[OJS 3.0.2] Can't switch to APA or CBE Citation Formats

[Ph-We]

If we try to switch to either APA or CBE the displayed format stays the same.
http://forum.pkp.sfu.ca/t/problem-with-first-name-displaying-in-citation/28641

[asmecher]

@Ph-We, I'm afraid I can't replicate this behavior -- both citation formats appear to work OK for me. Can you investigate further?

[Ph-We]

@asmecher, Yes, after some further investigation I've found that the issue is stably replicated only with Cyrillic entries. This might probably be something akin to that one:
http://forum.pkp.sfu.ca/t/problem-with-first-name-displaying-in-citation/28641/7

Here is an example
https://cmd-journal.hse.ru/article/view/2900
(switch to Русский)

[asmecher]

@Ph-We, can you check your browser's console for a Javascript error?

[asmecher]

Using Firebug, I see that the response is blank; did anything show up in the PHP error log?

[Ph-We]

Hi @asmecher, I've traced the issue might be related to this error in PHP log:
[19-May-2017 13:09:03 Europe/Berlin] Malformed UTF-8 characters, possibly incorrectly encoded

[asmecher]

@Ph-We, I think this means that a call to PHP's json_encode encountered invalid character encodings. Is it possible that there is non-UTF8 data in your database, or that your database connection parameters in config.inc.php aren't appropriate for UTF8?

[Ph-We]

@asmecher, I could reproduce this with three different installations: productive, test, localhost. On the latter there was only one article with most simple words in its title. And I did not touch config.inc.php. So I doubt there could be any non-UTF8 or malformed UTF8 data.

[asmecher]

@Ph-We, I suspect the message is coming from JSONMessage.inc.php. In order to confirm that this is the case, and check what content is causing trouble, can you modify the file and change...

error_log(json_last_error_msg());

...to...

error_log(json_last_error_msg());
error_log("Subject: " . print_r($jsonObject, true));

[Ph-We]

@asmecher, thanks, now I've got something more informative:

[26-May-2017 10:50:04 Europe/Berlin] Malformed UTF-8 characters, possibly incorrectly encoded

[26-May-2017 10:50:04 Europe/Berlin] Subject: Array
(
    [status] => 1
    [content] => 	

		Р�ванов, Р., & 		РЎРёРґРѕСЂРѕРІ, Р.
(2017).
 Статья о главном.
<em>Журнал о жизни, 1</em>(1), 1-33.
doi:10.1234/.v1i1.3
    [elementId] => 0
)

So it seems to count any Cyrillic characters as malformed UTF8, since they really get malformed at some stage.

[asmecher]

Ah, @Ph-We, I found it. The Smarty truncate modifier was being used to get the author's first initial for both of those formats, and it's not UTF-8 aware. Try the patch attached to this issue.