ePerl(3pm) | User Contributed Perl Documentation | ePerl(3pm) |
Apache::ePerl - Fast emulated Embedded Perl (ePerl) facility
# Apache's httpd.conf file # mandatory: activation of Apache::ePerl PerlModule Apache::ePerl <Directory /root/of/webmaster/area> <Files *.iphtml> Options +ExecCGI SetHandler perl-script PerlHandler Apache::ePerl </Files> </Directory> # optional: configuration of Apache::ePerl <Perl> $Apache::ePerl::Config->{'BeginDelimiter'} = '<?'; $Apache::ePerl::Config->{'EndDelimiter'} = '!>'; $Apache::ePerl::Config->{'CaseDelimiters'} = 0; $Apache::ePerl::Config->{'ConvertEntities'} = 1; </Perl> # optional: activation of Apache::Status for Apache::ePerl <Location /perl-status> Options +ExecCGI SetHandler perl-script PerlHandler Apache::Status </Location>
These packages provides a handler function for Apache/mod_perl which can be used to emulate the stand-alone Server-Side-Scripting-Language ePerl (see eperl(3) for more details) in a very fast way. This is not a real 100% replacement for nph-eperl because of reduced functionality under some special cases, principal runtime restrictions and speedup decisions. For instance this variant does not (and cannot) provide the SetUID feature of ePerl nor does it check for allowed filename extensions (speedup!), etc. Instead it uses further features like object caching which ePerl does not use.
But the accepted bristled source file format is exactly the same as with the regular ePerl facility, because Apache::ePerl uses the Parse::ePerl package which provides the original ePerl parser and translator. So, any valid ePerl which works under nph-eperl can also be used under Apache::ePerl.
The intent is to use this special variant of ePerl for scripts which are directly under control of the webmaster. In this situation no real security problems exists for him, because all risk is at his own hands. For the average user you should not use Apache::ePerl. Instead additionally install the regular stand-alone ePerl facility (nph-eperl) for those users.
So, the advantage of Apache::ePerl against the regular nph-eperl is better performance and nothing else. Actually scripts executed under Apache::ePerl are at least twice as fast as under nph-eperl. The reason its not that ePerl itself is faster. The reason is the runtime in-core environment of Apache/mod_perl which does not have any forking overhead.
First you have to install Apache::ePerl so that Apache/mod_perl can find it. This is usually done via configuring the ePerl distribution via the same Perl interpreter as was used when building Apache/mod_perl.
Second, you have to add the following config snippet to Apache's httpd.conf file:
PerlModule Apache::ePerl <Directory /root/of/webmaster/area> <Files *.iphtml> Options +ExecCGI SetHandler perl-script PerlHandler Apache::ePerl </Files> </Directory>
This forces all files under the directory /root/of/webmaster/area/ with extension .iphtml to be processed by the Apache::ePerl::handler function which emulates the runtime behavior of the stand-alone "eperl" program (when run as a SSSL) up to 90%.
If you're not paranoid about security (for instance driving a stand-alone webserver without user accounts) you can also just use
PerlModule Apache::ePerl <Files *.iphtml> SetHandler perl-script PerlHandler Apache::ePerl </Files>
which enables .iphtml files everywhere.
Third, when you want to change the defaults of the ePerl parser, you also can add something like this to the end of the snippet above.
<Perl> $Apache::ePerl::Config->{'BeginDelimiter'} = '<?'; $Apache::ePerl::Config->{'EndDelimiter'} = '!>'; $Apache::ePerl::Config->{'CaseDelimiters'} = 0; $Apache::ePerl::Config->{'ConvertEntities'} = 1; </Perl>
Fourth, you can additionally enable the mod_perl runtime status which then automatically enables an Apache::ePerl status handler:
<Location /perl-status> Options +ExecCGI SetHandler perl-script PerlHandler Apache::Status </Location>
This enables the URL "/perl-status" in general and the URL "/perl-status?ePerl" in special. Use it to see how much scripts where run and how much are still cached.
Ralf S. Engelschall rse@engelschall.com www.engelschall.com
Apache::ePerl was first implemented by Mark Imbriaco <mark@itribe.net> in December 1996 as a plain Perl module after he has seen the original ePerl from Ralf S. Engelschall. It implemented the ePerl idea, but was not compatible to the original ePerl. In May 1997 Hanno Mueller <hmueller@kabel.de> has taken over the maintenance from Mark I. and enhanced Apache::ePerl by adding caching for P-Code, adding the missing "chdir" stuff, etc.
Nearly at the same time Ralf S. Engelschall was unhappy of the old Apache::ePerl from Mark I. and already started to write this version (the one you are current reading its POD). He has rewritten the complete module from scratch, but incorporated the P-Code caching idea and the Apache::Status usage from Hanno M.'s version. The big difference between this one and Mark I.'s or Hanno M.'s versions are that this version makes use of the new Parse::ePerl module which itself incorporates the original ePerl parser. So this version is more compliant to the original ePerl facility.
Web-References:
Perl: perl(1), http://www.perl.com/ ePerl: eperl(1), http://www.engelschall.com/sw/eperl/ mod_perl: mod_perl(1), http://perl.apache.org/ Apache: httpd(7), http://www.apache.org/
2022-10-20 | perl v5.36.0 |