[Feechen]

Dear Roland,

I hope this is the right thread to report bugfixes and enhancements to your add-on!

If not, please bear with me (I'm new to this) and tell me the right place, ok?

I installed this add-on of yours recently and I'm VERY fond of it!

However, since I wanted to make some changes, I came across something that appears to possibly be a little bug.

The file "product-acronym_mod.xml" contains a Perl regular sub-expression which reads as follows: (([.,:;-?!()\s"<\[]|$))

However, the part ";-?" in it means "including all characters from ; to ?" in Perl, i.e., the characters "; < = > ?" in ASCII or ISO-Latin-1.

This is probably not what was intended, I guess the dash itself was meant instead. To achieve this, and to free the dash from its special meaning in character classes, you need to escape it with a backslash, like so: (([.,:;\-?!()\s"<\[]|$))

Moreover, I found out that the add-on also does replacements inside of HTML tags, which can cause problems like creating letter salad and make hyperlinks stop from working.

In my case it was the "ALT" tag, which incidentally corresponds to an abbreviation in our forum (= Alanine-Amino-Transferase).

Therefore I introduced an additional, negative zero-width look-ahead expression to prevent substitutions inside of HTML tags, directly after the match of the abbreviation (or acronym ), like this: (?![^<>]*>)

This verifies that the intended substitution is not taking place inside an HTML tag by verifying that no closing ">" follows.

As a downside of this, abbreviations followed by a plain ">" somewhere in the middle of normal text (before the next HTML tag) will not work anymore. I personally do not find this very limiting, in practice, though.

However, abbreviations can be made to work in these circumstances if the ">" that follows is in reality written as "& gt ;" (without the blanks, of course).

For this, the list of non-word characters or separating characters in the add-on needs to be adapted to include at least the ampersand "&".

I found it convenient to put ALL punctuation and special characters into that list of possible separator characters, except for the backslash, which for a reason I haven't investigated yet did not work, and the "at" sign "@", because it might conflict easily with email addresses.

I also found it useful to make the separators to the left and the right of the match symmetrical. This has the added benefit of making staggered abbreviations, such as for instance in "TPO-AB", possible, i.e., it allows "TPO" to be recognised as an abbreviation for "thyroid peroxidase" and "AB" as an abbreviation for "antibodies" (in theory this feature would also allow something like "XX-YY-ZZ" to work, although I guess this won't be needed very often in practice).

The whole line now reads as follows: $post['message'] = preg_replace('#
(?<=[\s!"\#\$%&\'()*+,\-.\/:;?^`{|}~>\]]|^)(' . $acros_final['acro'] . ')(?![^<>]*>)
(([\s!"\#\$%&\'()*+,\-.\/:;?^`{|}~<\[]|$))#siU',
'<acronym title="'.$acros_final['desc'].'"><span style="border-bottom: 1px dotted #082ceb; cursor: help;">\\1</span></acronym>\\2', $post['message']);

I am attaching the original zip archive of the add-on as a reference, and also the modified file "product-acronym_mod.xml" itself (if this forum's configuration will let me).

BTW, I also changed the font style to be just underlined, not green and bold, in order not to disrupt normal reading too much.

I hope that these changes may be useful to others as well!

Best regards,
ibeji
in the name of Feechen

P.S.: I have based my changes on the German version, but I guess it should be no big deal to transfer the modifications to the English version, if those were different at all.

P.P.S.: See also my posting in the German forum in the thread "[vB 3.5.x]Akronyme f?r vB".