[rt-users] JS minify broken (was: RE: RT4 slowness)

[Adam Thompson]

One additional note: the problem occurs under mod_perl, but NOT when 
running rt-server standalone.  That makes it an easy fix, for now :-)
I'm rather unclear on how to set up anything other than Apache2.2 
w/mod_perl and RT4.  Isn't that the default configuration now, anyway?
-Adam


> Kevin, I'm also experiencing the same problem as reported here.  (And
> yes, I just filed a new, duplicate bug about that, [rt3 #17673].) I can
> provide direct access (including root access) to the pre-production VM
> I'm working with, if that would help?
> Just for kicks, I tried substituting "/bin/cat" for 
> "/usr/local/bin/jsmin", but
> after turning on a lot of debugging I just get an error:
>
> Log::Dispatch::__ANON__('Log::Dispatch=HASH(0x2aef1bb0fe40)',
> 'failed to
> jsmin: /bin/cat: standard output: Bad file descriptor^J ') called at
> /opt/rt4/sbin/../lib/RT/Squish/JS.pm line 99
>
> With the same debug options on, I do _not_ get the same error when
> invoking jsmin.  I _do_ get the UI breakage, however :-).
>
> CentOS 5.6, CPAN packages from RPMForge where possible, CPAN
> otherwise (only as required).
> Rt-4.0.0 from tarball.
> Jsmin as downloaded from Crockford's site today (compiled with stock
> gcc on CentOS 5.6, no options).
>
> I added another warning line to JS.pm to tell me if jsmin was being 
> called
> on each and every request, and it indeed is being called on every single
> request.
>
> Furthermore, if I examine the minified JS using Chrome's developer
> tools, it's an empty file (http://**.****.***/NoAuth/js/squished-
> d41d8cd98f00b204e9800998ecf8427e.js
> = 0 bytes).
>
> There is no such cached object on disk under /opt/rt4/var/, which I
> suspect would be the problem.  My Apache (and consequently
> mod_perl) process does have write permission to
> /opt/rt4/var/mason_data to .../session_data.
>
> Don't know what would be useful to try next...
>
> -Adam Thompson
>  athompso at athompso.net
>
>