(In reply to neil@parkwaycc.co.uk from comment #16) > > Can nsHTMLEditor::GetInlineProperty be called from a mochitest? > We do call it from Composer UI (via EditorGetTextProperty), if that helps. Yes, I saw this in comm-central/editor/ui/composer/content/editorUtilities.js: GetCurrentEditor().getInlineProperty(propAtom, attribute, value, firstHas, anyHas, allHas); I've written one new Mochitest in my life (and modified a few), so I hope I can work something out with the help of Ehsan. I've got Firefox compiled on my travel laptop now. I'll look into it over the next few days.

In order to write a test for this, you need to ask help from someone who understand the comm-central side of how this functionality is called from there, I'm not the right person to answer questions about that code.

Are you saying this can't be tested using only Firefox in, say, a <div contenteditable>?

The code in question should very well be testable from a normal content editable div using document.queryCommandValue("fontname")

(In reply to Ehsan Akhgari from comment #20) > The code in question should very well be testable from a normal content > editable div using document.queryCommandValue("fontname") That's no use, this is equivalent to the function I'm using in my workaround.

OK, I'd like to summarise what we know so far and see where we go from here. We are interested in the getInlineProperty function, which is implemented here: https://dxr.mozilla.org/mozilla-central/source/editor/libeditor/nsHTMLEditorStyle.cpp#1100 For collapsed selection we find the code starting here: https://dxr.mozilla.org/mozilla-central/source/editor/libeditor/nsHTMLEditorStyle.cpp#1128 For non-collapsed selection, it starts here: https://dxr.mozilla.org/mozilla-central/source/editor/libeditor/nsHTMLEditorStyle.cpp#1180 We note that at this line in the non-collapsed case https://dxr.mozilla.org/mozilla-central/source/editor/libeditor/nsHTMLEditorStyle.cpp#1237 a value is assigned to *aValue. This is missing from the collapsed case, so Neil's patch adds this in right after this line: https://dxr.mozilla.org/mozilla-central/source/editor/libeditor/nsHTMLEditorStyle.cpp#1148 Now the question is how to test this. Obviously we want to call getInlineProperty in some JS Mochitest. We do call getInlineProperty in Thunderbird client code like this: GetCurrentEditor().getInlineProperty(...) so that makes me think that there is a JS interface to it. I tried to start writing a Mochitest taking Ehsan's editor/libeditor/tests/test_bug1100966.html as a starting point. I'll attach the attempt, which fails like follows: line 35: TypeError: editor.getInlineProperty is not a function Sadly, I have very poor JS skills, but a Google search showed, that is this may have to do with the function parameters. Obviously getInlineProperty doesn't take ("font", "face", value, firstHas, anyHas, allHas), but the first argument is an nsIAtom. So how do I pass that in? BTW, I found a description of the function here: http://doxygen.db48x.net/mozilla/html/interfacensIHTMLEditor.html The third argument is an "in", but in light of what I said above, for non-collapsed selections, the third argument can also return something. So am I on the right track here? Any help is welcome.

I'm not sure whether this should go into mochitest.ini or chrome.ini. I put it into chrome.ini, this is where Ehsan's test_bug1100966.html can be found.

(In reply to Jorg K (at the beach until 22nd April) from comment #22) > TypeError: editor.getInlineProperty is not a function In theory it's possible that the object you are working with is not an nsIHTMLEditor (getInlineProperty is defined on line 114 of nsIHTMLEditor.idl). However it's more likely that the object is an nsIHTMLEditor, but XPConnect doesn't know it yet. GetCurrentEditor(), used by EditorGetTextProperty(), uses editor instanceof Components.interfaces.nsIHTMLEditor to notify XPConnect but QueryInterface() should also work. (The difference between the two is that QueryInterface() returns the editor and throws on failure but instanceof returns a boolean flag.) > Obviously getInlineProperty doesn't take ("font", "face", value, firstHas, > anyHas, allHas), but the first argument is an nsIAtom. So how do I pass that > in? You need to get an atom from the atom service. (EditorGetTextProperty() does this too.)

> You need to get an atom from the atom service. (EditorGetTextProperty() does > this too.) Yes, I saw that, Thunderbird does this: gAtomService = Components.classes["@mozilla.org/atom-service;1"].getService(Components.interfaces.nsIAtomService); var propAtom = gAtomService.getAtom(property); GetCurrentEditor().getInlineProperty(propAtom, attribute, value, firstHas, anyHas, allHas); So I changed my test to: var editor = getEditor(); var gAtomService = Components.classes["@mozilla.org/atom-service;1"].getService(Components.interfaces.nsIAtomService); var propAtom = gAtomService.getAtom("font"); editor.getInlineProperty(propAtom, "face", value, firstHas, anyHas, allHas); Which fails with the same error. Let's see what Ehsan says.

As Neil pointed out, this is the definition: https://dxr.mozilla.org/mozilla-central/source/editor/nsIHTMLEditor.idl#114 void getInlineProperty(in nsIAtom aProperty, in AString aAttribute, in AString aValue, out boolean aFirst, out boolean aAny, out boolean aAll); So it's questionable whether or not something should be returned on the "in" parameter aValue. However, the behaviour is inconsistent. As detailed on comment #22, sometimes something is returned, other times not.

Like Neil said, you need to QueryInterface() the editor object for Components.interfaces.nsIHTMLEditor before calling its interface function.

I'm taking the liberty to supersede all previous patches. @Ehsan: Please review and keep in mind that I am *not* a JS programmer. I did a cut & paste job. The test fails, when the code change is removed, as getInlineProperty then always returned true on the three out parameters. Please ignore (as you already have) the nonsense I wrote about passing back something on the 3rd "in" argument. I was confused, but then I began to see the light. Try run here: https://treeherder.mozilla.org/#/jobs?repo=try&revision=018f3fb1f578

Comment on attachment 8592329 [details] [diff] [review] Fix by Neil and test by Jorg K >+ var div = document.getElementById("content"); >+ div.focus(); >+ synthesizeMouseAtCenter(div, {}); >+ synthesizeKey("VK_LEFT", {}); >+ >+ var sel = window.getSelection(); >+ var selRange = sel.getRangeAt(0); >+ is(selRange.endContainer.nodeName, "#text", "selection should be in text node"); >+ is(selRange.endOffset, 9, "offset should be 9"); [Rather than navigating to a particular selection, you could just explicitly select the node of interest.] >+ editor.getInlineProperty(propAtom, "face", "Arial", firstHas, anyHas, allHas); >+ is(firstHas.value, true, "true expected"); >+ is(anyHas.value, true, "true expected"); >+ is(allHas.value, true, "true expected"); >+ editor.getInlineProperty(propAtom, "face", "Courier", firstHas, anyHas, allHas); >+ is(firstHas.value, false, "false expected"); >+ is(anyHas.value, false, "false expected"); >+ is(allHas.value, false, "false expected"); [These messages are not very meaningful, I'm not the best person to ask here but something like "didn't expect allHas Courier" perhaps?]

(In reply to neil@parkwaycc.co.uk from comment #29) Thanks for the comments. > [Rather than navigating to a particular selection, you could just explicitly > select the node of interest.] As I said, copy/paste. Doing a mouse click gives a collapsed selected, going left by one wasn't necessary. > [These messages are not very meaningful, I'm not the best person to ask here > but something like "didn't expect allHas Courier" perhaps?] Correct. I'll fix them with whatever else Ehsan complains about. I'm not even sure this should be a "chrome" test as opposed to a "plain" test. Anyway, we'll get the bug squashed real soon now.

Corrections according to review and suggestions by Neil. Sorry, I'm not sure whether to supersede the previous patch, since it would lose the review+. Please do what needs to be done upon review of the new version.

