[Webkit-unassigned] [Bug 123314] New: <select size> broken on safari, chrome, silk smartphone webkit browsers

bugzilla-daemon at webkit.org bugzilla-daemon at webkit.org
Thu Oct 24 20:13:51 PDT 2013


https://bugs.webkit.org/show_bug.cgi?id=123314

           Summary: <select size> broken on safari, chrome, silk
                    smartphone webkit browsers
           Product: WebKit
           Version: 528+ (Nightly build)
          Platform: Unspecified
               URL: http://www.w3.org/MarkUp/Test/HTML401/current/tests/17
                    _6-BF-01.html
        OS/Version: Unspecified
            Status: UNCONFIRMED
          Severity: Critical
          Priority: P2
         Component: New Bugs
        AssignedTo: webkit-unassigned at lists.webkit.org
        ReportedBy: dgoldman at ehdp.com


1. Version

I do not know the webkit version, so I initially assigned to the most recent version. But I know 100% the bug happens under the following browsers:

safari browser on ipod 4G
silk browser in kindle fire hd
chrome browser on galaxy player 4.2

There is a chance that the bug has been fixed in the most recent version of webkit. This can only be confirmed by the developers, I have no way of telling if the bug exists in the most recent version of webkit. If the bug has already been fixed, that would be great.

The bug does NOT occur in any webkit desktop browser, based on my testing with safari and chrome.

2. Component

I am filing this under "New Bugs". It would seem to fall under "Forms" or possibly "Layout and Rendering". The problem is specifically that <select size> does not work correctly under the smartphone and other non-desktop webkit implementations, that SIZE is ignored.

3. Platform / OS

This bug occurs on at least the following platforms and operating systems, so I could not make a single selection:

safari browser on ipod 4G
silk browser in kindle fire hd
chrome browser on galaxy player 4.2

The bug does NOT occur in any webkit desktop browser, based on my testing with safari and chrome.

4. Priority

I would suggest P2 priority, for the following reasons:

The bug is 100% reproducible
The bug causes any site using <size select> to not function in a non-trivial way. The bug causes the site using <size select> to lose so much functionality to be essentially not usable.

Because this is a bug means that webkit is NOT compliant with a basic HTML functionality, it could easily be considered as P1 priority.

5. Severity

For any web site that relies on <select size> the severity would be the highest Blocker level, because the bug destroys the functionality of the page, makes it unusable. The priority is certainly not less than Normal. I would think this would an urgent bug to fix, because it means that webkit is NOT compliant with a basic HTML functionality, so I assigned Critical.

6. URL

http://www.w3.org/MarkUp/Test/HTML401/current/tests/17_6-BF-01.html is the key URL to demonstrate the bug. This is a w3c page to "Verify the functionality of the size attribute of the SELECT element." The test is "Three options shall be visible at the same time since "size = 3" is set on the SELECT element.". The smartphone webkit browsers fail the test, so webkit is not compliant with a basic HTML functionality.

7. Summary

<select size> broken on safari, chrome, silk smartphone webkit browsers

8. Description

First, do I also need to post this bug to the chromium (or blink) group? Please let me know.

------------------ Basic description

There is a problem how webkit-based smartphone browsers display the HTML "select" tag when the "size" attribute is used. No matter what "size" is set to, the smartphone browser displays it as one line, and totally empty if no option selected. The problem was noticed on kindle fire (amazon silk), ipod (safari) and galaxy player (chrome), all of which are webkit-based smartphone-type browsers.

Here are screenshots of a w3schools example that shows the problem (vs correct display):

http://www.ehdp.com/out/silk-select-size-fails.png - silk fails badly.
http://www.ehdp.com/out/desktop-select-size-works.jpg - Any desktop browser passes.

Here are the w3schools pages with the <select size="3"> example:

http://www.w3schools.com/tags/att_select_size.asp
http://www.w3schools.com/tags/tryit.asp?fi...tml_select_size

There is nothing special about w3schools here, or the number "3". It is just a convenient example to test <select size> HTML tag.

It is of course needed to modify browser behavior in some cases to make the user experience better on smartphone. But this is not one of those cases. This is not just a cosmetic difference. This severely damages or makes basically unusable any website designed to use the "size" attribute. For example, try the following (the ones of concern to bug reporter) on desktop vs smartphone:

http://www.ehdp.com/vn/ro/txu1/pd1/dgw/en_US/index.htm - SELECT boxes messed up
http://www.ehdp.com/vn/ro/txu1/pd1/dgw/en_US/ajax.htm - change age or "race of woman"

Below are two desktop screenshots showing correct appearance and function of "select size" situations. The pages are totally destroyed, both appearance and function, when the webkit smartphone browser ignores the "size" attribute.

http://www.ehdp.com/vw-ajx/img02-bb1.htm
http://www.ehdp.com/vw-ajx/img03-bb1.htm

If the developer wants the select box to be three lines tall, why should the smartphone or other device overrule that choice? Why not let the developer, decide? There is no advantage of wiping out "size", there is no user benefit. If it was just a cosmetic difference, it would not matter. But it is a huge functional difference. It makes the select box display and work incorrectly, severely messes up the interface.

