Greater Than & Less Than Symbols in Abstracts

[tshore]

I was quite surprised to find out about this bug from an author. They submitted an abstract containing this:

p<.05

and it worked fine. However, when one tries to view the abstract it's cut off, ending right before the <. I can get it to display properly by replacing < with &lt;, however, this needs to be fixed. I tried searching this forum and Bugzilla, but I had no luck.

[jmacgreg]

Thanks, Trudy, I've posted this as a bug:

http://pkp.sfu.ca/bugzilla/show_bug.cgi?id=3316

Cheers,
James

[tshore]

I thought that this had been fixed last year? Anyway, it's popped up again this year. The bug you refer to (#3316) never resulted in a patch and the bug that it refers to (#3807), I can't really figure out. What am I supposed to do with my last year's version of OCS to deal with this?

http://www.imrf.info/2009/

[mj]

Hi Trudy,

The resolution that resulted from this bug was part of an overhaul in character encoding handling which is now part of the 2.3 line of OCS, OJS, etc. Unfortunately we couldn't come up with a simple patch that would address the issue for the 2.1.x and 2.2.x releases without some major changes to the code. The good news is that OJS and OCS 2.3 will both be released later this year, but I understand that this won't help you in the interim.

What I might suggest is that, for the affected abstracts, replacing the unencoded greater-than > and less-than > signs with their encoded HTML entity references: &gt; and &lt; respectively. I also noticed that the current conference site you link to seems to be running OCS 2.0.0 -- character handling has been improved in the more recent releases (2.1.1), so you might want to try upgrading and see if that addresses the issue to any degree.

Hope this helps,
MJ

[tshore]

Me again. I thought this had been fixed? I'm using a copy of last year's OCS 2.3.3 installation for a current conference, and an author has just pointed out that her abstract has been truncated. And, as it turns out, it's caused by a less-than symbol. Oddly enough, I can't remember what I did to fix it in 2009 - three years ago. I just double-checked my OCS 2.3.4 conference and it still can't deal with < or > that are not followed by a space. Also, not sure if this is relevant, but to check this in OCS 2.3.4, I went in as a director and clicked on EDIT METADATA for a submission. And, when you do this, you don't get TinyMCE. So what I'm saying is, the problem isn't caused by TinyMCE.

[jmacgreg]

Hi Trudy,

I can't replicate either problem in OCS 2.3.4/5: pasting

< abstract < just trying > some of thing s out.

into the abstract field displays just fine later on; and the TinyMCE field appears as it should when I edit the submission metadata as a Director. Is this an upgraded OCS instance? Is your version of TinyMCE up-to-date? You may want to try upgrading as per http://pkp.sfu.ca/bugzilla/show_bug.cgi?id=7207 (which is a security bug as well, so definitely worth applying regardless).

Cheers,
James

[tshore]

This is the issue, from my post above:

can't deal with < or > that are not followed by a space

And I wasn't worried whether or not I should've gotten TinyMCE when I edited it as a director, I was just pointing that out to rule out TinyMCE as a cause of the bug.

[asmecher]

Hi Trudy,

You *should* get TinyMCE for that field, and if you don't, then the situation you describe (odd behavior depending on whether e.g. < or > are followed by a space) is expected behavior -- either due to the browser's handling of incomplete HTML tags or the unsafe HTML strip code. The fix will be to get TinyMCE working for that field, most likely by upgrading to the newest release.

Regards,
Alec Smecher
Public Knowledge Project Team

[tshore]

Can you test it again on your installation with no space after the < or >?

I've turned off TinyMCE for abstracts. In the past, I've had tonnes of problems with TinyMCE and formatting, especially with abstracts and addresses, so I turn it off for a number of fields.

I've temporarily solved the problem by replacing < by &lt; in the one abstract I know has the problem.

[jmacgreg]

Hi Trudy,

Just confirming that

<somthing> >the< <else>

also works just fine for me. As Alec says, if you turn off TinyMCE you will run into problems with those characters. OCS will normally identify them as potentially malicious and strip them; TinyMCE converts them to &lt; etc. in the background for you on submission so that OCS doesn't strip them.

Cheers,
James