Open Bug 191603 Opened 22 years ago Updated 4 years ago

When clicking in a form's textarea/input, the text inside that area is copied to the copy buffer

Categories

(Core :: DOM: Editor, defect, P5)

x86
Linux
defect

Tracking

()

People

(Reporter: woolite, Unassigned)

References

()

Details

(Keywords: dataloss)

Attachments

(1 file)

User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.2.1) Gecko/20021226 Debian/1.2.1-9 Build Identifier: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.2.1) Gecko/20021226 Debian/1.2.1-9 When clicking in a form's textarea or input field, the string: javascript:this.form.elements['name'].select(); will scribble over your copy buffer if you are using Linux. I have verified this in both WindowMaker and KDE, so I don't think it's a window manager problem. Reproducible: Always Steps to Reproduce: Steps to reproduce: 1) Visit http://www.darkpact.com/test4.html 2) Click on the textarea's scrollbar 3) Middle-click paste into another window to see what is in your copy buffer. If everything went according to my dark, devious plans, it should be the text that is featured inside of the textarea! 4) You can also reproduce this bug by clicking on the text in the textarea, or on the input field to the right of the text area. Actual Results: The copy buffer is filled with the text from the field. Expected Results: NOTHING! :) I really don't like my copy buffer being smashed by Javascript... Especially when scrolling though a textarea with scrollbars! This bug is probably only applicable to operating systems that have a highlight-copy mechanism.
I can't remember seeing that before. Do you have a live site sample? I can think of times i would find it useful, though.
I can't show you anything that is live, but I do know one project that uses it. PHPMyAdmin (http://phpmyadmin.sourceforge.net/) The onfocus highlight/copy bug is REALLY annoying when trying to middle-click paste into a text area. When you middle click paste, it will activate the 'onfocus' before performing the paste, so it will smash your copy buffer and repaste the text from the textarea.
Here is a textual example (so you don't have to veiw the source on the website: <form> <textarea name="test" onfocus="javascript:this.form.elements['test'].select();"> Text that will be copied. </textarea> </form>
Attached file testcase
kin, I thought you fixed this in bug 188440? Looks like you only got the cases when _we_ select the text, not the ones where the _website_ does.... (marking dataloss, since this wipes out the clipboard unexpectedly).
Assignee: rogerl → jfrancis
Severity: normal → critical
Status: UNCONFIRMED → NEW
Component: JavaScript Engine → Editor: Core
Ever confirmed: true
Keywords: dataloss
QA Contact: pschwartau → sujay
To kin. kin, could you please check this out?
Assignee: jfrancis → kin
bz, I think this is a dup of bug 188440. The build date used by the Anoakie is 2002-12-26, and I didn't fix bug 188440 until 2003-01-21. Anoakie, can you try a more a recent build? I did verify, in my TRUNK build, that the select() call in the test case triggers SelectAllContents() which was part of the fix for 188440, so it does look like a dup. *** This bug has been marked as a duplicate of 188440 ***
Status: NEW → RESOLVED
Closed: 22 years ago
Resolution: --- → DUPLICATE
kin, give me _some_ credit. 1) 2002-12-26 is before 2003-01-08, which is when bug 88049 was checked in, causing bug 188440 2) The problem with the testcase in this bug is quite present in linux trunk build 2003-02-01-22. Perhaps I should have clarified that this is a problem with builds with the fix for bug 188440 in them. So this problem is distinct from that of bug 188440 (and predates it). But it's certainly related; I'm surprised we're using a different codepatch in this case.... (Do we perhaps call SelectAllContents() and then _also_ call some other selection function?) Do you have access to a Linux build to debug with usefully, or should I try to see who's asserting PRIMARY?
Status: RESOLVED → REOPENED
Resolution: DUPLICATE → ---
I just gave 20030130 a whirl (mozilla-snapshot in debian unstable), and it exhibits the same problems as my 20021226 build. I also download the latest nightly (20030203), and it too shows this problem. Not as severe as the 20021226 build (scrolling with the scrollbars no longer smashes the copy buffer), but when you click in the text area, it will still smash the copy buffer. I think you should give this bug another try with the latest build before marking it as 'RESOLVED'. And I got a midair collision... It seems that someone else has reopened it. Thanks :)
Sorry for the pre-mature dup'ing bz. It looks like the problem is due to the fact that the AutoCopy code is looking for a selection change that is the result of a button up event. The problem with this test case is that the order of events happen like this: 1. Button Down (Selection is collapsed at point clicked) 2. A focus event is synthesized and processed synchronously (while post processing the Button Down event) which triggers the JS select() call in the testcase, causing all of the content in the widget to be selected. 3. The user releases the button and the auto copy code is triggered. Note, I am actually testing this on Win32 since I don't have a Linux build currently, but I imagine the same set of events happen on Linux too.
Status: REOPENED → ASSIGNED
Events aren't really the problem. Everything appears to be happening in the proper order, it's just that autocopy isn't checking whether the selection that was created is the result of a user input or not. As we talked about, I'd suggest having the place where mousemoves (and arrow keys) trigger selection set a boolean noting that the selection is the result of user interaction.
QA Contact: sujay → editor
Assignee: kinmoz → nobody
Status: ASSIGNED → NEW

Bulk-downgrade of unassigned, >=5 years untouched DOM/Storage bugs' priority.

If you have reason to believe this is wrong (especially for the severity), please write a comment and ni :jstutte.

Severity: critical → S4
Priority: -- → P5
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Creator:
Created:
Updated:
Size: