Bash too many arguments error

Shell Programming and Scripting
1

Hello, I've got this little script that gets a bunch of comics and puts them in an html file. However, when I check that the comic URL is there and that it's a new one, I get the "too many arguments" error. The script is this:

CADURL=`curl --silent http://www.cad-comic.com/cad/ | grep -o 'http://cdn.cad-comic.com/comics/.*.jpg'`
COMICJKURL=`curl --silent http://comicjk.com/comic.php/latest | grep -o '../Pics/.*\.gif' | sed 's|\.\.|http://comicjk.com|g'`
BUGURL=`curl --silent http://www.bugcomic.com/ | grep -o 'http://www.bugcomic.com/comics/.*.png'`
XKCDURL=`curl --silent http://www.xkcd.com/ | grep -o 'http://imgs.xkcd.com/comics/.*.png' | head -1`
DILBERTURL=`curl --silent http://www.dilbert.com/ | grep -o 'http://dilbert.com/dyn/str_strip/.*/.*/.*/.*/.*/.*/.*/.*/.*\.strip\..*\gif'`
CYANIDEURL=`curl --silent -L http://www.explosm.net/comics/new/ | grep -o 'http://www.explosm.net/db/files/Comics/.*/.*.png'`
HIJINKSURL=`curl --silent http://hijinksensue.com/ | grep -o 'http://hijinksensue.com/comics/.*.jpg'`

HEADER="<html>\n<head>\n\t<title>Home</title>\n</head>"
CADHTML='<FONT FACE="Lucida Grande" SIZE=20><br>CTRL-ALT-DEL<br></FONT>\n<img style="padding: 5px\; border: 1px solid \#000000\;" src="CADURL">'
COMICJKHTML='<FONT FACE="Lucida Grande" SIZE=20><br>ComicJK<br></FONT>\n<img style="padding: 5px\; border: 1px solid \#000000\;" src="COMICJKURL">'
BUGHTML='<FONT FACE="Lucida Grande" SIZE=20><br>Bug<br></FONT>\n<img style="padding: 5px\; border: 1px solid \#000000\;" src="BUGURL">'
HIJINKSHTML='<FONT FACE="Lucida Grande" SIZE=20>HijiNKS ENSUE<br></FONT>\n<img style="padding: 5px\; border: 1px solid \#000000\;" src="HIJINKSURL">'
DILBERTHTML='<FONT FACE="Lucida Grande" SIZE=20><br>Dilbert<br></FONT>\n<img style="padding: 5px\; border: 1px solid \#000000\;" src="DILBERTURL">'
CYANIDEHTML='<FONT FACE="Lucida Grande" SIZE=20><br>Cyanide & Happiness<br></FONT>\n<img style="padding: 5px\; border: 1px solid \#000000\;" src="CYANIDEURL">'
XKCDHTML='<FONT FACE="Lucida Grande" SIZE=20><br>xkcd<br></FONT>\n<img style="padding: 5px\; border: 1px solid \#000000\;" src="XKCDURL">'

echo -e "$HEADER" > dailycomic.html
if [ $XKCDURL != "" ] $$ [ "$XKCDURL" != `cat .previouscm.txt | grep $XKCDURL` ]; then
	echo -e "$XKCDHTML" | sed "s|XKCDURL|$XKCDURL|" >> dailycomic.html
fi
if [ $CYANIDEURL != "" ] $$ [ $CYANIDEURL != `cat .previouscm.txt | grep $CYANIDEURL` ]; then
	echo -e "$CYANIDEHTML" | sed "s|CYANIDEURL|$CYANIDEURL|" >> dailycomic.html
fi
if [ $BUGURL != "" ] $$ [ $BUGURL != `cat .previouscm.txt | grep $BUGURL` ]; then
	echo -e "$BUGHTML" | sed "s|BUGURL|$BUGURL|" >> dailycomic.html
fi
if [ $COMICJKURL != "" ] $$ [ $COMICJKURL != `cat .previouscm.txt | grep $COMICJKURL` ]; then
	echo -e "$COMICJKHTML" | sed "s|COMICJKURL|$COMICJKURL|" >> dailycomic.html
fi
if [ $DILBERTURL != "" ] $$ [ $DILBERTURL != `cat .previouscm.txt | grep $DILBERTURL` ]; then
	echo -e "$DILBERTHTML" | sed "s|DILBERTURL|$DILBERTURL|" >> dailycomic.html
fi
if [ $HIJINKSURL != "" ] $$ [ $HIJINKSURL != `cat .previouscm.txt | grep $HIJINKSURL` ]; then
	echo -e "$HIJINKSHTML" | sed "s|HIJINKSURL|$HIJINKSURL|" >> dailycomic.html
fi
if [ $CADURL != "" ] $$ [ $CADURL != `cat .previouscm.txt | grep $CADURL` ]; then
	echo -e "$CADHTML" | sed "s|CADURL|$CADURL|" >> dailycomic.html
fi

I have googled for the fix for that error but I was unable to find anything that helped.

2

What is in .previouscm.txt ? Does it exist what is the result of such a grep?

[ "$XKCDURL" != "`cat .previouscm.txt | grep $XKCDURL`" ]

In any case I would use double quotes to make sure the test has a single argument.

3

The problem is you're using too many arguments. There are limits on argument lists and variable lengths, and if anything bash is unusually generous. You should keep both below a couple kilobytes in size.

This means statements that slurp in and/or compare entire massive blocks of data, like:

DILBERTURL=`entirety of everything`
$DILBERTURL != `cat .previouscm.txt | grep $DILBERTURL`

...should be avoided. Save the results directly into files with redirection, then use the diff command to see if there's any changes instead.

4

Scrutinizer, Well until this works, .previouscm.txt is just a text file with random letters and numbers. Once this script works, I will add a few lines at the bottom writing all of the current url's to the text file, so if the URL appears in the text file it will do nothing.
When I grep it, nothing is returned, as shown in bash debug:

'[' http://imgs.xkcd.com/comics/audiophiles.png '!=' '' ']' 98109 '[' http://imgs.xkcd.com/comics/audiophiles.png '!=' '' ']'

I also tried the double quotes, didn't help.

Corona688, The Dilbert URL is 113 character, http://dilbert.com/dyn/str_strip/000000000/00000000/0000000/100000/00000/9000/300/109351/109351.strip.sunday.gif, and that is too large?
If so, I should change the code to output each url to separate files and have the old urls in separate files and diff them?

5

Oh. No, it's not. All the backticking and cat-ing and grep-ing had me thinking you were dealing with an entire list of comics.

6

Ah, ok. Damn, the error is still unexplained.

7

You have $$ in the if statements -- did you want && or || .

8

What are all those "$$" for? Did you mean that to be && or ||

It produces the comic list when I replace them with &&.

9

Well look at that. How stupid of me. That did it, thanks.

10

It LOOKS normal at first glance. :slight_smile: I didn't notice it until a few rereads.

11

Me neither :slight_smile: