I have a page in XPages that I use to open and edit a document. There are two ways to open a document in edit-mode: first in read-mode then click a button to put it in edit-mode, or open it directly in edit-mode. Both work in all browsers, yet IE seems to handle both cases differently. We found this out when working with the SWING API.
Opening directly in edit-mode in IE (8/9/10) works, via read-mode to edit-mode doesn't. What we found is that the internal representation of a textarea field differs: when opened in edit-mode, there are more properties, but most importantly, the return+linefeeds are correctly set in both the value and the innerText property.
The button just contains a simple Change Mode action.
Has anyone heard of this anomaly? And does someone know what we did wrong?
PS I'll try to build a simple XPage that shows this behaviour more clearly tomorrow.
For IE switching from read to edit mode, you need a full page refresh
Related
UPDATE: At time of asking this question, this was related to SignalR library and not plain WebSockets. I see correctly formatted messages now.
Is there any way to word-wrap messages in WS tab in Chrome Developer Tools or display JSON with formatting ? It's really annoying to scroll to right to see whole message.
Example with message selected and it's preview doesn't have any formatting or word wrapping applied:
Thank you in advance.
It's working fine here on Chrome/78.0.3904.97:
What I did:
Go to http://crawl.develz.org/play.htm
Open one of the listed servers
Start devtools
Go to the Application tab and add a cookie called "no-compression" with value "yeah no" to the relevant server. (Any truthy string should work, I just chose the least confusing one I could think of in about a minute.)
Otherwise, crawl's webtiles server can end up compressing messages even when browser supports RFC 7692's "permessage-deflate" extension, which ruins the demonstration.
Open the Network tab
Reload the page
Select the "socket" request, switch to the "Messages" tab, and pick a frame.
Start drilling down in the tree view in the bottom pane!
I'm trying to make a bookmarklet that will set an SVG filter over the entire page it is loaded into. To that end, it sets a style on the root element:
document.children[0].style.filter='url(data:image/svg+xml;base64,PHN2ZyB4…#h)';
This works well in Firefox, but Chrome seems to think that it is unsafe cross-site content, drops the filter, and outputs a warning to the console:
(source: codl.fr)
Is this expected behaviour? I would like to be sure before I file a bug. If so, how else can I acheve this?
Here is a test page containing the bookmarklet: http://f.codl.fr/1402/hacker/
I have a problem with a project that I have been working on for a while now. When I open an entry, the inputs all have data in them. However when I go to actually print the entry (using the print-mode, CTRL+P), I notice that the select, and textarea fields are missing the selected or input text on paper. What's going on here? I have googled for hours and can't come up with anything relevant... help! Thanks!
Link to Project (Click "Demo" to create a session - required)
Link to Entry (Click this after you click the 1st link)
Update : Before now I have just been using Google Chrome to test printing. However, when I tried FireFox, it printed alright, is there anything special I need to do to get Chrome to work??...
RESOLUTION : I am using jQuery and I found that by removing the ui-corner-all class from the inputs before printing, I could then print. Thanks anyways!
If you're using Internet Explorer, then this discussion thread may help you (summary: IE can get confused when printing malformed docs). Check your doctype tag matches your output (the w3c markup validator is very useful)
Edit - I see you're using chrome, but I suggest you still check your page is well formed :)
RESOLUTION: I'm using jQuery and I found that by removing the ui-corner-all class from the inputs before printing, I could then print. Thanks anyways!
This is something of a follow-up question to my question here. You can find the HTML source in a text file here.
When I load that page in IE8, I get the "Done, but with errors on page." message in my status bar. The detail view shows
Expected identifier
sms Line: 147
Code: 0 Char: 67
and I see absolutely no problems anywhere near there. In IE8, the page is still behaving erratically w/r/t the randomly losing focus as mentioned in my other question.
When I load the same exact page in Firefox (using Firebug) the console shows no errors and the page works perfectly. Any thoughts on what's going on here? This is driving me nuts and making me want to give up on even trying to write an IE friendly page.
Edit: Thanks for all the comments! This page is written as a JSP, so I edit in Eclipse. I found an Eclipse warning about the onblur event for the username field. I switched it from
onblur="alert(document.activeElement + ' class:' + document.activeElement.class)"
to
onblur="alert(document.activeElement)"
and that made the bizarre IE page error vanish. I had been trying to give more info (namely, its CSS class) about specifically which element is stealing focus - to my own detriment, apparently, since Javascript was interpreting the '.class' part in the Java(script) sense there is no class property (I should have been using className).
And, no, the page doesn't validate. But the errors were mostly/all ones that just didn't make sense to me, such as
Line 14, Column 41: Attribute "LANGUAGE" is not a valid attribute. Did you mean "language"?
But I'm still stuck trying to figure out why, as I enter text in the username & password fields, focus randomly switches to a div (working on figuring out which div currently).
Edit 2: It's the div between the two "global nav" comments, at the very top of the body. Still no idea why it's happening, though.
The problem appears to be the line
onBlur="alert(document.activeElement + ' class:' + document.activeElement.class)"
when you take off the ".class" it loads without issue. Are you sure ".class" is valid?
Does your HTML validate?
I agree, IE does a terrible job giving developers information about page errors. I only support IE on the principle that users shouldn't have to download twenty-odd browsers to go to their favorite websites. Web developers have a responsibility to make it "just work". Browser developers have a responsibility to communicate with each other and conform to standards.
Run your javascript through JSLint. You probably have a missing semicolon somewhere or a comma at the end of an object property that shouldn't be there. Firefox is more forgiving than IE when it comes to those types of syntax errors.
Edit: The inline js in your page seems to be OK. Check the contents of qm_scripts.js.
When I remove the .class from document.activeElement.class the error goes away.
The real issue is that you need to be able to debug your JS in IE 8, correct? There is a tool for that! :)
I am experience an issue using sIFR 3, jQuery BlockUI and an UpdatePanel where the page refreshes and the SIFR font disappears.
In the Javascript file I use to handle the blocking and unblocking of the UI, I detect the partial postback and unblock the UI:
prm.add_endRequest(function() {
$.unblockUI(); });
If I re-run sIFR.replace() in this code block the font sIFR font is rendered, but # characters are replacing my page title (in the top of the browser window/tab). This seems to only happen in IE (go figure.) Each time I refresh the page using the LinkButton an extra # character is added, so my page title ends up looking like "#########". Any idea why this is happening?
Is there a specific sIFR.update() function I should be running instead? If I execute sIFR.redraw() or sIFR() I get javascript run time errors.
This is caused by a bug in IE:
This is a very peculiar bug in IE where for each Flash movie in the page the fragment identifier part of the URI (#foo) is repeated in the document title. It has been reported by Macromedia that this bug can be fixed by updating the VBScript engine the browser uses. However, for those users who still have the wrong engine sIFR contains a workaround.
http://wiki.novemberborn.net/sifr/Fragment+Identifier+Bug
sIFR 3 has a workaround but it doesn't kick in unless the page is loaded with a fragment identifier (window.location.hash != ""). In your case, I assume the page is loaded without the identifier, then you add one, and then you re-run replacements, reinvoking the bug.
You can patch r436 by opening (the compressed version of) sifr.js and replacing this.fix=c.ua.ieWin&&window.location.hash!="" by this.fix=c.ua.ieWin.