slab / quill Goto Github PK
View Code? Open in Web Editor NEWQuill is a modern WYSIWYG editor built for compatibility and extensibility
Home Page: https://quilljs.com
License: BSD 3-Clause "New" or "Revised" License
Quill is a modern WYSIWYG editor built for compatibility and extensibility
Home Page: https://quilljs.com
License: BSD 3-Clause "New" or "Revised" License
To reproduce:
Expected behavior:
Actual behavior:
If the following lines do not begin with a tab, Scribe behaves as expected. However, if leading tabs are subsequently deleted, so that you only have lines with no leading tabs, the error behavior described persists.
Steps to reproduce in Firefox 17.0:
Note that after backspacing out, no text entry is allowed.
In Chrome: The pasted text assumes the font 'serif', but should remain 'san serif'.
In Firefox: Nothing is pasted.
The {color: 'red'} attribute is lost on the second "cd". Occurs in Chrome; Firefox is fine. I've added a test for this in the delete suite, which should help prevent regressing any more (it's been discovered/fixed before; see issue #12.
Expected: The cursor is at index 1, between "a" and "c"
Actual: The cursor is at index 2, after "ac"
Note that you must hit delete; backspace works fine.
Works fine in Chrome.
If editor's style is customized, the cursor container is off (showing cursors in the wrong place)
Chrome 26.
Expected: "\n"
Actual: "\n\n\n"
Hit this on beta.
Adding newlines at the end from user input should scroll the editor to keep the cursor in view
This isn't really a bug in Stypi but it's a bug when integrated into Salesforce Chatter. The _ is being overridden somehow by certain Chatter pages and we need to guard it from this.
Expected: Cursor is at index 1.
Actual: Cursor is at index 0
Version: IE9.
If you type multiple lines, then this is not an issue. Added a test to cursor.rb (see commit 2c69070) to prevent regression.
Ubuntu 12.04 LTS / Firefox 19.0.2
"debugger" break-point in fuzzer.rb prior to fuzzer logic.
To reproduce:
Expected behavior: Fixnum (value: 6)
Actual behavior: String (value: "")
To check values, simply remove ".class" from step 4
Start a new note, type "hello", then hit backspace 5 times. After that, hit backspace again and you'll see a revision was created, this is presumably because the \n is being deleted with the 6th backspace. After that, no syncing happens when you type.
To reproduce:
Expected behavior: Fonts stay monospaced.
Actual behavior: The first line reverts to default Sans Serif, second line stays monospaced.
Note that a tab is inserted instead of the focus being set on the previous element.
Browser: Chrome 26
Steps to reproduce:
Expected Result
"ef" with the previously applied background color.
Actual Result
"ef" with no background color.
Ubuntu 12.04 LTS / Firefox 19.0.2
"debugger" break-point in fuzzer.rb prior to fuzzer logic.
To reproduce:
Expected behavior: "abcdef" in the editor
Actual behavior: nothing in the editor
Additional notes:
In the Firefox window, the address bar has initial focus.
At step 2, if you CLICK back into the terminal instead of using alt-tab the behavior is as expected: "abcdef" prints in the editor. HOWEVER, Firefox steals focus the first time adapter.type_text is called, requiring you to alt-tab/click back into terminal.
At step 2, if you click the editor, and then alt-tab/click into the terminal the behavior is as expected: "abcdef" prints in the editor
Typing a character into Scribe on Firefox 20 yields the following error:
NotFoundError: Node was not found
.
The referenced line is in scribe.all.js, so it's probably easiest for you to just check it out yourself. The error is reproducible on both beta.stypi.com and the head of the Scribe repo.
Let me know if you need any further info.
Expected: "abc" gets indented in two tab widths
Actual: The editor loses focus after the first tab, and the URL bar gains it.
Expected: 'links', all hyperlinked.
Actual: 'link' is hyperlinked, 's' is not.
All other formatting types have the behavior that appending maintains the formatting; I don't think that links should be any different, but I could be overlooking a good reason for this?
On beta, to reproduce:
Expected result: "abc\ndef"
Actual result: "efabc\ndef"
Expected: "\t1\n\t2\n3"
Actual: "\t1\n\t2\n\t3"
Expected: Cursor moves down one line after every enter, ending on the third line.
Actual: Cursor stays in place.
If you click into the editor after the typing, the cursor will jump to the correct position. Also, editor.getDelta() does yield the correct delta. Unfortunately this makes it pretty tough to automate this in webdriver.
Expected: The bold button is not activated, and clicking it applies bold to 'ab'.
Actual: The bold button is activated, and clicking it does nothing.
The hot keys work fine. As does highlighting in the other direction, i.e., highlighting 'a' then 'b'.
When Scribe first loads, the hyperlinking UI flickers in the editor. It seems like it needs to be display: none'd sooner or something.
Issue has altered since fix in focus issue
Follow steps 1-7 below, the updated behavior is as follows:
Expected behavior: "abcdefZZZ"
Actual behavior: "aZZZbcdef"
Expected behavior: "abcdefZZZqq"
Actual behavior "aZZZbcdefqq"
Ubuntu 12.04 LTS / Firefox 19.0.2
"debugger" break-point in fuzzer.rb prior to fuzzer logic.
