[tinkeruntilitworks]

if there were no tracking and no chance of malware a reasonable amount of ads would be fine

until that day

lol

[SurprisedItWorks]

Confession: going back to edit the last posted version just a wee bit to eliminate the stupid use of cat from the big pipeline. Works fine, but it's pointless, as are most cats in pipelines. When you are tempted to have a cat feed a pipe, stop and think.

[Alozaros]

well, back to the thing that put me off...
first my sbc.io has most of the lines needed...and according to my script it Dl about 56k lines...so...
i print it out and i can see any other lines than 0.0.0.0...there is nothing alien...with your script
it trims down to 36k lines and i do not see 20k in trimmings, if you know what i mean..yep many lines are missing...for 20k reduction this is almost a half
not to mention that if i use 2 sources than all the lines are not more than 39k to be precise...

[SurprisedItWorks]

[Alozaros]

yep sry now it shows 42286
you are wright and this is with my script
few days ago it was showing 56031 so it vary's
there was a something with your script i use a very short version of it

( cd /tmp; touch badhosts; sleep 30
( wget -O - http://winhelp2002.mvps.org/hosts.txt
wget -O - http://sbc.io/hosts/hosts ) 2>badhosts.log \
| sed 's/\t/ /g; /^0\.0\.0\.0 /!d; s/ *\#.*$//; s/\r//' | sort -u \
| sed -E '/\.hulu(|ad)\./d' > badhosts
) && killall -HUP dnsmasq &

now i installed curl so i can use both

anyways my mistake i ll stick to my script until you
finalize yours, i cant play much with it as im not around the router, and i cant leave my colleges without internet..

[SurprisedItWorks]

Edited 25 September 2019 (date of my post below pointing out the correction) to add the -f flag to the curl, to provide correct behavior when a download site doesn't exist at the URL given.

I think functionally my script is stable, but I did do a cleanup pass on the code this morning to make it more compact on the page and so use less startup-code storage, to make it easier to add or remove download targets, and to make it easier to switch from curl -k to wget -O - with the latter now only a matter of doing exactly that substitution in exactly one place.

Of course by now everyone has noticed that I keep thinking that tweaking is done, then I think of a new tweak.

[Alozaros]

this one does some strange errs, i got only one line facebook bla bla bla in cat tmp/badhosts..
it says err code 527 or something like..
tried with curl and wget too...

[SurprisedItWorks]

[Alozaros]

user.notice startup-code adblocker: badhosts has 1 lines, exit codes 127 127

cat /tmp/dlhosts.codes 127 127
/tmp/custom.sh: line 11: curl: not found
/tmp/custom.sh: line 11: curl: not found

hmm it seems this curl i installed via opkg is
not working as it should..

and weget -O does this
user.notice startup-code adblocker: dlhosts has 1 lines, exit codes 255 255

[SurprisedItWorks]

[Edit: I responded to the below too quickly. Actually, the fact that you had two wget error codes in the log points to wget being run twice. The only way that could happen is if using the function is fine, since the function is called twice. So check carefully the use of wget. Careful that "-O" is followed, after a space, by "-" since omitting that would certainly crash things.]

[Alozaros]

yep some builds are like that things are missin...
and this is 8MB flash size device, so it's normal...
some things are stripped off in order to fit in the flash size...

no output of that function, so this is the bugger

it says bbx 1.31 built-in shell (ash) ...


yep im ok with the old script..or yours...the short version of it, its also fine...

[SurprisedItWorks]

Bug fix!
I'm editing it into the earlier posted versions of the script in this thread.

Replace "curl -k" in the adblocker with "curl -kf" so that if the site doesn't exist, a nonzero exit code will be returned and no stdout will be produced. I mistakenly assumed curl did this anyway, like wget did in my early versions. Turns out it doesn't though. Without the -f a nonexistent site makes it return a zero exit code and send an "OMG THERE'S NO SUCH SITE?" or some such (OK, with more conventional wording), in html, to stdout. The sed commands will filter out the nonsense output, so it shouldn't contaminate the badhosts file. But the error codes will not show anything is wrong.

[SurprisedItWorks]

Edited 9 Dec 2019 to add wait loop before curls, to make sure the network is available, and an outer subshell (first two noncomment lines and a final closing parenthesis) that can be exited early if a lockfile shows it's already been run. You likely won't need either tweak in a shell version that you run with cron.

So while I am in a fixing mood, and continuing my idiotic tradition of posting new versions way too soon and after minimal testing of changes (a dd-wrt tradition!), here's the latest adblocker code for GUI>Administration>Commands, Startup. Note the first comment re the other GUI edit required.

The main point here is just cleanup for readability. However, there are actual changes here to that earlier hurried kluge of syslog-entry machinery. This version is cleaner. Pointless group-command braces in the Download function have been removed as well. While I've badmouthed cats in pipelines, even in this very thread, I don't actually see a cleaner way here (in the syslog-entry code). So there it lurks.

[pikasso]

Hi there..
SurprisedItWorks.. I registered today to ddwrt just to say thank you.
This method is so simple and automatic, no need of flash drive, tuning something in addition, and just simply works.
I must say I had a lot of fun watching the posts updates, from one method then using curl.. then wget.. then optimizing the sequences.. really perfect mindset and script.

Thank you !
Pierrick

[SurprisedItWorks]

Thank you, Pierrick! Unusual that people bother with that. It is appreciated.

And so here's a bonus, something I just added to my routers a couple of weeks ago. Add this as a new line after the sleep and before the first sed:

until ping -c 1 -w 10 &> /dev/null cloudflare.com; do sleep 110; done

This prevents things from going further until the router is able to ping cloudflare.com. Tries the ping every two minutes.

I was moved to add this when I had a power failure at home for an hour, and when the power came back up, the router booted before the modem. This meant the router was trying to do the Download steps before success was possible, and as a result all the Downloads failed and badhosts remained empty. This way it won't try the downloads until the network is up with DNS functioning.