Comment on attachment 8592656 [details] [diff] [review] Fix by Neil and test by Jorg K - take 2 Review of attachment 8592656 [details] [diff] [review]: ----------------------------------------------------------------- FYI, r+ means "I trust you to make the necessary changes, so I don't need to review them again unless you decide to make substantial changes to the patch."

OK, but where do we go from here? Should I r+ the new patch myself and supersede the old? Who does the check-in? In the past you marked it for "check-in needed" or so. I'm happy to learn ;-)

You just need to mark the bug as checkin-needed. No need to manually toggle the Bugzilla flags!

OK, to whomever does the check-in: Please check in attachment 8592656 [details] [diff] [review], NOT attachment 8592329 [details] [diff] [review].

Comment on attachment 8592656 [details] [diff] [review] Fix by Neil and test by Jorg K - take 2 Carrying forward ehsan's r+

@Aryeh: Can you please help me with this: M2 fails on this run https://treeherder.mozilla.org/#/jobs?repo=try&revision=018f3fb1f578 and also another run for bug 1154791 https://treeherder.mozilla.org/#/jobs?repo=try&revision=6e599d277f32 where I accidentally submitted the patch from this bug here again. The confusing failure is: 1345 INFO TEST-UNEXPECTED-FAIL | dom/imptests/editing/conformancetest/test_runtest.html | [["inserttext","a"]] "foo<font color=brown><a href=http://www.google.com>[bar]</a></font>baz" compare innerHTML - [["inserttext","a"]] "foo<font color=brown><a href=http://www.google.com>[bar]</a></font>baz" compare innerHTML: assert_equals: Unexpected innerHTML (after normalizing inline style) expected "foo<font color=\"brown\"><a href=\"http://www.google.com\">a</a></font>baz" but got Looks like it timed out. I can't see what the changes from this bug or the other bug 1154791 could have to do with it.

Thank you for the explanation. I also need to be looking for GetInlineProperty or GetInlinePropertyWithAttrValue, right?. There are a few more calls: http://mxr.mozilla.org/mozilla-central/search?string=GetInlineProperty&find=&findi=&filter=^[^\0]*%24&hitlimit=&tree=mozilla-central Only the calls are interesting, where the third argument is not null or not empty, since only in the case were a "real" third parameter is passed, the behaviour has changed. Looking at the calls there is only one call which passes a third argument, it can be found here: http://mxr.mozilla.org/mozilla-central/source/editor/libeditor/nsHTMLEditRules.cpp#7562 Not a surprise, since you mentioning 'typing an "a"'. There doesn't seem to be anything wrong with this call, it passes what was previously retrieved: &(mCachedStyles[i].value However, the retrieval at http://mxr.mozilla.org/mozilla-central/source/editor/libeditor/nsHTMLEditRules.cpp#7496 may not be correct, at least it's no symmetric to the setting, ie. it uses IsTextPropertySetByContent. It may be a little hard to debug, since it only fails in the conformance test. I forgot to mention: How has the behaviour of GetInlineProperty changed: Well, when a third argument is passed, it now really checks whether the property really has the value. Before it always returned "true". In other words, for example: Before GetInlineProperty ("font", "face", "Arial", first, any, all) and GetInlineProperty ("font", "face", "Courier", first, any, all) *always* returned true, now true is returned if the value is really present. So in case of the failed test, it will ask for "font", "color", "something", and the answer will be false, so the "brown" is reapplied, as we saw. Most likely it doesn't ask for "brown", or some code further down is broken that now returns false despite the "brown" being present. Any hints/help/etc. are welcome ;-)

(In reply to Jorg K (at the beach until 22nd April) from comment #42) > It may be a little hard to debug, since it only fails in the conformance test. Actually, I could test it in a <div contenteditable> foo<font color=brown><a href=http://www.google.com>bar</a></font>baz <div> by typing on top of the "bar" or in the Midas demo, where I can also check the HTML generated.

Nope, that ain't working. If I select the "bar" and type over it, the whole link goes. So any suggestions on how to debug this further?

OK, here some findings. For my test I cut down to conformance test to the one case that fails, so: dom/imptests/editing/conformancetest/data.js var browserTests = [ ["foo<font color=brown><a href=http://www.google.com>[bar]</a></font>baz", [["inserttext","a"]], "foo<font color=\"brown\"><a href=\"http://www.google.com\">a[]</a></font>baz", [true], {"inserttext":[false,false,"",false,false,""]}] ] Then I added debug to the function I mentioned: nsHTMLEditRules.cpp: nsresult nsHTMLEditRules::CacheInlineStyles(nsIDOMNode *aNode) ... mCachedStyles[j].mPresent = true; mCachedStyles[j].value.Assign(outValue); printf ("\n\n\nfound style |%s|\n", NS_ConvertUTF16toUTF8(outValue).get()); and nsresult nsHTMLEditRules::ReapplyCachedStyles() ... nsresult res = mHTMLEditor->GetInlinePropertyBase(mCachedStyles[i].tag, &(mCachedStyles[i].attr), &(mCachedStyles[i].value), &bFirst, &bAny, &bAll, &curValue, false); NS_ENSURE_SUCCESS(res, res); printf ("\n\n\nchecked style |%s|, result %d", NS_ConvertUTF16toUTF8(mCachedStyles[i].value).get(), bAny); Sadly, this is the result from the two print statements: found style |brown| checked style |brown|, result 0 So, GetInlineProperty does actually not work. It never got noticed, since before, when the selection was collapsed, it also didn't work but always returned "true", which coincidentally is the correct answer here, but not in other cases. So, what a nice can of worms :-(( @Ehsan, if you find the time, your comment would be appreciated. Maybe just read from comment #40 if this comment alone doesn't tell you everything.

Printing the "atom", the attribute and the value gives this: found style |10415456|color|brown| found style |10415456|color|brown|, result 0

How about marking this test as known failure and be done with it? It's not anything that the user can do, select the text part of a link only and type over it (see comment #43 and comment #44). In another bug 1141017 we removed 114 known failures, so here we would add one.

OK, here is what happens: Instead of looking up brown, we look up red and don't find it: nsHTMLEditRules::CacheInlineStyles - found style |color|brown| nsHTMLEditRules::ReapplyCachedStyles - looking up |color|brown| nsHTMLEditor::GetInlinePropertyBase (before IsCSSEquivalentToHTMLInlineStyleSet) |color|brown| nsHTMLEditor::GetInlinePropertyBase (after IsCSSEquivalentToHTMLInlineStyleSet) |color|rgb(0, 0, 255)|, isSet 0 nsHTMLEditRules::ReapplyCachedStyles - looked up |color|brown|, result 0 Without the fix, the "if (aValue) {tOutString.Assign(*aValue);", the debug looks like this: nsHTMLEditRules::CacheInlineStyles - found style |color|brown| nsHTMLEditRules::ReapplyCachedStyles - looking up |color|brown| nsHTMLEditor::GetInlinePropertyBase (before IsCSSEquivalentToHTMLInlineStyleSet) |color|| <-- no value here nsHTMLEditor::GetInlinePropertyBase (after IsCSSEquivalentToHTMLInlineStyleSet) |color|rgb(0, 0, 255)|, isSet 1 nsHTMLEditRules::ReapplyCachedStyles - looked up |color|brown|, result 1 Not so jolly good that 1 is always returned regardless from IsCSSEquivalentToHTMLInlineStyleSet if no value is passed in. Basically, the error that was corrected discovered an error elsewhere. So Ehsan and Aryeh: I need further instructions.

