A simple doubt :-/

[Vijoy Varghese]

Hello Group

I have a doubt.

The programs written in 'C' language *.c is compiled to *.obj and then
linked to *.exe which is the 'machine language' substitute for our *.c
program, right? Now for a processor to execute any program, the
program should be coded in a language which the processor can
understand, and here in our case the *.exe is such a language (or may
be one level above the 'language processor can understand', and our
Operating System does the translation work for the processor). So what
ever it be, the *.exe is close to the language which the processor can
understand.

Now, a CGI program written in Perl, *.cgi there is no compilation or
linking. That is each time we try to activate this program the Perl
interpreter have to convert it to the *.exe format(or directly to hard
core machine language of 0's and 1's?) and this process is repeated
during all activation of the *.cgi script, right?

So this is a kind of overhead when a same script is activated some
1000 time in a minute, right? But still people are using Perl for CGI
scripting, why? Yes, Perl is fun, its always fun to program using a
language which remind us more of a 'natural language' rather than a
'programming language'. But still, why don’t some one make a
'Compiled Perl' with all features of current Perl, but which can be
compiled to machine language, so that there is no need to 'translate'
it each time its activated.

Or is there is already one available? I have heard a lot about
'mod_perl' and that it increases the speed of CGI scripts by some 100
times. Is that have something to do with the 'compiled' version :-/

I am expecting a lot of '@#$#$%$@' replies, because I know this is not
the way to ask questions in a group. But friends, this doubt... its
hunting me for a long time, and i cant find a convincing answer
anywhere. :-(

Waiting for your replies or some pointers where I can find answer to
my doubt.

Thanking you all

Regards
Vijoy Varghese

 

[Brian McCauley]

The programs written in 'C' language *.c is compiled...

[snip how compiled lanuages work ]

Now, a CGI program written in Perl, *.cgi there is no compilation or
linking. That is each time we try to activate this program the Perl
interpreter have to convert it to the *.exe format(or directly to hard
core machine language of 0's and 1's?) and this process is repeated
during all activation of the *.cgi script, right?

Actually in converts it into an intermediate form - Perl byte code.
The Perl by code is then interpreted by a virtual machine.

So this is a kind of overhead when a same script is activated some
1000 time in a minute, right? But still people are using Perl for CGI
scripting, why?

Be aware that a lot of the overhead is also in the process creation
overhead. Even if you use native executable CGIs you still have to
pay.

If you want speed one way is to load your web server apps into the
web server itself. For example, in Apache you do this with Apache
loadable modules.

Yes, Perl is fun, its always fun to program using a
language which remind us more of a 'natural language' rather than a
'programming language'. But still, why don’t some one make a
'Compiled Perl' with all features of current Perl, but which can be
compiled to machine language, so that there is no need to 'translate'
it each time its activated.

Attempts have been made but they've not managed to produce significant
speed improvements.

Or is there is already one available?

There are, AFAIK, no production-ready Perl compilers or byte-code loaders.

I have heard a lot about 'mod_perl' and that it increases the speed
of CGI scripts by some 100 times. Is that have something to do with
the 'compiled' version :-/

Yes. In Apache mod_perl the Perl interpreter is kept running and not
restarted for each web transaction so once any module (or pseudo CGI
script) has been loaded and compiled once it need not be again until
the web server is restarted. Actually there's a pool of several Perl
interpreters servicing requests so each module is compiled once for
each interpreter. Also because Perl programs sometime leak memory
each interpreter is killed off and replaced periodically.

If you know in advance what modules are going to be used the you can
preload them at server startup and then they don't need to be loaded
during the first transaction that needs them.

Be aware that mod_perl is not CGI. It is possible to write Perl
source code that can run both as CGI scripts and as mod_perl handler
modules using the 'Registry' mechanisms provided by mod_perl. This is
the mode in which most people (including myself) use mod_perl most of
the time. Such programs need to obey a few simple rules above and
beyond those which real CGI scripts would.

However you can do a lot more in mod_perl than you can in CGI.

Also running mod_perl in the mode where in can use these dual-role
programs will slow it down.

For information about Apache+mod_perl go see the Apache+mod_perl
websites. (Like, D'uh!).

There's some real exciting stuff in Apache2+mod_perl2.

Note: there are other alternatives to CGI that also eleminiate the
process creation-per-transation overhead, e.g. FastCGI.

--
\\ ( )
. _\\__[oo
.__/ \\ /\@
. l___\\
# ll l\\
###LL LL\\

 

[Sam Holden]

I haven't found a way to use mod_perl that wasn't horribly difficult.
This may be my missing something obvious, and if it is, I'd very much
love to be hit with the clue stick. None of the tutorials or documentation
I've seen talks about how to take a functioning, cleanly written CGI
program and make it run correctly under mod_perl.

http://perl.apache.org/docs/1.0/guide/porting.html

Covers most of the issues with moving from CGI to mod_perl. Though
you've probably read it.

 

[pkent]

IME if your program really is a cleanly written CGI it runs Just Fine
under mod_perl. But if that's not your experience I fully second the
recommendation of:

http://perl.apache.org/docs/1.0/guide/porting.html

My hint would be to completely ignore handlers and Apache::* stuff until
you're au fait with the environment... and even then only if you really
want to or need to. I've very very rarely had to do anything other that
required mod_perl handlers, for example. At my place we write stuff so
that it runs as a CGI or under mod_perl just the same, with no changes
(of course running it under mod_perl is generally faster and more
efficient and so on)

P

 

[test]

Hello,



I want to make a little script in perl .
.... but I'm very newbie in this langage



For each sub-directory of a given directory ...

(and for each file in the directory)

.... I want to write in a file something like that :




< <topic>
<baseName>
<baseNameString>nameOfTheSubDirectory</baseNameString>
</baseName>
<occurrence>
<resourceRef
xlink:href="file:///fullNameOfTheFirstFileInTheSubDirectory"/>
</occurrence>
<occurrence>
<resourceRef
xlink:href="file:///fullNameOfTheSecondInTheSubDirectory"/>
</occurrence>
...
<occurrence>
< resourceRef
xlink:href="fille:///nameOfTheLastFileInTheSubDirectory"/ >
</occurrence>
</topic>



(Where fullNameOfEachFile is the name of the file with his path)



Any idea to help me to start ?



Regards



Bernard