directedition - Developer IT

iptables to play nice with tor and ntpd

- by directedition

I'm setting up a server to operate as a tor relay and nothing else. I setup iptables to only allow talk on port 9001 and it worked fine, but there was an issue, the clock needs to be properly set and maintained for the relay to work properly, so I needed ntpd setup and running, but for some reason I can't get iptables to work as I want it. I'm trying to have it allow only tor and ntpd to talk over the network, but when I set it up to allow port 123 using udp, suddenly it ignores my -A OUTPUT ! -s 127.0.0.1 -j DROP and allows everything through. How should I go about this? Please excuse my ignorance, I've brand new to iptables.

Read the article

LXML E builder for java?

- by directedition

There is one thing I really love about LXML, and that the E builder. I love that I can throw XML together like this: message = E.Person( E.Name( E.First("jack") E.Last("Ripper") ) E.PhoneNumber("555-555-5555") ) To make: <Person> <Name> <First>Jack</First> <Last>Ripper</Last> </Name> <PhoneNumber>555-555-5555</PhoneNumber> </Person> As opposed to the painstaking way DOM works. I am going to be moving a bunch of my software to Java soon and it is very very heavy on its usage of E. Does Java have anything near equivalent to that usage?

Read the article

Python mechanize - two buttons of type 'submit'

- by directedition

I have a mechanize script written in python that fills out a web form and is supposed to click on the 'create' button. But there's a problem, the form has two buttons. One for 'add attached file' and one for 'create'. Both are of type 'submit', and the attach button is the first one listed. So when I select the forum and do br.submit(), it clicks on the 'attach' button instead of 'create'. Extensive Googling has yielded nothing useful for selecting a specific button in a form. Does anyone know of any methods for skipping over the first 'submit' button and clicking the second?

Read the article

Fast way to test if a port is in use using Python

- by directedition

I have a python server that listens on a couple sockets. At startup, I try to connect to these sockets before listening, so I can be sure that nothing else is already using that port. This adds about three seconds to my server's startup (which is about .54 seconds without the test) and I'd like to trim it down. Since I'm only testing localhost, I think a timeout of about 50 milliseconds is more than ample for that. Unfortunately, the socket.setdefaulttimeout(50) method doesn't seem to work for some reason. How I can trim this down?

Read the article

Programmatically adding a trusted cert in Java

- by directedition

I use SSL to communicate between two components written in Java. I can't use a CA, so I have to self-sign everything. Unfortunately, this means that when I try to handshake, I get a SunCertPathBuilderException. I can create my own X509TrustManager that just trusts everything, but that sort of defeats the purpose of having a signed cert. I would like, when first making the connection, to prompt the user with "SSL handshake with invalid cert. Add cert to store?" or something so they could have it added for them to their certificate store, like web browsers do at sites with invalid certs. I can find plenty of examples online of adding a cert to the store through the commandline, but I can't figure out how to do it programmatically. Is there a way to do this?

Read the article

Safe image uploading with PHP

- by directedition

I want my site to have a feature for users to upload images, but I want to do it safely. Namely, I want to strip the EXIF data and any malicious code that could be in their image. I was thinking of a way to do this involving direct manipulation of the file, but it struck me, would it make more sense to just convert the image they provide to BMP and then convert that back to the original format? It would suffer a generational loss, but I believe that would meet my project's requirements. BMP doesn't contain EXIF, does it? And the reprocessing should strip any malicious content.

Read the article

pyOpenSSL and the WantReadError

- by directedition

I have a socket server that I am trying to move over to SSL on python 2.5, but I've run into a snag with pyOpenSSL. I can't find any good tutorials on using it, so I'm operating largely on guesses. Here is how my server sets up the socket: ctx = SSL.Context(SSL.SSLv23_METHOD) ctx.use_privatekey_file ("mykey.pem") ctx.use_certificate_file("mycert.pem") sock = SSL.Connection(ctx, socket.socket(socket.AF_INET, socket.SOCK_STREAM)) sock.setsockopt(socket.SOL_SOCKET, socket.SO_REUSEADDR, 1) addr = ('', int(8081)) sock.bind(addr) sock.listen(5) Here is how it accepts clients: sock.setblocking(0) while True: if len(select([sock], [], [], 0.25)[0]): client_sock, client_addr = sock.accept() client = ClientGen(client_sock) And here is how it sends/receives from the connected sockets: while True: (r, w, e) = select.select([sock], [sock], [], 0.25) if len(r): bytes = sock.recv(1024) if len(w): n_bytes = sock.send(self.message) It's compacted, but you get the general idea. The problem is, once the send/receive loop starts, it dies right away, before anything has been sent or received (that I can see anyway): Traceback (most recent call last): File "ClientGen.py", line 50, in networkLoop n_bytes = sock.send(self.message WantReadError The manual's description of the 'WantReadError' is very vague, saying it can come from just about anywhere. What am I doing wrong?

Read the article

Pitfalls when converting C++/CLI to C++

- by directedition

I have a library written in C++/CLI and I want to open it up. I want it to be as cross-platform as possible and be able to write bindings to it for other languages to use (Java, Python, etc, etc). To do this, the library needs to be in plain C++ for maximum flexibility. I figure that the logical structures are already there, I just need to replace the .NET libraries it uses with the standard C++ ones. Is this a misguided notion? What should I watch out for when making this transition?

Read the article

Poll for pressed buttons in Java

- by directedition

I have a WorldWind application build based on the Java SDK. It has a great event handler for detecting when you click on objects, but I've run into a snag. While I can click on and select individual objects, I can't determine if the user is pressing the control key while they click (if they want to select multiple objects). I can implement event handlers for both the mouse and the keyboard, but I can't for the life of me figure out how to tie the two together. How could I make my mouse listener poll the system for a list of currently depressed keys?

Developer IT