Further analysis in/of IsCSSEquivalentToHTMLInlineStyleSet: If no value is passed in, it always returns true, jolly good: http://mxr.mozilla.org/mozilla-central/source/editor/libeditor/nsHTMLCSSUtils.cpp#1114 Then, it retrieves the colour from the element, it returns (0,0,255) - which is BLUE!! (not red as stated in comment #48). Yes, the link is BLUE, it's not brown despite the inline style saying so. Here is some more debug: nsHTMLEditRules::CacheInlineStyles - found style |color|brown| nsHTMLEditRules::ReapplyCachedStyles - looking up |color|brown| nsHTMLEditor::GetInlinePropertyBase (before IsCSSEquivalentToHTMLInlineStyleSet) |color|brown| Before GetCSSEquivalentToHTMLInlineStyleSet value |brown| After GetCSSEquivalentToHTMLInlineStyleSet value |rgb(0, 0, 255)| <--- brown is turned to blue comparing (1) |rgb(165, 42, 42)| to |rgb(0, 0, 255)| <--- the RGB equivalent of brown is compared to blue nsHTMLEditor::GetInlinePropertyBase (after IsCSSEquivalentToHTMLInlineStyleSet) |color|rgb(0, 0, 255)|, result 0 nsHTMLEditRules::ReapplyCachedStyles - looked up |color|brown|, result 0 So there is some logic error here. We expect brown to be returned from an element that is essentially BLUE!! This HTML: foo<font color=brown><a href=http://www.google.com>bar</a></font>baz produces "bar" in BLUE, not brown. Anyway, I would mark the test as "expected fail", since it's pretty much nonsense, and it can't happen in real life. No user can select the text alone and type over it. Any selection will always select the whole link. The only fix could be to check nsHTMLEditRules::CacheInlineStyles and not pick up "brown" for something which is blue. If blue were picked up and cached to start with, it would all work well.

Pursuing the last idea a little further, therefore some more debug. nsHTMLEditRules::CacheInlineStyles - found style via IsTextPropertySetByContent |color|brown| nsHTMLEditRules::CacheInlineStyles - found style |color|brown| nsHTMLEditRules::ReapplyCachedStyles - looking up |color|brown| nsHTMLEditor::GetInlinePropertyBase (before IsCSSEquivalentToHTMLInlineStyleSet) |color|brown| Before GetCSSEquivalentToHTMLInlineStyleSet value |brown| After GetCSSEquivalentToHTMLInlineStyleSet value |rgb(0, 0, 255)| comparing (1) |rgb(165, 42, 42)| to |rgb(0, 0, 255)| nsHTMLEditor::GetInlinePropertyBase (after IsCSSEquivalentToHTMLInlineStyleSet) |color|rgb(0, 0, 255)|, result 0 nsHTMLEditRules::ReapplyCachedStyles - looked up |color|brown|, result 0 In nsHTMLEditRules::CacheInlineStyles we're apparently in the branch which retrieves via IsTextPropertySetByContent: http://mxr.mozilla.org/mozilla-central/source/editor/libeditor/nsHTMLEditRules.cpp#7499 So in other words: the supposed colour gets picked up via IsTextPropertySetByContent which is than compared to the real colour retrieved by GetCSSEquivalentToHTMLInlineStyleSet. The only option I see is to make this change: Before: if (!useCSS || (mCachedStyles[j].tag == nsGkAtoms::font && mCachedStyles[j].attr.EqualsLiteral("size"))) { New: if ((mCachedStyles[j].tag == nsGkAtoms::font && mCachedStyles[j].attr.EqualsLiteral("size"))) { In this case the debug looks like this and the test succeeds: nsHTMLEditRules::CacheInlineStyles - found style via IsCSSEquivalentToHTMLInlineStyleSet ||| nsHTMLEditRules::CacheInlineStyles - found style ||underline| nsHTMLEditRules::CacheInlineStyles - found style via IsCSSEquivalentToHTMLInlineStyleSet |face|| nsHTMLEditRules::CacheInlineStyles - found style |face|serif| nsHTMLEditRules::CacheInlineStyles - found style via IsCSSEquivalentToHTMLInlineStyleSet |color|| nsHTMLEditRules::CacheInlineStyles - found style |color|rgb(0, 0, 255)| nsHTMLEditRules::ReapplyCachedStyles - looking up ||underline| nsHTMLEditor::GetInlinePropertyBase (before IsCSSEquivalentToHTMLInlineStyleSet) ||underline| nsHTMLEditor::GetInlinePropertyBase (after IsCSSEquivalentToHTMLInlineStyleSet) ||underline|, result 1 nsHTMLEditRules::ReapplyCachedStyles - looked up ||underline|, result 1 nsHTMLEditRules::ReapplyCachedStyles - looking up |face|serif| nsHTMLEditor::GetInlinePropertyBase (before IsCSSEquivalentToHTMLInlineStyleSet) |face|serif| nsHTMLEditor::GetInlinePropertyBase (after IsCSSEquivalentToHTMLInlineStyleSet) |face|serif|, result 1 nsHTMLEditRules::ReapplyCachedStyles - looked up |face|serif|, result 1 nsHTMLEditRules::ReapplyCachedStyles - looking up |color|rgb(0, 0, 255)| nsHTMLEditor::GetInlinePropertyBase (before IsCSSEquivalentToHTMLInlineStyleSet) |color|rgb(0, 0, 255)| comparing (2) |rgb(0, 0, 255)| to |rgb(0, 0, 255)| nsHTMLEditor::GetInlinePropertyBase (after IsCSSEquivalentToHTMLInlineStyleSet) |color|rgb(0, 0, 255)|, result 1 nsHTMLEditRules::ReapplyCachedStyles - looked up |color|rgb(0, 0, 255)|, result 1

OK, trying the previous patch with the additional change in editor/libeditor/nsHTMLEditRules.cpp - if (!useCSS || (mCachedStyles[j].tag == nsGkAtoms::font && - mCachedStyles[j].attr.EqualsLiteral("size"))) { + if (mCachedStyles[j].tag == nsGkAtoms::font && + mCachedStyles[j].attr.EqualsLiteral("size")) { This might be a true disaster: https://treeherder.mozilla.org/#/jobs?repo=try&revision=4edfeeead3b9

Oops, src/editor/libeditor/nsHTMLEditRules.cpp:7489:8: error: unused variable 'useCSS' [-Werror=unused-variable] - compiled on Windows ;-) Trying again with less tests: https://treeherder.mozilla.org/#/jobs?repo=try&revision=ea0fa5712f63 Questions: 1) What is this used for? bool useCSS = mHTMLEditor->IsCSSEnabled(); Why not always use CSS? 2) How to print the string content of a nsIAtom? Some hours later: Always using CSS produced 66, failures in the conformance test. So, let's try something else: Only always use CSS for font color. I guess that this will also give some errors, a subset of the 66.

