Hello,
I am writing a tool to package several files into a single file. It is
quite similar to the 'phar' PEAR package but with more features, as I
want to connect it with an autoload mechanism. The problem I have today
and it is the same in the PHAR package, is that I don't know how to
restrict the PHP parser to the bootstrap code of the package file.
I explain : As in the shar format, the idea is to put some bootstrap
code first, and the rest of the data after it. The problem is that I
don't want PHP to consider any '<?php' string appearing in the data as
some PHP code during the initial phase. An example : if I have
something like :
<?php Bootrap code ?> ...Anything... <?php ... Anything...
In this case, I want the 2nd '<?php' to be considered as data and I
don't want the tokenizer to go and look inside this section before
executing the bootstrap code. I have the possibility to replace any
'<?' string by another one, say '</?' and then translate them when they
are extracted, but it is not easy to do because it changes every file
offsets ! And, of course, I reference the files through their offsets
in the package file.
And compressing the files is not the perfect solution because :
1. I want the compression to be optional for each file in the archive
(and also the compression method)
2. and the result of the compression process can contain a '<?' string,
which would automatically lead to a syntax error (with short tags
enabled).
I tried to put a chr(0) after my code but PHP doesn't care. If there is
a string or something allowing to stop the tokenizer at a given point
in a script, or if you have a smarter idea, please tell me.
Thanks in advance.
François
I am writing a tool to package several files into a single file. It is
quite similar to the 'phar' PEAR package but with more features, as I
want to connect it with an autoload mechanism. The problem I have today
and it is the same in the PHAR package, is that I don't know how to
restrict the PHP parser to the bootstrap code of the package file.
I explain : As in the shar format, the idea is to put some bootstrap
code first, and the rest of the data after it. The problem is that I
don't want PHP to consider any '<?php' string appearing in the data as
some PHP code during the initial phase. An example : if I have
something like :
<?php Bootrap code ?> ...Anything... <?php ... Anything...
In this case, I want the 2nd '<?php' to be considered as data and I
don't want the tokenizer to go and look inside this section before
executing the bootstrap code. I have the possibility to replace any
'<?' string by another one, say '</?' and then translate them when they
are extracted, but it is not easy to do because it changes every file
offsets ! And, of course, I reference the files through their offsets
in the package file.
And compressing the files is not the perfect solution because :
1. I want the compression to be optional for each file in the archive
(and also the compression method)
2. and the result of the compression process can contain a '<?' string,
which would automatically lead to a syntax error (with short tags
enabled).
I tried to put a chr(0) after my code but PHP doesn't care. If there is
a string or something allowing to stop the tokenizer at a given point
in a script, or if you have a smarter idea, please tell me.
Thanks in advance.
François
Comment