[Xerte-dev] Re: FW: Further Xerte Data Loss

Julian Tenney Julian.Tenney at nottingham.ac.uk
Wed May 21 09:24:03 BST 2014


In Bootstrap, all text is CDATA (I think), so it's not duff chars.

-----Original Message-----
From: xerte-dev-bounces at lists.nottingham.ac.uk [mailto:xerte-dev-bounces at lists.nottingham.ac.uk] On Behalf Of Ron Mitchell
Sent: 20 May 2014 17:29
To: 'For Xerte technical developers'
Subject: [Xerte-dev] Re: FW: Further Xerte Data Loss

Oh dear maybe this isn't adding to the uncertainty or maybe it's a clue - just had the following reply from Alistair...

"I have still experienced some loss since my Broadband was stabilised. I was working on the Bootstrap template using Chrome browser and lost a set of pages about 2 weeks ago."

That's the first time I'm aware of this happening with Bootstrap too but I guess that's also writing xml so perhaps not really any different?

Ron

-----Original Message-----
From: xerte-dev-bounces at lists.nottingham.ac.uk [mailto:xerte-dev-bounces at lists.nottingham.ac.uk] On Behalf Of Julian Tenney
Sent: 20 May 2014 16:01
To: For Xerte technical developers
Subject: [Xerte-dev] Re: FW: Further Xerte Data Loss

> One of the changes I've committed to github / a pull request is that there's better validation on the XML. i.e. if it doesn't validate then XOT won't overwrite what ever is in the USER-FILES/$whatever/(data|preview).xml file(s).

Cheers.

It does seem the data loss issue has the remote user as the common factor.

-----Original Message-----
From: xerte-dev-bounces at lists.nottingham.ac.uk [mailto:xerte-dev-bounces at lists.nottingham.ac.uk] On Behalf Of David Goodwin
Sent: 20 May 2014 15:47
To: xerte-dev at lists.nottingham.ac.uk
Subject: [Xerte-dev] Re: FW: Further Xerte Data Loss

<snip/>


>
> A student on one of the HEA projects reported this recently and 
> responded as follows:
>
> “No text was copied and pasted,  so I doubt there would be HTML, plus 
> no hyper linking was included....however pages were lost, but only 
> when working or accessing the project at my house anywhere else it 
> 'seemed' to be fine... But I can't be too sure”
>
> This mirrors the only consistent cause Alistair and I were able to 
> identify e.g. he was having connectivity problems and regularly 
> getting corrupt LO's or lost pages. Since his connectivity was fixed I 
> don't think this has happened but I'll check. All that said I've never 
> once had this happen to me even when on trains or using mobile 
> connections including the same codebase or even installations as 
> Alistair.
>
> Even if a new solution is put in place to validate before saving etc 
> it seems to me that there will still be high risk for all those users 
> and installations unlikely to be able to upgrade immediately and would 
> still be good to try to nail down the cause of this.
>

I've certainly seen corruption due to invalid characters being embedded into the LO. I can't replicate this, regardless of what sort of formatting I embed in, which is a pain.


One of the changes I've committed to github / a pull request is that there's better validation on the XML. i.e. if it doesn't validate then XOT won't overwrite what ever is in the USER-FILES/$whatever/(data|preview).xml file(s).

(via plugins/xml_parsing_check.php when/if merged).


David.

--
Pale Purple Ltd

PHP Web application development and support

http://www.palepurple.co.uk
@PalePurpleLtd
07792 380669 / 0845 0046746

_______________________________________________
Xerte-dev mailing list
Xerte-dev at lists.nottingham.ac.uk
http://lists.nottingham.ac.uk/mailman/listinfo/xerte-dev
_______________________________________________
Xerte-dev mailing list
Xerte-dev at lists.nottingham.ac.uk
http://lists.nottingham.ac.uk/mailman/listinfo/xerte-dev


_______________________________________________
Xerte-dev mailing list
Xerte-dev at lists.nottingham.ac.uk
http://lists.nottingham.ac.uk/mailman/listinfo/xerte-dev


More information about the Xerte-dev mailing list