Try tree is closed, so I'm attaching for feedback instead. For the record, here the 66 failures of the previous try push: Line 1: 1345 INFO TEST-UNEXPECTED-FAIL | dom/imptests/editing/conformancetest/test_runtest.html | [["fontname","sans-serif"],["delete",""]] "foo[]bar" queryCommandValue("fontname") after - [["fontname","sans-serif"],["delete",""]] "foo[]bar" queryCommandValue("fontname") after: assert_equals: Wrong result returned expected "sans-serif" but got "serif" Line 10: 1347 INFO TEST-UNEXPECTED-FAIL | dom/imptests/editing/conformancetest/test_runtest.html | [["fontname","sans-serif"],["delete",""],["inserttext","a"]] "foo[]bar" compare innerHTML - [["fontname","sans-serif"],["delete",""],["inserttext","a"]] "foo[]bar" compare innerHTML: assert_equals: Unexpected innerHTML (after normalizing inline style) expected "fo<font face=\"sans-serif\">a</font>bar" but got "foabar" Line 19: 1349 INFO TEST-UNEXPECTED-FAIL | dom/imptests/editing/conformancetest/test_runtest.html | [["fontname","sans-serif"],["delete",""],["inserttext","a"]] "foo[]bar" queryCommandValue("fontname") after - [["fontname","sans-serif"],["delete",""],["inserttext","a"]] "foo[]bar" queryCommandValue("fontname") after: assert_equals: Wrong result returned expected "sans-serif" but got "serif" Line 28: 1351 INFO TEST-UNEXPECTED-FAIL | dom/imptests/editing/conformancetest/test_runtest.html | [["fontname","sans-serif"],["formatblock","<div>"]] "foo[]bar" queryCommandValue("fontname") after - [["fontname","sans-serif"],["formatblock","<div>"]] "foo[]bar" queryCommandValue("fontname") after: assert_equals: Wrong result returned expected "sans-serif" but got "serif" Line 37: 1353 INFO TEST-UNEXPECTED-FAIL | dom/imptests/editing/conformancetest/test_runtest.html | [["fontname","sans-serif"],["formatblock","<div>"],["inserttext","a"]] "foo[]bar" compare innerHTML - [["fontname","sans-serif"],["formatblock","<div>"],["inserttext","a"]] "foo[]bar" compare innerHTML: assert_equals: Unexpected innerHTML (after normalizing inline style) expected "<div>foo<font face=\"sans-serif\">a</font>bar</div>" but got "<div>fooabar</div>" Line 46: 1355 INFO TEST-UNEXPECTED-FAIL | dom/imptests/editing/conformancetest/test_runtest.html | [["fontname","sans-serif"],["formatblock","<div>"],["inserttext","a"]] "foo[]bar" queryCommandValue("fontname") after - [["fontname","sans-serif"],["formatblock","<div>"],["inserttext","a"]] "foo[]bar" queryCommandValue("fontname") after: assert_equals: Wrong result returned expected "sans-serif" but got "serif" Line 55: 1357 INFO TEST-UNEXPECTED-FAIL | dom/imptests/editing/conformancetest/test_runtest.html | [["fontname","sans-serif"],["forwarddelete",""]] "foo[]bar" queryCommandValue("fontname") after - [["fontname","sans-serif"],["forwarddelete",""]] "foo[]bar" queryCommandValue("fontname") after: assert_equals: Wrong result returned expected "sans-serif" but got "serif" Line 64: 1359 INFO TEST-UNEXPECTED-FAIL | dom/imptests/editing/conformancetest/test_runtest.html | [["fontname","sans-serif"],["forwarddelete",""],["inserttext","a"]] "foo[]bar" compare innerHTML - [["fontname","sans-serif"],["forwarddelete",""],["inserttext","a"]] "foo[]bar" compare innerHTML: assert_equals: Unexpected innerHTML (after normalizing inline style) expected "foo<font face=\"sans-serif\">a</font>ar" but got "fooaar" Line 73: 1361 INFO TEST-UNEXPECTED-FAIL | dom/imptests/editing/conformancetest/test_runtest.html | [["fontname","sans-serif"],["forwarddelete",""],["inserttext","a"]] "foo[]bar" queryCommandValue("fontname") after - [["fontname","sans-serif"],["forwarddelete",""],["inserttext","a"]] "foo[]bar" queryCommandValue("fontname") after: assert_equals: Wrong result returned expected "sans-serif" but got "serif" Line 82: 1363 INFO TEST-UNEXPECTED-FAIL | dom/imptests/editing/conformancetest/test_runtest.html | [["fontname","sans-serif"],["indent",""]] "foo[]bar" queryCommandValue("fontname") after - [["fontname","sans-serif"],["indent",""]] "foo[]bar" queryCommandValue("fontname") after: assert_equals: Wrong result returned expected "sans-serif" but got "serif" Line 91: 1365 INFO TEST-UNEXPECTED-FAIL | dom/imptests/editing/conformancetest/test_runtest.html | [["fontname","sans-serif"],["indent",""],["inserttext","a"]] "foo[]bar" compare innerHTML - [["fontname","sans-serif"],["indent",""],["inserttext","a"]] "foo[]bar" compare innerHTML: assert_equals: Unexpected innerHTML (after normalizing inline style) expected "<blockquote>foo<font face=\"sans-serif\">a</font>bar</blockquote>" but got "<blockquote>fooabar</blockquote>" Line 100: 1367 INFO TEST-UNEXPECTED-FAIL | dom/imptests/editing/conformancetest/test_runtest.html | [["fontname","sans-serif"],["indent",""],["inserttext","a"]] "foo[]bar" queryCommandValue("fontname") after - [["fontname","sans-serif"],["indent",""],["inserttext","a"]] "foo[]bar" queryCommandValue("fontname") after: assert_equals: Wrong result returned expected "sans-serif" but got "serif" Line 109: 1369 INFO TEST-UNEXPECTED-FAIL | dom/imptests/editing/conformancetest/test_runtest.html | [["fontname","sans-serif"],["inserthorizontalrule",""]] "foo[]bar" queryCommandValue("fontname") after - [["fontname","sans-serif"],["inserthorizontalrule",""]] "foo[]bar" queryCommandValue("fontname") after: assert_equals: Wrong result returned expected "sans-serif" but got "serif" Line 118: 1371 INFO TEST-UNEXPECTED-FAIL | dom/imptests/editing/conformancetest/test_runtest.html | [["fontname","sans-serif"],["inserthorizontalrule",""],["inserttext","a"]] "foo[]bar" compare innerHTML - [["fontname","sans-serif"],["inserthorizontalrule",""],["inserttext","a"]] "foo[]bar" compare innerHTML: assert_equals: Unexpected innerHTML (after normalizing inline style) expected "foo<hr><font face=\"sans-serif\">a</font>bar" but got "foo<hr>abar" Line 127: 1373 INFO TEST-UNEXPECTED-FAIL | dom/imptests/editing/conformancetest/test_runtest.html | [["fontname","sans-serif"],["inserthorizontalrule",""],["inserttext","a"]] "foo[]bar" queryCommandValue("fontname") after - [["fontname","sans-serif"],["inserthorizontalrule",""],["inserttext","a"]] "foo[]bar" queryCommandValue("fontname") after: assert_equals: Wrong result returned expected "sans-serif" but got "serif" Line 136: 1375 INFO TEST-UNEXPECTED-FAIL | dom/imptests/editing/conformancetest/test_runtest.html | [["fontname","sans-serif"],["insertimage","/img/lion.svg"]] "foo[]bar" queryCommandValue("fontname") after - [["fontname","sans-serif"],["insertimage","/img/lion.svg"]] "foo[]bar" queryCommandValue("fontname") after: assert_equals: Wrong result returned expected "sans-serif" but got "serif" Line 145: 1377 INFO TEST-UNEXPECTED-FAIL | dom/imptests/editing/conformancetest/test_runtest.html | [["fontname","sans-serif"],["insertimage","/img/lion.svg"],["inserttext","a"]] "foo[]bar" compare innerHTML - [["fontname","sans-serif"],["insertimage","/img/lion.svg"],["inserttext","a"]] "foo[]bar" compare innerHTML: assert_equals: Unexpected innerHTML (after normalizing inline style) expected "foo<img src=\"/img/lion.svg\"><font face=\"sans-serif\">a</font>bar" but got "foo<img src=\ Line 154: 1379 INFO TEST-UNEXPECTED-FAIL | dom/imptests/editing/conformancetest/test_runtest.html | [["fontname","sans-serif"],["insertimage","/img/lion.svg"],["inserttext","a"]] "foo[]bar" queryCommandValue("fontname") after - [["fontname","sans-serif"],["insertimage","/img/lion.svg"],["inserttext","a"]] "foo[]bar" queryCommandValue("fontname") after: assert_equals: Wrong result returned expected "sans-serif" but got "serif" Line 163: 1381 INFO TEST-UNEXPECTED-FAIL | dom/imptests/editing/conformancetest/test_runtest.html | [["fontname","sans-serif"],["insertorderedlist",""]] "foo[]bar" queryCommandValue("fontname") after - [["fontname","sans-serif"],["insertorderedlist",""]] "foo[]bar" queryCommandValue("fontname") after: assert_equals: Wrong result returned expected "sans-serif" but got "serif" Line 172: 1383 INFO TEST-UNEXPECTED-FAIL | dom/imptests/editing/conformancetest/test_runtest.html | [["fontname","sans-serif"],["insertorderedlist",""],["inserttext","a"]] "foo[]bar" compare innerHTML - [["fontname","sans-serif"],["insertorderedlist",""],["inserttext","a"]] "foo[]bar" compare innerHTML: assert_equals: Unexpected innerHTML (after normalizing inline style) expected "<ol><li>foo<font face=\"sans-serif\">a</font>bar</li></ol>" but got "<ol><li>fooabar</li></ol>" Line 181: 1385 INFO TEST-UNEXPECTED-FAIL | dom/imptests/editing/conformancetest/test_runtest.html | [["fontname","sans-serif"],["insertorderedlist",""],["inserttext","a"]] "foo[]bar" queryCommandValue("fontname") after - [["fontname","sans-serif"],["insertorderedlist",""],["inserttext","a"]] "foo[]bar" queryCommandValue("fontname") after: assert_equals: Wrong result returned expected "sans-serif" but got "serif" Line 190: 1387 INFO TEST-UNEXPECTED-FAIL | dom/imptests/editing/conformancetest/test_runtest.html | [["fontname","sans-serif"],["insertparagraph",""]] "foo[]bar" queryCommandValue("fontname") after - [["fontname","sans-serif"],["insertparagraph",""]] "foo[]bar" queryCommandValue("fontname") after: