[rt-devel] RT pages without links

[CLOSE Dave]

Robert Nesius wrote:

> Are you using the system perl?   If so, I'd try installing the
> version of perl that was on your old machine with perlbrew, update
> its modules as necessary, and have RT run over that perl.

Yes, I am using the system Perl. I suspect your idea is exactly what I 
was missing. Thanks.

Surprisingly, at least to me, I hadn't heard of perlbrew. I think that's 
a great idea and I'm getting it installed now. But since RT doesn't run 
as an "application" but simply as a series of Apache pages and CGI 
scripts, I'm wondering how to get it to use the older Perl. I don't want 
Apache to use the old version for other pages. Any ideas?

Maybe this will be as simple as changing the shabang lines. But I 
haven't tried it yet so maybe someone can save me a little time.

> I wrote:
>
>> I recently moved our copy of RT 3.8.7 from Fedora 8 to Fedora 19.
>> I would have preferred to concurrently upgrade to the latest RT
>> but had some problems converting the data base. So in the interest
>> of time, I took the simpler approach of just copying the entire
>> installation.
>>
>> It works, mostly. In fact, the only problem I see is that ticket
>> lists on the dashboards are not links. Comparing the old and new
>> installations (both are running though DNS leads users to the new
>> one), I see almost exactly the same page layout and content. There
>> are two obvious differences.
>>
>> 1. Font selections are different. This is not critical and the new
>> fonts are usable. It's clearly an Apache difference.
>>
>> 2. On the default home page, the tickets listed under "highest
>> priority" and "newest unowned" are not active links. Comparing the
>> page source for old and new installations, it is obvious that the
>> href tags are simply not present on the new output. Other items,
>> like the queue summary, do have working links as they did before.
>>
>> Digging through the source code, I quickly got lost. I can see
>> where the collection-as-table class is specified but not where the
>> associated variables are added. It seems likely to me that
>> something in F19 is confusing RT but thus far I haven't found what.
>> If I could see where the variable information was added to the
>> output, I suspect I'd see some kind of test that is now getting a
>> different answer and that that test would lead me to the solution.
>>
>> So, if anyone can tell me the cause of my problem right away, that
>> would be much appreciated. But, if not, then a clue on where to
>> look in the source code might help almost as much.
>
> I'm surprised and disappointed that no one seems able to help with
> this issue. Did I omit something important when describing the
> problem?
--
Dave Close