To reproduce:
Expected behavior: "abcdefZZZ"
Actual behavior: "aZZZcdef"
Expeced behavior: "abcdefZZZqq"
Actual behavior: "aZZZqqdef"
Additional notes:
At step 2, if you click the editor, and then alt-tab/click into the terminal, behavior is as expected.
specifically, if you use APPLE KEY + (direction) to get to the beginning / end of the document or line the cursor doesn't update, presumably because you are not listening to these keystroke combos, or whatever.
User must explicitly click in for cursor to be detected
Instead of rendering Scribe inside it's iframe, IE instead displays a page reading "Internet Explorer cannot display the webpage."
Occurs in IE 9 & 10.
This issue IS NOT isolated to firefox.
"debugger" break-point in fuzzer.rb prior to fuzzer logic.
attr_accessor :cursor_pos in selenium_adapter.rb
To reproduce:
Expected behavior at step x:
5. cursor_pos should be 0 because it is an empty document
7. cursor_pos should be 6
Actual behavior at step x:
5. cursor_pos is 10
7. cursor_pos is 16
doc_length correctly outputs 6.
Browser: Firefox 19.0
Steps to reproduce:
Expected Result:
"123\n"
Actual Result:
The lines "1\n", "23"
The tail is always pushed to the next line. This issue does not occur in Chrome.
Once it has stabilized, we should document all the options that the Scribe constructor accepts.
Expected: "a"
Actual: "\na"
See commit 418e0d5 for the accompanying test case.
On the advanced demo:
This does not happen in the index demo or if you type out Link, set the format and then insert the link. This indicates that the author module is probably the culprit.
Expected: "c\n" is highligted
Actual: "c\nde" is highlighted.
Internet Explorer 9, 10.
After about 35 characters the title field will drop to the next line and disappear from the view.
The color will remain red. This is the case for all default options. I think it's related to the value=false change, see fffa6dd
Copying "\nHey" only pastes "Hey"
This seems to only apply when modifying a mid-string char; i.e., the 'b' in 'abc'.
The dropdown should read 'San Serif', but still reads 'Serif'. editor.getOps() indicates that the delta has been updated correctly though.
The same issue occurs with all other dropdowns when switching back to a default style; for example, when switching back to the "Normal" size or the "White" background color.
Ubuntu 12.04 LTS / Firefox 19.0.2
"debugger" break-point in fuzzer.rb prior to fuzzer logic.
attr_accessor :cursor_pos in selenium_adapter.rb
To reproduce:
Expected behavior at step x:
5. doc_length should be 6
6. cursor_pos should be 6
8. "ZZZabcdef"
Actual behavior at step x:
5. doc_length is 0
6. cursor_pos is 0
7. "aZZZbcdef"
The error: "IndexSizeError: Index or size is negative or greater than the allowed amount" is raised.
If you move the cursor around after hitting enter, it does seem that the newline was actually added. However, if you instead type any further characters they won't appear.
I think a necessary precondition is that we start with the empty document. I can't reproduce this in demo/index.html, but it does occur in demo/pretty.html and also in fuzzer.html.
If the document is long enough that the scroll bar is introduced, and if you are scrolled to the bottom, pasting text anywhere within the current view seems to automatically reset the scroll position to 0, at the very beginning of the doc.
I experienced this when pasting the same bullet points in a previous bug. But this time I only selected a subset of the line of text so as not to include any line delimiters and this caused the resulting paste to be broken into multiple lines.
This seems to be the case when the font is the only attribute, however. If you have {font: monospace, bold: true}, for example, then paste works as expected.
Expected: ab is bold, c is not
Acceptable: abc are not bold
Actual: a is bold, bc are not
Occurs in Firefox, not Chrome, who knows about IE
Currently set at 100% min height thus padding guarantees scrollbar
The dropdown's value should switch to 'Normal' but it remains 'Huge'. The editor's delta is correct though.
This is more than just a cosmetic issue, because if you were to highlight '123' with the intention of making it 'Huge', you wouldn't be able to (it's technically already selected and selecting it does nothing). To force the matter, you currently have to first set the size to something other than 'Huge' or 'Normal', and then switch back to what you actually want.
Expected: "abc" is highlighted.
Actual: "b" is highlighted.
The behavior in Firefox and IE is that it highlights a character, advances the cursor; highlights a character, advances the cursor... making it impossible to highlight more than 1 char at a time.
If I paste bullet points from outlook for example, pasting does nothing, the result is like pasting blank characters.
If user specifies http, leave as http. We force a protocol to prevent xss. Perhaps a better algorithm is keep either http or https, otherwise, strip the protocol?
A declarative, efficient, and flexible JavaScript library for building user interfaces.
๐ Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
The Web framework for perfectionists with deadlines.
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. ๐๐๐
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
We are working to build community through open source technology. NB: members must have two-factor auth.
Open source projects and samples from Microsoft.
Google โค๏ธ Open Source for everyone.
Alibaba Open Source for everyone
Data-Driven Documents codes.
China tencent open source team.