assert_equals: Wrong result returned expected "sans-serif" but got "serif" Line 199: 1389 INFO TEST-UNEXPECTED-FAIL | dom/imptests/editing/conformancetest/test_runtest.html | [["fontname","sans-serif"],["insertparagraph",""],["inserttext","a"]] "foo[]bar" queryCommandValue("fontname") after - [["fontname","sans-serif"],["insertparagraph",""],["inserttext","a"]] "foo[]bar" queryCommandValue("fontname") after: assert_equals: Wrong result returned expected "sans-serif" but got "serif" Line 208: 1391 INFO TEST-UNEXPECTED-FAIL | dom/imptests/editing/conformancetest/test_runtest.html | [["fontname","sans-serif"],["insertunorderedlist",""]] "foo[]bar" queryCommandValue("fontname") after - [["fontname","sans-serif"],["insertunorderedlist",""]] "foo[]bar" queryCommandValue("fontname") after: assert_equals: Wrong result returned expected "sans-serif" but got "serif" Line 217: 1393 INFO TEST-UNEXPECTED-FAIL | dom/imptests/editing/conformancetest/test_runtest.html | [["fontname","sans-serif"],["insertunorderedlist",""],["inserttext","a"]] "foo[]bar" compare innerHTML - [["fontname","sans-serif"],["insertunorderedlist",""],["inserttext","a"]] "foo[]bar" compare innerHTML: assert_equals: Unexpected innerHTML (after normalizing inline style) expected "<ul><li>foo<font face=\"sans-serif\">a</font>bar</li></ul>" but got "<ul><li>fooabar</li></ul>" Line 226: 1395 INFO TEST-UNEXPECTED-FAIL | dom/imptests/editing/conformancetest/test_runtest.html | [["fontname","sans-serif"],["insertunorderedlist",""],["inserttext","a"]] "foo[]bar" queryCommandValue("fontname") after - [["fontname","sans-serif"],["insertunorderedlist",""],["inserttext","a"]] "foo[]bar" queryCommandValue("fontname") after: assert_equals: Wrong result returned expected "sans-serif" but got "serif" Line 235: 1397 INFO TEST-UNEXPECTED-FAIL | dom/imptests/editing/conformancetest/test_runtest.html | [["fontname","sans-serif"],["justifycenter",""]] "foo[]bar" queryCommandValue("fontname") after - [["fontname","sans-serif"],["justifycenter",""]] "foo[]bar" queryCommandValue("fontname") after: assert_equals: Wrong result returned expected "sans-serif" but got "serif" Line 244: 1399 INFO TEST-UNEXPECTED-FAIL | dom/imptests/editing/conformancetest/test_runtest.html | [["fontname","sans-serif"],["justifycenter",""],["inserttext","a"]] "foo[]bar" queryCommandValue("fontname") after - [["fontname","sans-serif"],["justifycenter",""],["inserttext","a"]] "foo[]bar" queryCommandValue("fontname") after: assert_equals: Wrong result returned expected "sans-serif" but got "serif" Line 253: 1401 INFO TEST-UNEXPECTED-FAIL | dom/imptests/editing/conformancetest/test_runtest.html | [["fontname","sans-serif"],["justifyfull",""]] "foo[]bar" queryCommandValue("fontname") after - [["fontname","sans-serif"],["justifyfull",""]] "foo[]bar" queryCommandValue("fontname") after: assert_equals: Wrong result returned expected "sans-serif" but got "serif" Line 262: 1403 INFO TEST-UNEXPECTED-FAIL | dom/imptests/editing/conformancetest/test_runtest.html | [["fontname","sans-serif"],["justifyfull",""],["inserttext","a"]] "foo[]bar" queryCommandValue("fontname") after - [["fontname","sans-serif"],["justifyfull",""],["inserttext","a"]] "foo[]bar" queryCommandValue("fontname") after: assert_equals: Wrong result returned expected "sans-serif" but got "serif" Line 271: 1405 INFO TEST-UNEXPECTED-FAIL | dom/imptests/editing/conformancetest/test_runtest.html | [["fontname","sans-serif"],["justifyleft",""]] "foo[]bar" queryCommandValue("fontname") after - [["fontname","sans-serif"],["justifyleft",""]] "foo[]bar" queryCommandValue("fontname") after: assert_equals: Wrong result returned expected "sans-serif" but got "serif" Line 280: 1407 INFO TEST-UNEXPECTED-FAIL | dom/imptests/editing/conformancetest/test_runtest.html | [["fontname","sans-serif"],["justifyleft",""],["inserttext","a"]] "foo[]bar" queryCommandValue("fontname") after - [["fontname","sans-serif"],["justifyleft",""],["inserttext","a"]] "foo[]bar" queryCommandValue("fontname") after: assert_equals: Wrong result returned expected "sans-serif" but got "serif" Line 289: 1409 INFO TEST-UNEXPECTED-FAIL | dom/imptests/editing/conformancetest/test_runtest.html | [["fontname","sans-serif"],["justifyright",""]] "foo[]bar" queryCommandValue("fontname") after - [["fontname","sans-serif"],["justifyright",""]] "foo[]bar" queryCommandValue("fontname") after: assert_equals: Wrong result returned expected "sans-serif" but got "serif" Line 298: 1411 INFO TEST-UNEXPECTED-FAIL | dom/imptests/editing/conformancetest/test_runtest.html | [["fontname","sans-serif"],["justifyright",""],["inserttext","a"]] "foo[]bar" queryCommandValue("fontname") after - [["fontname","sans-serif"],["justifyright",""],["inserttext","a"]] "foo[]bar" queryCommandValue("fontname") after: assert_equals: Wrong result returned expected "sans-serif" but got "serif" Line 307: 1427 INFO TEST-UNEXPECTED-FAIL | dom/imptests/editing/conformancetest/test_runtest.html | [["forecolor","#0000FF"],["delete",""]] "foo[]bar" queryCommandValue("forecolor") after - [["forecolor","#0000FF"],["delete",""]] "foo[]bar" queryCommandValue("forecolor") after: assert_equals: Wrong result returned (after color normalization) expected "rgb(0, 0, 255)" but got "rgb(0, 0, 0)" Line 316: 1429 INFO TEST-UNEXPECTED-FAIL | dom/imptests/editing/conformancetest/test_runtest.html | [["forecolor","#0000FF"],["delete",""],["inserttext","a"]] "foo[]bar" queryCommandValue("forecolor") after - [["forecolor","#0000FF"],["delete",""],["inserttext","a"]] "foo[]bar" queryCommandValue("forecolor") after: assert_equals: Wrong result returned (after color normalization) expected "rgb(0, 0, 255)" but got "rgb(0, 0, 0)" Line 325: 1431 INFO TEST-UNEXPECTED-FAIL | dom/imptests/editing/conformancetest/test_runtest.html | [["forecolor","#0000FF"],["formatblock","<div>"]] "foo[]bar" queryCommandValue("forecolor") after - [["forecolor","#0000FF"],["formatblock","<div>"]] "foo[]bar" queryCommandValue("forecolor") after: assert_equals: Wrong result returned (after color normalization) expected "rgb(0, 0, 255)" but got "rgb(0, 0, 0)" Line 334: 1433 INFO TEST-UNEXPECTED-FAIL | dom/imptests/editing/conformancetest/test_runtest.html | [["forecolor","#0000FF"],["formatblock","<div>"],["inserttext","a"]] "foo[]bar" queryCommandValue("forecolor") after - [["forecolor","#0000FF"],["formatblock","<div>"],["inserttext","a"]] "foo[]bar" queryCommandValue("forecolor") after: assert_equals: Wrong result returned (after color normalization) expected "rgb(0, 0, 255)" but got "rgb(0, 0, 0)" Line 343: 1435 INFO TEST-UNEXPECTED-FAIL | dom/imptests/editing/conformancetest/test_runtest.html | [["forecolor","#0000FF"],["forwarddelete",""]] "foo[]bar" queryCommandValue("forecolor") after - [["forecolor","#0000FF"],["forwarddelete",""]] "foo[]bar" queryCommandValue("forecolor") after: assert_equals: Wrong result returned (after color normalization) expected "rgb(0, 0, 255)" but got "rgb(0, 0, 0)" Line 352: 1437 INFO TEST-UNEXPECTED-FAIL | dom/imptests/editing/conformancetest/test_runtest.html | [["forecolor","#0000FF"],["forwarddelete",""],["inserttext","a"]] "foo[]bar" queryCommandValue("forecolor") after - [["forecolor","#0000FF"],["forwarddelete",""],["inserttext","a"]] "foo[]bar" queryCommandValue("forecolor") after: assert_equals: Wrong result returned (after color normalization) expected "rgb(0, 0, 255)" but got "rgb(0, 0, 0)" Line 361: 1439 INFO TEST-UNEXPECTED-FAIL | dom/imptests/editing/conformancetest/test_runtest.html | [["forecolor","#0000FF"],["indent",""]] "foo[]bar" queryCommandValue("forecolor") after - [["forecolor","#0000FF"],["indent",""]] "foo[]bar" queryCommandValue("forecolor") after: assert_equals: Wrong result returned (after color normalization) expected "rgb(0, 0, 255)" but got "rgb(0, 0, 0)" Line 370: 1441 INFO TEST-UNEXPECTED-FAIL | dom/imptests/editing/conformancetest/test_runtest.html | [["forecolor","#0000FF"],["indent",""],["inserttext","a"]] "foo[]bar" queryCommandValue("forecolor") after - [["forecolor","#0000FF"],["indent",""],["inserttext","a"]] "foo[]bar" queryCommandValue("forecolor") after: assert_equals: Wrong result returned (after color normalization) expected "rgb(0, 0, 255)" but got "rgb(0, 0, 0)" Line 379: 1443 INFO TEST-UNEXPECTED-FAIL | dom/imptests/editing/conformancetest/test_runtest.html | [["forecolor","#0000FF"],["inserthorizontalrule",""]] "foo[]bar" queryCommandValue("forecolor") after - [["forecolor","#0000FF"],["inserthorizontalrule",""]] "foo[]bar" queryCommandValue("forecolor") after: assert_equals: Wrong result returned (after color normalization) expected "rgb(0, 0, 255)" but got "rgb(0, 0, 0)" Line 388: 1445 INFO TEST-UNEXPECTED-FAIL | dom/imptests/editing/conformancetest/test_runtest.html | [["forecolor","#0000FF"],["inserthorizontalrule",""],["inserttext","a"]] "foo[]bar" queryCommandValue("forecolor") after - [["forecolor","#0000FF"],["inserthorizontalrule",""],["inserttext","a"]] "foo[]bar" queryCommandValue("forecolor") after: assert_equals: Wrong result returned (after color normalization) expected "rgb(0, 0, 255)" but got "rgb(0, 0, 0)" Line 397: 1447 