I'm using VIM and I have a screwy html file.
It is in Unix format... but a stupid piece of crap Microsoft program added a bunch of carriage returns CR to it. Most of the lines are ended with a CR + LF but the microcrap proggy added an extra CR to some lines.

I just want to search and replace the lines using regular expressions in VIM. But the search fails whenever it encounters the crap that micro$oft put in.

Here's what it looks like in both ascii and hex.

<p>
<pre>

here's the hex stuff that you can't see (btw 0d = CR, 0a = LF)

<p>0d 0a 0d
<pre>

the extra '0d' at the end was inserted by the micro$oft program. Normal Unix files have just 0d 0a to end and start a line.

the regex command i used in ViM to try to find it was:
/<p>(.|\n)*<pre>

but it fails to find the pattern... therefore I can't use a regex substitution expression on it. (plan to do it on multiple files)

Does anyone know how to use the \x hexadecimal character class that Vim's regex comes with?

can you not just use dos2unix or something similar?

i am pretty sure that you can open it in pico and it will remove them then just save it

When I usually vim a windows formatted text file I see the "^M" character at the end of the line. I'm not sure why you do not see that. I use the command ":s/^V^M//g" to get rid of the windows crap. Note that the <ctrl>V will not appear, it just allows you to enter the <ctrl>M correctly.

try:

cat file | tr -d '\r'

or

fromdos < file.dos > file.unix


hth

vfs

Thanks for all the excellent help.

I found out that I was just using the OR '|' pipe incorrectly in vim. Didn't know i had to escape the pipe using \I

So my Vim regular expression search command was:
<p>\(.\|\n\)*<pre>

Vim has that magic and nomagic thing that confuses me since I use other regular expression programs.

Anyway, the above found all the extra '0d' CR's that the micro$oft program put in.

Thanks for that neat trick crabboy, with the <ctrl>+V and <ctrl>+M. That worked too.
the search pattern was <p>\(\n\|^M\)*<pre>
Using the ^V^M thing.