No more "HTML5 Nu Validator"? #44
Comments
I gather @menovak cannot reproduce this -- @stevefaulkner do you get this error in the latest WAT release? |
There seems to be something wrong with referring to this host as the validator; it doesn't seem accessible to me:
I don't know why it is accessible to @menovak but not me, but it looks like we should be using the main host instead of the subdomain? |
I would ask Steve again or others in the group, I do not seem to have the On Mon, Dec 22, 2014 at 5:59 AM, Matthew Tylee Atkinson <
|
@menovak what do you get when you run the same commands on your system? (The Windows equivalent command is @stevefaulkner: if you're reading this, could you try the same please? Either way, it does seem that we need to get this changed somehow, as Shirley I can't be the only person in the world who can't access that host? |
@matatk yes i get same error when i use (as html5) options, suggest that this option is not needed in the wat, just have the w3c nu validation service function (and same with open in new window) |
Sounds fine to me; I guess the latest checker does all dialects anyway. I suggest that we assign this to @menovak and also ask @jun325 to remove the corresponding code, which appears to be in the DLL file (it is referenced from Translation.ini). |
just so I am clear, are we removing both the W3C Nu Validation Service (as On Mon, Dec 22, 2014 at 10:18 AM, Matthew Tylee Atkinson <
|
@menovak yes :) |
I am experiencing a problem with the new Nu Markup Checker...
IE11, Win 8.1 |
If the frame is used, some windows are opened. Please confirm the WAT command in the traslation.ini. |
hi Jun In the Accessibility_Toolbar.xml file, the two remaining menu validate Which translate to in the Translation.ini file validatorNU=nu_auto So if I understand correctly, the "validatorNU" should be changed to Regards, Mark On Thu, Jan 15, 2015 at 2:10 AM, Jun notifications@github.com wrote:
|
@menovak yes if you don't want the validation of other frames |
My understanding (probably wrong) was that the whole code of the page (including content of iframes) would be POSTed (i.e. via HTTP) to the validator and checked as one page. It sounds, from @jun325, like each iframe actually has to be validated separately. We should definitely validate all of the page content, but currently it's confusing as to why all those tabs open -- though documentation is the answer to that perhaps. The main problem I am having is that one of those tabs that's opened seems to overwrite the current one (I would like the current page to stay as it is, and then an extra tab/tabs open to show the validation results). When only one extra tab opens to show validation results, it works just fine, but when more than one tab opens to show validation results, it seems to overwrite the original page's tab. |
Thanks Jun. Matt, assume you are following this thread, so we need to add an item to Mark, On Thu, Jan 15, 2015 at 10:03 AM, Jun notifications@github.com wrote:
|
hi Matt, Opening in a new window is also an easy fix, again a simple change to the Perhaps something added to the documentation as to the difference then Mark On Thu, Jan 15, 2015 at 10:08 AM, Matthew Tylee Atkinson <
|
I just want to recap what the actual issue(s) are here, before we make any changes...
So we have two things to fix: one is a matter of documentation and another seems to be a bug, but I'm not sure if it's a bug in the DLL or in our the scripts (I guess it may be a bug in the DLL, as that's where the function that calls the validation service lives?) |
hi Matt, What we currently have is: A. If you select W3C Nu Markup Checker it B. If you select W3C Nu Markup Checker (new window) it There are no bugs, this behavior is doing what the menu selections are We can change A above to open in a new tab window as B does, and it would If we 'always' want to validate the main window and any iframe(s) found, I was suggesting some documentation to alert users to the difference in Note, as it currently works, if there are no iframe(s) found, then the Does that help? Mark On Thu, Jan 15, 2015 at 11:02 AM, Matthew Tylee Atkinson <
|
Aha, this is interesting… I do not get the behaviour you described. The DLLs have just been updated to allow Translation.ini to support UTF-8. I wonder if this may be related? Using the latest build (just uploaded to GitHub), I get the following behaviour on IE 11 when looking at the IE default home page: http://www.msn.com/en-gb/?ocid=iehp Using “W3C Nu Markup Checker”:
Using the “W3C Nu Markup Checker (new window)”:
Maybe there’s an option in IE for opening new windows as tabs — this is now a default in most browsers, so maybe IE is following suit. |
hi Matt 1 - you sure pick 'interesting' sites for review, note, this site has at 2 - when I said a 'new tab window', that is the same as you saying 'tab Using my last build: (Win 7, IE 11, comments in bold below) Using “W3C Nu Markup Checker”:
Using the “W3C Nu Markup Checker (new window)”:
Can you look at your copy of Accessibility_Toolbar.xml, line 24 and 25 I I suspect that your's got changed to the following, if the first check is MarkOn Thu, Jan 15, 2015 at 12:25 PM, Matthew Tylee Atkinson <