INFO TEST-UNEXPECTED-FAIL | dom/imptests/editing/conformancetest/test_runtest.html | [["forecolor","#0000FF"],["insertimage","/img/lion.svg"]] "foo[]bar" queryCommandValue("forecolor") after - [["forecolor","#0000FF"],["insertimage","/img/lion.svg"]] "foo[]bar" queryCommandValue("forecolor") after: assert_equals: Wrong result returned (after color normalization) expected "rgb(0, 0, 255)" but got "rgb(0, 0, 0)" Line 406: 1449 INFO TEST-UNEXPECTED-FAIL | dom/imptests/editing/conformancetest/test_runtest.html | [["forecolor","#0000FF"],["insertimage","/img/lion.svg"],["inserttext","a"]] "foo[]bar" queryCommandValue("forecolor") after - [["forecolor","#0000FF"],["insertimage","/img/lion.svg"],["inserttext","a"]] "foo[]bar" queryCommandValue("forecolor") after: assert_equals: Wrong result returned (after color normalization) expected "rgb(0, 0, 255)" but got "rgb(0, 0, 0)" Line 415: 1451 INFO TEST-UNEXPECTED-FAIL | dom/imptests/editing/conformancetest/test_runtest.html | [["forecolor","#0000FF"],["insertorderedlist",""]] "foo[]bar" queryCommandValue("forecolor") after - [["forecolor","#0000FF"],["insertorderedlist",""]] "foo[]bar" queryCommandValue("forecolor") after: assert_equals: Wrong result returned (after color normalization) expected "rgb(0, 0, 255)" but got "rgb(0, 0, 0)" Line 424: 1453 INFO TEST-UNEXPECTED-FAIL | dom/imptests/editing/conformancetest/test_runtest.html | [["forecolor","#0000FF"],["insertorderedlist",""],["inserttext","a"]] "foo[]bar" queryCommandValue("forecolor") after - [["forecolor","#0000FF"],["insertorderedlist",""],["inserttext","a"]] "foo[]bar" queryCommandValue("forecolor") after: assert_equals: Wrong result returned (after color normalization) expected "rgb(0, 0, 255)" but got "rgb(0, 0, 0)" Line 433: 1455 INFO TEST-UNEXPECTED-FAIL | dom/imptests/editing/conformancetest/test_runtest.html | [["forecolor","#0000FF"],["insertparagraph",""]] "foo[]bar" queryCommandValue("forecolor") after - [["forecolor","#0000FF"],["insertparagraph",""]] "foo[]bar" queryCommandValue("forecolor") after: assert_equals: Wrong result returned (after color normalization) expected "rgb(0, 0, 255)" but got "rgb(0, 0, 0)" Line 442: 1457 INFO TEST-UNEXPECTED-FAIL | dom/imptests/editing/conformancetest/test_runtest.html | [["forecolor","#0000FF"],["insertparagraph",""],["inserttext","a"]] "foo[]bar" queryCommandValue("forecolor") after - [["forecolor","#0000FF"],["insertparagraph",""],["inserttext","a"]] "foo[]bar" queryCommandValue("forecolor") after: assert_equals: Wrong result returned (after color normalization) expected "rgb(0, 0, 255)" but got "rgb(0, 0, 0)" Line 451: 1459 INFO TEST-UNEXPECTED-FAIL | dom/imptests/editing/conformancetest/test_runtest.html | [["forecolor","#0000FF"],["insertunorderedlist",""]] "foo[]bar" queryCommandValue("forecolor") after - [["forecolor","#0000FF"],["insertunorderedlist",""]] "foo[]bar" queryCommandValue("forecolor") after: assert_equals: Wrong result returned (after color normalization) expected "rgb(0, 0, 255)" but got "rgb(0, 0, 0)" Line 460: 1461 INFO TEST-UNEXPECTED-FAIL | dom/imptests/editing/conformancetest/test_runtest.html | [["forecolor","#0000FF"],["insertunorderedlist",""],["inserttext","a"]] "foo[]bar" queryCommandValue("forecolor") after - [["forecolor","#0000FF"],["insertunorderedlist",""],["inserttext","a"]] "foo[]bar" queryCommandValue("forecolor") after: assert_equals: Wrong result returned (after color normalization) expected "rgb(0, 0, 255)" but got "rgb(0, 0, 0)" Line 469: 1463 INFO TEST-UNEXPECTED-FAIL | dom/imptests/editing/conformancetest/test_runtest.html | [["forecolor","#0000FF"],["justifycenter",""]] "foo[]bar" queryCommandValue("forecolor") after - [["forecolor","#0000FF"],["justifycenter",""]] "foo[]bar" queryCommandValue("forecolor") after: assert_equals: Wrong result returned (after color normalization) expected "rgb(0, 0, 255)" but got "rgb(0, 0, 0)" Line 478: 1465 INFO TEST-UNEXPECTED-FAIL | dom/imptests/editing/conformancetest/test_runtest.html | [["forecolor","#0000FF"],["justifycenter",""],["inserttext","a"]] "foo[]bar" queryCommandValue("forecolor") after - [["forecolor","#0000FF"],["justifycenter",""],["inserttext","a"]] "foo[]bar" queryCommandValue("forecolor") after: assert_equals: Wrong result returned (after color normalization) expected "rgb(0, 0, 255)" but got "rgb(0, 0, 0)" Line 487: 1467 INFO TEST-UNEXPECTED-FAIL | dom/imptests/editing/conformancetest/test_runtest.html | [["forecolor","#0000FF"],["justifyfull",""]] "foo[]bar" queryCommandValue("forecolor") after - [["forecolor","#0000FF"],["justifyfull",""]] "foo[]bar" queryCommandValue("forecolor") after: assert_equals: Wrong result returned (after color normalization) expected "rgb(0, 0, 255)" but got "rgb(0, 0, 0)" Line 496: 1469 INFO TEST-UNEXPECTED-FAIL | dom/imptests/editing/conformancetest/test_runtest.html | [["forecolor","#0000FF"],["justifyfull",""],["inserttext","a"]] "foo[]bar" queryCommandValue("forecolor") after - [["forecolor","#0000FF"],["justifyfull",""],["inserttext","a"]] "foo[]bar" queryCommandValue("forecolor") after: assert_equals: Wrong result returned (after color normalization) expected "rgb(0, 0, 255)" but got "rgb(0, 0, 0)" Line 505: 1471 INFO TEST-UNEXPECTED-FAIL | dom/imptests/editing/conformancetest/test_runtest.html | [["forecolor","#0000FF"],["justifyleft",""]] "foo[]bar" queryCommandValue("forecolor") after - [["forecolor","#0000FF"],["justifyleft",""]] "foo[]bar" queryCommandValue("forecolor") after: assert_equals: Wrong result returned (after color normalization) expected "rgb(0, 0, 255)" but got "rgb(0, 0, 0)" Line 514: 1473 INFO TEST-UNEXPECTED-FAIL | dom/imptests/editing/conformancetest/test_runtest.html | [["forecolor","#0000FF"],["justifyleft",""],["inserttext","a"]] "foo[]bar" queryCommandValue("forecolor") after - [["forecolor","#0000FF"],["justifyleft",""],["inserttext","a"]] "foo[]bar" queryCommandValue("forecolor") after: assert_equals: Wrong result returned (after color normalization) expected "rgb(0, 0, 255)" but got "rgb(0, 0, 0)" Line 523: 1475 INFO TEST-UNEXPECTED-FAIL | dom/imptests/editing/conformancetest/test_runtest.html | [["forecolor","#0000FF"],["justifyright",""]] "foo[]bar" queryCommandValue("forecolor") after - [["forecolor","#0000FF"],["justifyright",""]] "foo[]bar" queryCommandValue("forecolor") after: assert_equals: Wrong result returned (after color normalization) expected "rgb(0, 0, 255)" but got "rgb(0, 0, 0)" Line 532: 1477 INFO TEST-UNEXPECTED-FAIL | dom/imptests/editing/conformancetest/test_runtest.html | [["forecolor","#0000FF"],["justifyright",""],["inserttext","a"]] "foo[]bar" queryCommandValue("forecolor") after - [["forecolor","#0000FF"],["justifyright",""],["inserttext","a"]] "foo[]bar" queryCommandValue("forecolor") after: assert_equals: Wrong result returned (after color normalization) expected "rgb(0, 0, 255)" but got "rgb(0, 0, 0)" Line 541: 1503 INFO TEST-UNEXPECTED-FAIL | dom/imptests/editing/conformancetest/test_runtest.html | [["delete",""],["inserttext","a"]] "foo_[bar]baz" compare innerHTML - [["delete",""],["inserttext","a"]] "foo_[bar]baz" compare innerHTML: assert_equals: Unexpected innerHTML (after normalizing inline style) expected "foo_abaz" but got "fooabaz" Line 550: 1505 INFO TEST-UNEXPECTED-FAIL | dom/imptests/editing/conformancetest/test_runtest.html | [["delete",""],["inserttext","a"]] "foo^[bar]baz" compare innerHTML - [["delete",""],["inserttext","a"]] "foo^[bar]baz" compare innerHTML: assert_equals: Unexpected innerHTML (after normalizing inline style) expected "foo^abaz" but got "fooabaz" Line 559: 1511 INFO TEST-UNEXPECTED-FAIL | dom/imptests/editing/conformancetest/test_runtest.html | [["delete",""],["inserttext","a"]] "foo_{<font size=3>[bar]</font>}baz" compare innerHTML - [["delete",""],["inserttext","a"]] "foo_{<font size=3>[bar]</font>}baz" compare innerHTML: assert_equals: Unexpected innerHTML (after normalizing inline style) expected "foo_{<font size=\"3\">a</font>}baz" but got "foo<font size=\"3\">a</font>baz" Line 568: 1521 INFO TEST-UNEXPECTED-FAIL | dom/imptests/editing/conformancetest/test_runtest.html | [["delete",""],["inserttext","a"]] "foo_[barbaz]" compare innerHTML - [["delete",""],["inserttext","a"]] "foo_[barbaz]" compare innerHTML: assert_equals: Unexpected innerHTML (after normalizing inline style) expected "foo_a" but got "fooa" Line 577: 1523 INFO TEST-UNEXPECTED-FAIL | dom/imptests/editing/conformancetest/test_runtest.html | [["delete",""],["inserttext","a"]] "foo^[barbaz]" compare innerHTML - [["delete",""],["inserttext","a"]] "foo^[barbaz]" compare innerHTML: assert_equals: Unexpected innerHTML (after normalizing inline style) expected "foo^a" but got "fooa" Line 586: 1529 INFO TEST-UNEXPECTED-FAIL | dom/imptests/editing/conformancetest/test_runtest.html | [["delete",""],["inserttext","a"]] "foo_{<font size=3>[bar</font>}baz]" compare innerHTML - [["delete",""],["inserttext","a"]] "foo_{<font size=3>[bar</font>}baz]" compare innerHTML: assert_equals: Unexpected innerHTML (after normalizing inline style) expected "foo_{<font size=\"3\">a</font>}" but got "foo<font size=\"3\">a</font>"

