Streamline your headache - Create Error reports while you build!

[theninjamanz 29]

This is a quick tutorial I hope will help new initiates to ubot get on quicker with building.

 

Basically the easiest way for your to test your bots is to have it create submission/error logs as the bot is run. This will mean that, if you sell your bot, you'll be able to ask customers to send you the submission log allowing you to verify what the problem is, and also allow you to understand the problems with the bots as your build them.

 

It will mean that you don't have to spend so much time clicking between scripts in the hope of being able to scroll down to the script location to see the bot code, you think has a problem, run in time.

 

In fact it's not so much error reporting here, but just getting a document of what the bot is ACTUALLY doing, compared to what THINK it should be doing. There is often a gap in between these two states that can turn a clear day into a dark and overcast one.

 

So, here's what I do.

 

First if I am building a bot that has an external list of targets - An example might be a comment "expediter" - I will set a variable to a UI (open File) called ErrorMessages.

 

With that variable set I will then create a list called "errormessages". Each time the bot completes an action, I will add a text descriptor to the list and save the file.

 

Now what's nice is that you don't just have to save static messages, you can add something dynamic. So, if you are cycling through a list or sites, you could set the URL to a varible and that would give you a way to save off the current URL and a message with it.

 

In the attachement is a bit of a workflow. You will see that If Either the words "Impossible" or "Impossibile" (I need to cover Italian bases!) are present on the page Then to the list ErrorMessages will be added the line

 

"CURRENTURL - Server too slow waiting for response"

 

In this case Current URL will be something like www.blogger.com/sdfkalewe/ or whatever.

 

In this case the IF checks to make sure that there is content on the page (Impossible, is an extract from the 404 message served by Ubots browser).

 

You might have to sketch out the exact logic of how you are going to report the errors, but you will find that it is a really great way of seing trends, rather that going completely mad (like me in the past) and analyzing things one by one.

 

Another great thing about doing things this way, is that if the bot crashes, you can have the last target the bot was working on, already saved to a list. This makes debugging much easier.

 

Hope this was helpful.

[Net66 54]

This is a quick tutorial I hope will help new initiates to ubot get on quicker with building.

 

Basically the easiest way for your to test your bots is to have it create submission/error logs as the bot is run. This will mean that, if you sell your bot, you'll be able to ask customers to send you the submission log allowing you to verify what the problem is, and also allow you to understand the problems with the bots as your build them.

 

It will mean that you don't have to spend so much time clicking between scripts in the hope of being able to scroll down to the script location to see the bot code, you think has a problem, run in time.

 

In fact it's not so much error reporting here, but just getting a document of what the bot is ACTUALLY doing, compared to what THINK it should be doing. There is often a gap in between these two states that can turn a clear day into a dark and overcast one.

 

So, here's what I do.

 

First if I am building a bot that has an external list of targets - An example might be a comment "expediter" - I will set a variable to a UI (open File) called ErrorMessages.

 

With that variable set I will then create a list called "errormessages". Each time the bot completes an action, I will add a text descriptor to the list and save the file.

 

Now what's nice is that you don't just have to save static messages, you can add something dynamic. So, if you are cycling through a list or sites, you could set the URL to a varible and that would give you a way to save off the current URL and a message with it.

 

In the attachement is a bit of a workflow. You will see that If Either the words "Impossible" or "Impossibile" (I need to cover Italian bases!) are present on the page Then to the list ErrorMessages will be added the line

 

"CURRENTURL - Server too slow waiting for response"

 

In this case Current URL will be something like www.blogger.com/sdfkalewe/ or whatever.

 

In this case the IF checks to make sure that there is content on the page (Impossible, is an extract from the 404 message served by Ubots browser).

 

You might have to sketch out the exact logic of how you are going to report the errors, but you will find that it is a really great way of seing trends, rather that going completely mad (like me in the past) and analyzing things one by one.

 

Another great thing about doing things this way, is that if the bot crashes, you can have the last target the bot was working on, already saved to a list. This makes debugging much easier.

 

Hope this was helpful.

 

Definate +1 rep point for that. Excellent advice that I am going to put to work in my own bots. Also its very easy to put together a sub that does the work so you can just pop in 'rub sub' and pass the error message as a parameter.

 

 

Andy

[UBotBuddy 331]

Dang it! You beat me! I have been doing this as well. DEFINITELY, a good thing to do. You can also search for and include other things as well. In my use, when I encounter an error I will also add to the list the tags. I also do this when things are good (in the event I need to show an inventory of my positive hits).

 

You get a +1 from me as well.

 

Buddy

[theninjamanz 29]

 

Hey Buddy you know that.

 

I tend to chuck in an additional comma and hardcode a string delimeter alongside the variable which is instructional for me.

 

such as:

 

RTY (retry)

ERR (some kind of error)

OK. (passed)

 

That way when the bot has finished, I just drop it all into Excel and can sort by classification.

 

Thanks for the Rep points, no idea what these do though.

 

[Eric 15]

Yup, very nice advise..

For me, whenever I setup a important action or task in a sub/script (like register confirmation,submission confirmation,etc), I will take into account what the site/page will show if the action is 1)successful, 2)failed. Then I will just add in the end result in a list, with the list items being the URL,success/error message,other remarks etc.. and then save these list items into a list and output to a txt file. So it will be easier to do analysis on the sites that show errors, etc.

 

eg:

URL1,success,link submitted

URL2,error,captcha error

URL3,error,email not avail

 

Maybe I should add in a 3rd error category which is the 404 error, which show up whenever a site failed to loads..