feature request: add progress bar for blockcheck.sh

[pohmelie]

I'm watching some probes for a hour and I don't know how long will it goes.

[bol-van]

It's not easy.

Check process depends on previous results.
For example, it will not test "fake,split" if just "fake" works.

[pohmelie]

If you can calculate total possible number of requests, then you can make progress on this. No need to be very precise it just for user to understand is it a minute or is it an hour. If some requests are skipped progress just will be faster.

[bol-van]

Theoretically it's possible but if you look at pktws_check_domain_http_bypass function you will see how many conditions and branches are there.
For every code I need to maintain another code that counts how many tests first code launches.
And keep them synchronized after any change.

I dont think this complexity worth it

[bol-van]

If you want to speed up execution there're some possibilities

SKIP_TPWS=1 CURL_MAX_TIME=1 PARALLEL=1 ./blockcheck.sh

If you set >1 attempts PARALLEL launches many attempts at once (can trigger ddos protection)

[pohmelie]

This problem is known well. You just need to collect all tasks descriptions at first step, and then iterate over it evaluating them. So, before actual requests you will know the total amount of requests

[bol-van]

Yes. But it's not enough to count total maximum tests.
I need to profile every branch how many max tests if would launch. So if a branch is not executed I know by what value to increase current test counter.

[pohmelie]

Hm, I don't get it. You can't without any requests build a total list of required requests, is it?

[bol-van]

I can dry-run it with silent fail and count total possible maximum that would be with force mode..
But I need to know current test number at every test to display it correctly.
This is not trivial. If all test would execute I just increase counter every test.
But if not ?

[pohmelie]

I mean, you can assume that every possible test will be evaluated and start from this count. Then, when you decide to skip test, you will mark this test as done in progress.

[bol-van]

Logic of test skipping is not straightforward.
See the function I mentioned.

[pohmelie]

Yes, I understand that in current implementation this can be hard or impossible to implement and probe code needs to be refactored.