Multi-line regex

saty has asked for the wisdom of the Perl Monks concerning the following question:

Hello monks, I have a small problem with multi-line regexs, I had posted this question before but the code I posted was incomplete. I am performing this multi-line regex on a C map file. Here is the perl code

#!/usr/local/bin/perl -w
# analyse_map.perl
# Generate a snapshot of audio registers using register log file,
# Open map files (map_merc and rom.dld)
open(MAP, "<map_merc") || die "Unable to open map_merc";
# Read map_merc :
while(<MAP>)
{
  if( ($titi, $toto) = /^\.(\w+).+$\^\t+\.\w+\W+(\w+)/m)
  {
    print($titi, $toto, " ", $., "\n");
  }
}
close(MAP);
[download]

The target I am trying to scan for data is given below (just a dew lines as the file is too big)


output        input        virtual
section        section        address        size     file

.begin                  00002000    00000068
        .begin          a0000000    00000068 begin_flash.o
             _START a0000000    00000000


.copy_flash             00000000    00001488
        .copy_flash     a0001000    000001ac libbsp_qt1_aspic32.a[copy
+_flash.o]
        copy_from_flash a0001000    00000000

        .copy_flash     a00011ac    00000978 libbsp_qt1_aspic32.a[cach
+3940.o]
           SysDisableDCache a00011ac    00000000
               disableD a0001220    00000000
           SysDisableICache a0001294    00000000
               disableI a0001308    00000000
        SysEnableDCache a000137c    00000000
            enableD a00013f8    00000000
        SysEnableICache a0001464    00000000

.text                   00003000    00052be4
        .boot           80003000    0000009c libbsp_qt1_aspic32.a[boot
+.o]
            HdwInit 80003000    00000000
            end_SIF 80003080    00000000

.endtext                80055be4    00000000
        .endtext        80055be4    00000000 end.o
            TextEnd 80055be4    00000000


.data                   00180000    000004d0
        .data           80201000    00000000 begin_flash.o
              StartData 80201000    00000000
[download]

The thing I am trying to do is scan the line that begins with "." eg .begin etc and put it in the scalar $titi and then if there is a line just below which starts with a tab and then ".", get the second part which is the address. I hope this explaination is clear enough. If you need more explaination mail me at satyajeet.navalkar@philips.com Thanx. Saty

Comment on Multi-line regex Select or Download Code

Replies are listed 'Best First'.
RE: Multi-line regex by Corion (Patriarch) on Aug 28, 2000 at 12:32 UTC
Please use `<CODE>` to start your code and `</CODE>` to end your code postings. Your regex in question is : `/^\.(\w+).+$\^\t+\.\w+\W+(\w+)/m` [download] This looks like a regular expression that was thought out with a small error and then mangled until Perl would accept it. The `^` and `$` special characters in a regular expression do not mean 'start of line' (and 'end of line') in multiline context, but 'start of string' (and 'end of string'). The backslashes before them quote them as literal characters, which is not at all what you are looking for. If you are using the `/m` multiline mode, you can embed `\n` newline metacharacters into the string right away : `/(?:^\|\n)\.(\w+).+\n\t+\.\w+\W+(\w+)/m` [download] I didn't test this RE, but it looks more like what you described. Here's a breakdown of the RE : `/(?:^\|\n) # Start matching either at the start of the string # or at a newline. The ?: part means, that # the parentheses I use don't get saved in $1 \.(\w+).+\n # match a line starting with a dot-word and some # more stuff \t+\.\w+ # The line after this must start with at least one # tab and then a dot-word \W+(\w+) # and contain some other stuff as well ... /mx` [download] I hope that helps you a bit.	[reply] [d/l] [select]
Re: Multi-line regex by merlyn (Sage) on Aug 28, 2000 at 14:40 UTC
Besides the above comment, I also didn't see anything even in your mangled code that reads the file. Must have that. And if you expect more than one line to be scanned by a regex, you have to read a bunch of lines: `open FOO, "somewhere" or die "cannot open somewhere: $!"; $_ = join "", <FOO>; if (/ (?:^\|\n) # Start matching either at the start of the string # or at a newline. The ?: part means, that # the parentheses I use don't get saved in $1 \.(\w+).+\n # match a line starting with a dot-word and some # more stuff \t+\.\w+ # The line after this must start with at least one # tab and then a dot-word \W+(\w+) # and contain some other stuff as well ... /mx ) { ... you found one ... }` [download] Something like that. -- Randal L. Schwartz, Perl hacker	[reply] [d/l]