Convert screen log to html

Question

I would like to log actions made in a terminal window, and convert the log to html on fly.

I already tried commands script/screen + ansi2html, but the result is not perfect: the escape sequences for line editing is not handled correctly, so in the resulting html I have a mixture for the old and the new version of the line. So if the output of the terminal is like

$> echo Original text
Original text
$> echo Other
Other

The resulting html is

 $> echo Original text
 Original text
 $> echo Othernal text
 Other

Any idea, how to solve the issue?

Refer this link and remove junk char from script output http://unix.stackexchange.com/questions/86901/script-command-without-junk-character — Rahul Patil, Sep 21 '13 at 22:16
@Rahul: the link you send is interesting, but still the result is not perfect, if I reedit a line. — Attila Szlovencsák, Sep 22 '13 at 09:00
Can you Update in Question what you have tried so far and what is not working... ? — Rahul Patil, Sep 22 '13 at 09:12
@Rahul I tried the perl script ( `perl -pe 's/\e([^\[\]]|\[.*?[a-zA-Z]|\].*?\a)//g'`), on a script log, where I have rewritten the line "echo Original text" to "echo Other". In the output, I saw "echo Othernal text" — Attila Szlovencsák, Sep 22 '13 at 20:11
Can you post sample file ? you can use http://paste.ubuntu.com/ for the same — Rahul Patil, Sep 23 '13 at 10:41
See it [here](https://www.dropbox.com/s/7ka1i8ous6zrfpv/sample.zip) — Attila Szlovencsák, Sep 23 '13 at 12:31

score 0 · Answer 1 · answered Aug 26 '16 at 23:26

The problem is not your Perl script (which should be unnecessary) but ansi2html which is incomplete.

The logfile shows that something printed some text and then

repeatedly added the escape sequence for clearing the remainder of the line \E[K
alternately adding a backspace.

In ansi2html, the script finds escape sequences, and then ignores all but a few of those which it finds, i.e.,

self.ansi_codes_prog = re.compile('\033\\\[' '(\[\\d;\]*)' '(\[a-zA-z\])')

if command not in 'mMA':

Getting rid of the repeated stuff is simplest in a loop. Here is an alternate script which works around the problems seen in ansi2html for your example:

#!/usr/bin/perl -w

while (<>) {
    my $save = "";
    while ( $save ne $_ ) {
        $save = $_;
        s/[^\010]\010\e\[K//;
    }
    $save = "";
    while ( $save ne $_ ) {
        $save = $_;
        s/[^\010]\010//;
    }
    s/\e\[\?[;0-9]*[\100-\176]//g;
    s/\e\][^\a]*\a//g;
    print;
}

and (calling that filterit), use it like this:

./filterit < 1.log | ansi2html >1.html

Convert screen log to html

1 Answers1

Linked