[RESOLVED] preg_replace rtf

[dagon]

for the code below I'm trying to clean out the rtf, i can get it to work in an online regex engine ut not with php, im confused

$rtf="{\rtf1\ansi\ansicpg1252\deff0\deftab1134{\fonttbl{\f0\fswiss\fprq2\fcharset0 Arial;}{\f1\froman\fprq2\fcharset0 Times New Roman;}}
\viewkind4\uc1\pard\sb100\sa100\tx720\tx1440\tx2160\tx2880\tx3600\tx4320\tx5040\tx5760\tx6480\tx7200\tx7920\tx8640\tx9360\tx10080\lang5129\f0\fs20 Accounting and Consulting services for the period ended XX March XXXX. In Particular;\f1\fs24\par
\f0\fs20 Finalisation of XXX financial statements and preparation and finalisation of XXX taxation return. Arranging for taxation return to be filed with the XXX. \f1\fs24\par
}
";

$matches = array('/({\\)(.+?)(})/','/(\\)(.+?)(\b)/');

echo preg_replace($matches,'',$rtf);

preg_replace() [<a href='function.preg-replace'>function.preg-replace</a>]: Compilation failed: missing ) at offset 12

[jgetner]

i found that PCRE has trouble with symbols used in php, html, javascript becuase but i have heard that converting to binary fixes the problem.(not sure!)

[Weedpacket]

The main hassle is that if you want to match a literal backslash you end up needing to escape it for both PHP and PCRE - especially if it's followed by another PCRE-significant character. So [/b] becomes \ so that PCRE recognises it as a literal backslash instead of an escape; but PHP's string parsing will see that as an escaped single backslash and only pass one of them on to PCRE - unless you escape them. Upshot is that you end up with \\.

Some of your parentheses (the outer pairs in both expressions) are redundant; the inner pairs are probably unnecessary also.

Ergo,

$matches = array('/\{\\\\(.+?)\}/','/\\\\(.+?)\b/'); 

Another ways around it would be to use chr(92) to sneak the backslash past PHP; using \x5c instead; or using \Q..\E to get PCRE to treat stuff between them literally (no metacharacter behaviour).

The other thing that's likely to be causing you grief (at least in what is posted) is that $rtf is double-quoted - so PHP will be parsing all those \r and \t sequences before the string even gets into the variable.

Maybe a more format-aware parser might be simpler to use (e.g., that {\fonttbl{...}{...}} can go, I'm guessing); I'm too rusty with rtf, but it occurs to me that using preg_replace_callback to match a more straightforward pattern that gets refined further in the callback function might be something worth tackling.

[dagon]

Thanks a lot Weedpacket that has me moving in the right direction. I did spend most of the day trying to find a php (non windows) way to convert the rtf to html, but gave up in the end and switched t stripping it of.

D.