(In reply to Jorg K from comment #52) > 2) How to print the string content of a nsIAtom? Easiest way from C++ is to use nsAtomString(aAtom).get(); which is a const char* that you can feed to %s in printf etc.

(In reply to Jorg K (at the beach until 22nd April) from comment #44) > Nope, that ain't working. If I select the "bar" and type over it, the whole > link goes. > > So any suggestions on how to debug this further? Copy this to the URL bar: data:text/html,<!doctype html> <div contenteditable>foo<font color=brown>< a href=http://google.com>bar</a></font>baz</div> <script>var text = document.querySelector("a").firstChild; getSelection().collapse(text, 0); getSelection().extend(text, 3); document.execCommand("inserttext", false, "a");</script> (In reply to Jorg K (at the beach until 22nd April) from comment #47) > How about marking this test as known failure and be done with it? > It's not anything that the user can do, select the text part of a link only > and type over it (see comment #43 and comment #44). > In another bug 1141017 we removed 114 known failures, so here we would add > one. I don't think this is a good idea at all. Fixing bugs is good, but it doesn't usually justify adding regressions. Even if this specific test is unrealistic, if your patch regresses this case, it's likely that it also regresses others, and we don't know it won't affect users. Users can deal with bugs that have always been there, but they get very aggravated if something used to work and no longer does. But it's Ehsan's decision to make. (I suggest that if you want to get Ehsan's feedback promptly, you make sure to write your questions very concisely, because he's extremely busy! Lots of long comments like these take time to read through.) (In reply to Jorg K (at the beach until 22nd April) from comment #49) > Further analysis in/of IsCSSEquivalentToHTMLInlineStyleSet: > > If no value is passed in, it always returns true, jolly good: > http://mxr.mozilla.org/mozilla-central/source/editor/libeditor/ > nsHTMLCSSUtils.cpp#1114 > > Then, it retrieves the colour from the element, it returns (0,0,255) - which > is BLUE!! (not red as stated in comment #48). > Yes, the link is BLUE, it's not brown despite the inline style saying so. Without reading the code -- I would think the link is blue in CSS mode, but brown in non-CSS mode. <a href> is blue, and it's inside the <font>, so the innermost style wins. Non-CSS mode in Firefox currently ignores the color generated by <a href>, so it will think the link is brown (which it's not). The way Firefox behaves in non-CSS mode is essentially one giant bug IMO and my spec says not to behave that way, but that's the status quo. (In reply to Jorg K (at the beach until 22nd April) from comment #52) > 1) What is this used for? bool useCSS = mHTMLEditor->IsCSSEnabled(); Why not > always use CSS? Look up the command styleWithCSS. The author can do execCommand("styleWithCSS", false, "true") to make output use CSS styles, like <span style="color: brown">; or execCommand("styleWithCSS", false, "false") to make output use non-CSS styles, like <font color=brown>. If styleWithCSS is false, IIRC, Gecko ignores all CSS, which IMO is a major bug. But it means Gecko really thinks the color here is brown: data:text/html,<!doctype html> <div contenteditable>foo<font color=brown>< a href=http://google.com>bar</a></font>baz</div> <script>var text = document.querySelector("a").firstChild; getSelection().collapse(text, 0); getSelection().extend(text, 3); document.body.textContent = document.queryCommandValue("forecolor");</script> outputs "rgb(0, 0, 238)", i.e., brown. Although actually, it seems to also do that if styleWithCSS is true. :( Maybe I misremembered how Gecko works here.

