RESOLVED FIXED 10145
REGRESSION: Allow line breaks between adjacent popups and buttons 
https://bugs.webkit.org/show_bug.cgi?id=10145
Summary REGRESSION: Allow line breaks between adjacent popups and buttons
mitz
Reported 2006-07-28 08:03:17 PDT
Lines should be allowed to break between adjacent <select> elements, even if there is no white space between them. The test case shows that WebKit no longer allows it. Firefox does. I think the reason is that the native Mac popups are adjusted to white-space:pre.
Attachments
Test case (312 bytes, text/html)
2006-07-28 08:03 PDT, mitz 		no flags
Details
Test of other replaced elements (809 bytes, text/html)
2006-08-18 12:34 PDT, mitz 		no flags
Details
Test as rendered by WinIE (3.31 KB, image/png)
2006-08-18 12:35 PDT, mitz 		no flags
Details
Test as rendered by Firefox (4.59 KB, image/png)
2006-08-18 12:35 PDT, mitz 		no flags
Details
Patch that I think may be what we want (12.38 KB, patch)
2007-01-13 00:18 PST, Dave Hyatt 		aroben: review+
DetailsFormatted DiffDiff
Really get every case. (10.82 KB, patch)
2007-01-14 17:18 PST, Dave Hyatt 		oliver: review+
DetailsFormatted DiffDiff
Show Obsolete (5) View All Add attachment proposed patch, testcase, etc.
mitz
Comment 1 2006-07-28 08:03:47 PDT
Created attachment 9745 [details] Test case
mitz
Comment 2 2006-08-18 12:34:58 PDT
Created attachment 10129 [details] Test of other replaced elements
mitz
Comment 3 2006-08-18 12:35:36 PDT
Created attachment 10130 [details] Test as rendered by WinIE
mitz
Comment 4 2006-08-18 12:35:56 PDT
Created attachment 10131 [details] Test as rendered by Firefox
mitz
Comment 5 2006-08-18 12:36:55 PDT
As the extended test shows, WebKit has the same problem with <input type="button">.
Stephanie Lewis
Comment 6 2006-11-06 19:43:26 PST
radar 4823049
Dave Hyatt
Comment 7 2007-01-13 00:18:14 PST
Created attachment 12411 [details] Patch that I think may be what we want I'm a little hesitant about this, but it does seem to match Firefox. Basically we use the parent white-space when looking at replaced elements. This seems to be what Firefox is doing.
Darin Adler
Comment 8 2007-01-13 07:27:11 PST
Comment on attachment 12411 [details] Patch that I think may be what we want You attached the wrong patch. This patch is removing the marquee unfurl style!
Dave Hyatt
Comment 9 2007-01-13 13:12:55 PST
It's not the wrong patch. I had to remove the whiteSpace() flag from marquees since with this patch it's no longer needed. While in there I decided to yank the unfurl thing, since it was just for Safari RSS prototyping and never used. The important bit is in bidi.cpp near the bottom.
Dave Hyatt
Comment 10 2007-01-13 15:37:22 PST
Comment on attachment 12411 [details] Patch that I think may be what we want Setting review flag (timidly).
Adam Roben (:aroben)
Comment 11 2007-01-13 23:14:37 PST
Comment on attachment 12411 [details] Patch that I think may be what we want Turn the testcase into a layout test, and r=me.
Dave Hyatt
Comment 12 2007-01-14 02:00:27 PST
I landed this patch, but as Mitz pointed out, there's still the problem of replaced elements followed by non-replaced elements.
Dave Hyatt
Comment 13 2007-01-14 02:04:05 PST
NEver mind, it's ok. I'm going to add an additional test.
mitz
Comment 14 2007-01-14 06:31:00 PST
This case is still misbehaving: <div style="width: 5px"><select></select><select></select><select></select></div>
Dave Hyatt
Comment 15 2007-01-14 17:18:55 PST
Created attachment 12436 [details] Really get every case. Here we go.
Dave Hyatt
Comment 16 2007-01-14 17:45:48 PST
Fixed in r18851.
Note You need to log in before you can comment on or make changes to this bug.