Bug 21645 – REGRESSION (r37469-r37605): File Input not Uploaded in Multipart/Form-data Form

• Status: RESOLVED
• Resolution: FIXED
• Component: Forms

[john]

The $_FILES variable in php is not getting populated at all when a multipart/form-data ecntype form is submitted.  It works fine in the shipped version of Safari 3.1.2, but the nightly build fails to populate it (which tells me the form is not being submitted properly?).

[Robert Blaut]

Last working nightly r37469. Confirmed. I saw the bug for the first time in Textpattern control panel during trying uploading file with multipart/form-data form.

Possible duplication: bug 21518.

[Robert Blaut]

Possible duplication: bug 21707

[Robert Blaut]

Possible duplication: bug 21730

[Timothy Hatcher]

This prevents uploading patches to Bugzilla too. Sam thinks it was a change Darin did.

[Darin Adler]

Should be very quick to fix.

[Darin Adler]

Created attachment 24497 [details]
test

[Darin Adler]

Created attachment 24498 [details]
patch

Comment on attachment 24498 [details]
patch

+        This "two at a time" a confusing way to do things, which explains why I got

Typo (missing "is"?).

r=me

[Darin Adler]

http://trac.webkit.org/changeset/37685

[Mark Rowe (bdash)]

*** Bug 21707 has been marked as a duplicate of this bug. ***

[Mark Rowe (bdash)]

*** Bug 21734 has been marked as a duplicate of this bug. ***

[Mark Rowe (bdash)]

*** Bug 21730 has been marked as a duplicate of this bug. ***

[john]

It appears that this bug has reappeared in in r39433.  The $_FILES variable in php is not getting populated on form submit if no files are chosen for the file form elements.  Works fine in the current release of Safari and Firefox, but is broken in the nightly build.

[Darin Adler]

I understand that a bug was introduced by the new check-in, but I think that reopening this old bug report wasn't the best way to deal with that. Even though the symptom is the same.

[Darin Adler]

Nico, George, it looks like the refactoring broke file uploading.

[George Staikos]

Oops.  We could back out, or fix it later today.  Which do you prefer?

[Nikolas Zimmermann]

(In reply to comment #16)
> Oops.  We could back out, or fix it later today.  Which do you prefer?
> 

I'll investigate in the problem. We clearly need a layout test (!) - it bothers me stuff like that is not covered by layout tests.

[Darin Adler]

(In reply to comment #17)
> I'll investigate in the problem. We clearly need a layout test (!) - it bothers
> me stuff like that is not covered by layout tests.

Me too.

I'm hoping this is quick to fix -- we probably don't need to back out if we can figure it out today.

[Darin Adler]

Created attachment 26217 [details]
test

[Darin Adler]

Hmm, a single file upload to bugzilla just worked for me. So this new bug does not have the same scope as the old one. Could someone please file a new bug report instead of reopening this one, since the new bug seems quite different from the old?

[Nikolas Zimmermann]

(In reply to comment #20)
> Hmm, a single file upload to bugzilla just worked for me. So this new bug does
> not have the same scope as the old one. Could someone please file a new bug
> report instead of reopening this one, since the new bug seems quite different
> from the old?
> 

I think I found the bug, I'll file a new seperated one.

[Nikolas Zimmermann]

(In reply to comment #21)
> (In reply to comment #20)
> > Hmm, a single file upload to bugzilla just worked for me. So this new bug does
> > not have the same scope as the old one. Could someone please file a new bug
> > report instead of reopening this one, since the new bug seems quite different
> > from the old?
> > 
> 
> I think I found the bug, I'll file a new seperated one.

Bug 22971 will cover this new seperated issue. Let's move there.