|
hi Matt Just retested that IE home page, and found 12 iframes, so indeed it is Mark On Thu, Jan 15, 2015 at 12:25 PM, Matthew Tylee Atkinson <
|
OK, agreed re test page, but isn’t it odd that choosing the “(new window)” option doesn’t open all those extra tabs — it sounds like it should! I’m now trying http://bbc.co.uk/ and I find the same behaviour as I described above, but the page only appears to have itself and one iframe, so when I use the ‘normal’ checker, I get two validator tabs (one which overwrites the BBC home page). But when I use the “(new window)” checker, I get only one extra tab. One is missing, then? Sounds like an off-by-one error in the DLL? Here’s the XML file, seems it has an extra line, but otherwise seems correct: https://github.com/ThePacielloGroup/WebAccessibilityToolbar/blob/2015-01-14a/wat/Accessibility_Toolbar.xml#L25 (at the time of writing, this is the current version on master, too). |
I think what is more interesting, but also somewhat expected, is that you ""OK, agreed re test page, but isn’t it odd that choosing the “(new
its not an 'off by one error in the DLL', two different functions are being On Thu, Jan 15, 2015 at 2:05 PM, Matthew Tylee Atkinson <
|
It has always been the case that running the checking tools will overwrite the current page with the results. The new window (now really new tab option) has always left the original page in current tab and opened results in a new tab (if pages contained (iframes) in new tabs.) Sounds like @matatk does not like the overwriting bit. I am OK with having 2 options What would be nice if we could have all results included in one new page with results being included in multiple iframes, which would overcome the annoyance of many tabs opening. |
@menovak I'm not choosing good test sites, sorry -- US version will be different, d'oh! @stevefaulkner I agree with all that you've said, except there is still a bug here -- if you try the BBC site, I believe you'll see that...
This seems to indicate that there is a bug (off-by-one error?) when using the 'new window' option. So there are potentially two things to change here...
If we can confirm that there is an off-by-one (or similar bug) then I think we'd need @jun325 to fix it. However I think myself and/or @menovak should implement the new results scheme (as we have all the kit to test it to ensure it's accessible). Does this sound OK? First step is: can @stevefaulkner confirm the difference in number of tabs opened when testing the BBC site. Also if someone could find a test site that has at least one iframe and is the same worldwide, so all of us get the same code, that would be great :-). |
hi Matt and Steve, I don't believe we are 'off by one', I think the two different functions I also prefer to leave as is, two options... but add some documentation to Mark On Fri, Jan 16, 2015 at 8:00 AM, Matthew Tylee Atkinson <
|
From re-reading the thread and looking at the code, I can see that the numbers of tabs are intentional and not an off-by-one error — sorry for ‘not getting this’ before. (There are also quite a lot of now-redundant lines in Translation.ini relating to the validators which, with some testing, and now that we have Translation.ini in a format git sees as text, it will be safe to remove.) However, I find it very confusing that the options do two different things to each other — I would like results from the validation to always appear in a new tab. Then the only difference between the options would be that one checks only the main page and one checks all frames. Does this sound OK? If so, in order to close this bug, I propose that:
As we are perennially short on time, we should record @stevefaulkner’s good idea of including all the results in one tab as a separate enhancement issue for… later… :-) |
Hi Matt, ""However, I find it very confusing that the options do two different I'm OK with this .. to always have the results in a new tab.. ""If so, in order to close this bug, I propose that:
We don't need Jun to do anything, I just tested with this change (see ""
Be sure you get them in the correct order, right now the first one on the ""
Thanks for volunteering to update DOCS. PLEASE be careful removing items ""As we are perennially short on time, we should record @stevefaulkner Agree, can make this change another time, as then the validator(s) would be Thanks Mark On Fri, Jan 16, 2015 at 10:10 AM, Matthew Tylee Atkinson <
|
@menovak thanks for all that information. Also very-much noted with respect to the .ini file -- the good news is that from no on, we have Git on hand to make fixing any such problems easy (though I don't anticipate them this time; will test it a lot, first, too ;-)). @jun325 looks like we do not need you to do any work on this one after all, thanks for putting up with this thread! |
When I try to validate a page with either of the Nu Validator HTML5 options, I get a page, generated by the Nu Validator, with the following message on it:
Could it be that the separate HTML5 mode has been retired now HTML5 is a recommendation?
The text was updated successfully, but these errors were encountered: