Scratch

Pitusky12 wrote:

bobbybee wrote:

I was the one that mentioned that security hole, right?

Hello,sorry for posting this here but im looking for the best scratchers in history,could you join my collab heres a link to it: http://scratch.mit.edu/forums/viewtopic.php?pid=1363435#p1363435
Post there.

Hey, around here, posting offtopic advertising is considered spam. Perhaps you should contact them on one of their projects or something. 

Hey, sparks, could you update the main goals you have for 2.0 on the original post? That way other people can see right off the bat what this is about. And I'm kinda lazy so I don't want to go through 14 pages to catch up. 

I just had an idea!

I presume a very large portion of the images created on the site will be fairly process-expensive to generate (PHP's image library isn't very efficient and we do allow a lot of resizing, moving, etc).

We have no control over how many times a day these images would be generated, and for frequent posters with many posts scattered around highly viewed forums, these requests could become too many for the servers to handle.

Therefore, at the cost of some disk space (which we don't lack), we could save the images as PNGs and have the viewer fetch those instead of computing them at each request.

Even keeping our servers on 000wh, I expect to be able to allocate the space for at least one hundred thousand images (and leaving 500MB left over for stuff). What do you think?

EDIT
I should also mention that if we do exceed CPU limits on our current host, we're getting the site shut down without database or file recovery.

EDIT2
I completely forgot though, the whole point of a self updating image 
I guess we could optimise it, but for now we would have to leave it like this.

Last edited by LS97 (2012-08-21 10:51:36)

Yay! I finished the parser/viewer of images!
fg123, you can test it manually if you wish by editing the elements table in phpmyadmin. I have thought of security and error handling too.

I implemented a censoring mechanism. Currently, no bad words are listed (to keep the code "clean") and two words are white listed: grape and skyscraper 

URLs that lead to external images are also briefly checked for format and are replaced by a local "bad URL" image if they don't meet my reg exp.

size and position are replaced by defaults if bad format, and font name, size, etc do likewise.

***

Now, I will try myself to make a (maybe non-drag n drop) image editor so we can start using and testing the API.

I have two questions about the viewer though. First of all, what fonts should we support, and how should we implement the censoring? will we have the words right in the code, or will we fetch them from a database table?

Neat! There's been some good work here recently and I'm so sorry I don't have the time atm to help :'(

There was talk on the last page about me updating the first post to show a link to the site and some key aims of the project. I've done an update to the first post now but please feel free to suggest changes to it!

sparks wrote:

Neat! There's been some good work here recently and I'm so sorry I don't have the time atm to help :'(

There was talk on the last page about me updating the first post to show a link to the site and some key aims of the project. I've done an update to the first post now but please feel free to suggest changes to it!

Nice! I think it neatly sums it up. However the bold tag didn't work.

Mmhm, would you like the registering code as well?

Also, I'll be taking a look at your changes now, looks promising. 

Why is the creation date in Unix and not the standard mysql datetime?

Looks really nice! How did you do the edit/12, is it making a file each time or using Apache rewrite?

Last edited by fg123 (2012-08-27 12:54:14)

technoboy10 wrote:

fg123 wrote:

Mmhm, would you like the registering code as well?

Also, I'll be taking a look at your changes now, looks promising. 

Why is the creation date in Unix and not the standard mysql datetime?

Looks really nice! How did you do the edit/12, is it making a file each time or using Apache rewrite?

Are you talking to me? 

Yea, the register code part.

Ah, basically, you store all the user details in the database when they register, and have it check the database to see if the user is in the database yet, if so, check their password. Since we are using MD5 as a hashing tool, passwords stored in the database are in MD5, therefore, we need to convert their entered password against the MD5 in the database. If you would like to do the authentication now, just perform it when the user is not found or the password is incorrect, as that means the user hasn't already been registered.

I try making a user account there and it says: username cannot include special characters
but i didn't use any!

LS97 wrote:

iTweak0r wrote:

I try making a user account there and it says: username cannot include special characters
but i didn't use any!

Maybe a 0 includes a special character?

i just tried "itweak" and it still didn't work!

iTweak0r wrote:

LS97 wrote:

iTweak0r wrote:

I try making a user account there and it says: username cannot include special characters
but i didn't use any!

Maybe a 0 includes a special character?

i just tried "itweak" and it still didn't work!

I got the same error, no useful explanation.

Tried clearing cache, cookies, and using a simpler username and password than I had been, but it didn't work.  I suppose I'm a special character  .

Last edited by LiquidMetal (2012-10-07 11:24:23)