------- This DOES break a basic HTML functionality, really is a bug.

 http://www.w3.org/TR/html401/interact/forms.html#adef-size-SELECT 4.01 says:

"If a SELECT element is presented as a scrolled list box, this attribute specifies the number of rows in the list that should be visible at the same time. Visual user agents are not required to present a SELECT element as a list box; they may use any other mechanism, such as a drop-down menu." 

The HTML 4 "not required" text seems to permit SIZE to be ignored. Based on that some might argue that there is no bug, that <select size> is not broken on webkit browsers. However, this is not a sustainable argument, for three reasons:

1) "not required" almost certainly pertains to devices that do not allow a scrolled list box. So there seems no excuse for webkit not to honor the SIZE attribute. The text does not encourage SIZE being ignored, just allows it under certain cases. This is NOT one of those cases.

2) http://www.w3.org/MarkUp/Test/HTML401/current/tests/17_6-BF-01.html HTML 4 test case clearly says "Three options shall be visible at the same time since "size = 3" is set on the SELECT element". This clearly shows the intent that SIZE be honored by HTML 4. The webkit smartphone browser implementations fails the test, and thus are not compliant with HTML 4.

3) http://www.w3.org/TR/html5/forms.html#the-select-element HTML 5 standard just says "The size attribute gives the number of options to show to the user." and dispenses with the "not required" text. Therefore, the webkit smartphone browsers are not compliant with HTML 5.

------ If a "design decision", it is wrong decision, still breaks HTML, needs to be corrected

Was this a "design decision"? If not, then this is 100% a bug, and this section need to read no further.

1) If a "design decision", it might be argued that the purpose was to prevent a select element from using too much space. But this argument holds no water:

Just because something "could" happen, and something relatively minor (too much space), it doesn't logically follow to totally ignore SIZE attribute in smartphone, especially because the downside (wreck the interface usability) is so destructive. A developer uses <select size> for a special purpose, they have a reason, but the webkit smartphone browsers ignore the developer's need, destroy the functionality. The designers of HTML included <select size> for good reasons, but the webkit browser arbitrarily removes that functionality. Finally, any developer is aware that a smartphone has a smaller interface. If they choose, the developer could modify the interface for the smartphone (eg, use one line SELECT) if they want it to take less space, so there is no need to throw SIZE out.

2) Again, if a design decision, it might be argued that the basic functionality of a select element is to allow the user to choose one of the option elements, and that basic functionality is retained even is SIZE is ignored. But this argument also holds no water:

It is ridiculous to suggest that if the "basic functionality" of a tag is met, independent of the attributes, everything is fine. If that were true, what other attributes might webkit be comfortable with throwing out next? The following screenshot shows one example of many that could be cited (any analytical page using <select size>) where the webkit smartphone ignoring SIZE is destructive, basically makes the interface unusable. Furthermore, SELECT can do a lot more than "choose one". Take a close look below, you will see that it actually can let the user select a group of option elements, through programming the DOM, so there is more functionality to SIZE that many realize:

http://www.ehdp.com/vw-ajx/img02-bb1.htm

The bottom line is that, if a "design decision", it breaks HTML functionality, and destroys just about any real-world web page using <select size>, so it needs to be corrected. No decisions are set in stone, any decent organization reviews and corrects bad past decisions.

----- <select size> is OK non-webkit smartphone browsers, so webkit is "out of step"

The opera and firefox browsers were tested on smartphone devices, display select size perfectly, and provide the needed functionality, allow use of pages with <select size>. So far the webkit smartphone browsers are the "odd man out", the only ones to ignore SIZE, and thus destroy the page functionality. This makes the webkit smartphone browsers NOT compliant with HTML. Internet Explorer on the iphone has not been tested yet, but a message has been posted to a newsgroup asking someone to check the HTML 4 test page. Here is example of perfect performance with <select size> of opera on ipod:

http://www.ehdp.com/out/ipod-opera-select-is-perfect.png

9. Keywords

Possible keywords to consider, to be determined by whoever reviews this bug report:

AndroidHitList - The bug shows up in the google browser for galaxy player.

EasyFix - Since <size select> works perfectly on the webkit desktop browsers, it may be an easy fix to correct the error in the smartphone browser implementations.

GoogleBug - The bug shows up in the google browser for galaxy player.

InChromiumBugs - I do not know about submitting to the chromium bug database. Could you please advise?

NativeListBox - I have no idea exactly what this is, but it might be related to the SELECT tag, so I listed it in case relevant.

W3Ctest - http://www.w3.org/MarkUp/Test/HTML401/current/tests/17_6-BF-01.html is the w3c test to confirm the bug.

-- 
Configure bugmail: https://bugs.webkit.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.



More information about the webkit-unassigned mailing list