Comment on attachment 8594341 [details] [diff] [review] Proposed patch to fix the colour issue Review of attachment 8594341 [details] [diff] [review]: ----------------------------------------------------------------- ::: editor/libeditor/nsHTMLEditRules.cpp @@ +7498,5 @@ > + // Don't use CSS for <font size>, we don't support it usefully (bug 780035) > + useCSS = false; > + } else if (mCachedStyles[j].tag == nsGkAtoms::font && > + mCachedStyles[j].attr.EqualsLiteral("color")) { > + // Do use CSS for <font color>. This doesn't look like the right approach. Why only font color? If we want to always use CSS, we should always use it. If not, we shouldn't. I think we should always use CSS for detecting the current style, but it will probably need more changes than just here. I don't know if it makes sense within the scope of this bug. Ehsan may be able to help here, because he's much more familiar with the bigger picture of how the Gecko editor works -- I know how my spec and tests work, but much less about Gecko.

First of all: Thanks to Neil and Aryeh for their comments. Here a summary: So far I see the following options forward with this bug: 1) Do nothing in core and work around it in the Thunderbird code. Neil already submitted a patch which we can revive. This would leave nsHTMLEditor::GetInlineProperty broken. 2) Accept the patch that was proposed (and curiously received review+) and additionally add one more "expected fail" to the conformance test. I'll personally will pay $10 to every user who is affected ;-) As you pointed our, the regression is in non-CSS mode, so who cares about this? 3) Fix the regression. For Nr. 3 I've already done a try run *always* using CSS. That produced 66 more nonconformities (see comment #53. As soon as the try server goes online again, I will run a second attempt with only using CSS for color as proposed in attachment 8594341 [details] [diff] [review] (which Aryeh didn't like). I'll have to see how many more errors we get from this. The other option to kill the regression of course would be a real HACK. When "color" is queried by calling nsHTMLEditor::GetInlineProperty, not pass the color to check, but pass Null to that nsHTMLEditor::GetInlineProperty will work as before. Or not call nsHTMLEditor::GetInlineProperty at all and always assume "true". As soon as I have the full picture, I'll write it up "short and sweet" for Ehsan to comment.

Next try with the patch from attachment 8594341 [details] [diff] [review] (use CSS for color) gave 36 M2 failures: https://treeherder.mozilla.org/#/jobs?repo=try&revision=a51422c19a7b So still now good. Here goes the next try, as mentioned, the last resort is hacking: https://treeherder.mozilla.org/#/jobs?repo=try&revision=8a3fab9201bd

https://treeherder.mozilla.org/#/jobs?repo=try&revision=8a3fab9201bd surprisingly worked, so here goes a broader test to prepare for review: https://treeherder.mozilla.org/#/jobs?repo=try&revision=04f4171470af

Actually, perhaps even better like this: https://treeherder.mozilla.org/#/jobs?repo=try&revision=a74537f27567

Carrying forward Ehsan's review+. Clear try: https://treeherder.mozilla.org/#/jobs?repo=try&revision=b79864136873 *Note to the person checking this in:* This patch modifies editor/libeditor/tests/chrome.ini. Recently something was landed in bug 1154791 which also modified this file: https://hg.mozilla.org/integration/mozilla-inbound/rev/85187d5c5219 So there may be a reject when applying this patch. The desired end result is: [test_bug1100966.html] [test_bug1102906.html] [test_bug1101392.html] [test_bug1140105.html] <-- new line added in this bug [test_bug1154791.html] <-- new line added in bug 1154791 [test_composition_event_created_in_chrome.html] [test_contenteditable_text_input_handling.html] [test_dragdrop.html] I've refreshed my local tree, but the change for bug 1154791 hasn't arrived, so at the moment I can't provide a different patch here. I hope you can sort this out.