit into the POST and web side programmers
can more easily use pre-built MIME parser contents to handle the
multipart/form-data post.

Comment 1

[Adrian Havill]

Attachment: adds "Content-Transfer-Encoding: binary" to file uploads

Comment 2

[Jacob Steenhagen]

Confirming this as it was split from bug 58033

Comment 3

[rods (gone)]

reassigning

Comment 4

[Dawn Endico]

this bug has a patch which needs to be reviewed. adding keywords

Comment 5

[Eric Pollmann]

r=pollmann for this change, but it needs to be thoroughly tested with some
existing file upload sites.  Adrian, what kind of testing have you done with
this change?

Comment 6

[Adrian Havill]

Eric Pollman wrote:
> r=pollmann for this change, but it needs to be thoroughly tested with some
> existing file upload sites.  Adrian, what kind of testing have you done with
> this change?

Currently testing against the file upload (multiform/form-data) parser library
routines included with Perl, JSP, Java servlets, and PHP. I need help testing
against the file upload (multipart/form-data) parser libraries for MS ASP-based
IIS scriptlets as I don't have access to that architecture.

In the meantime, I've deactivated this code included in bug 44464 by making a
NS4_COMPATIBLE_MULTIPART_FORM macro. To test/enable this functionality, undefine
the above or set it to false.

Comment 7

[Eric Pollmann]

Will try for Moz 0.9

Comment 8

[Eric Pollmann]

Attachment: test before checkin

Comment 9

[Eric Pollmann]

Fix checked in.  Thanks for the patch and help Adrian, sorry it took so long to 
get this checked in!

Do you have any test cases that could be used to easily verify that this bug 
was fixed?  If not I can cook one up, but let me know.  :)

Comment 10

[Vladimir Ermakov]

Verifying on build 2001-05-03-04-trunk windows 98

Comment 11

[Eric Pollmann]

See bug 83065, a regression caused by this patch.  (Of course, it was only a
matter of time before these little darlings started crawling out of the woodwork
with their beady little assumptions.)  I think the correct fix in this case is a
server-side fix, but I'm still investigating...