Comments for Staccato Signals http://artins.org/ben Just another WordPress weblog Sat, 04 Jul 2009 10:26:44 +0000 http://wordpress.org/?v=2.3.3 Comment on How to create a multihomed certificate with OpenSSL by Colum Paget http://artins.org/ben/how-to-create-a-multihomed-certificate-with-openssl#comment-2611 Colum Paget Tue, 09 Jun 2009 11:12:20 +0000 http://artins.org/ben/how-to-create-a-multihomed-certificate-with-openssl#comment-2611 <p>Thanks for this tip! I'm hoping that I can use this with outlook so that people who access a machine by its different IP's on different networks can all avoid being told 'certificate name does not match the server'.</p> Thanks for this tip! I’m hoping that I can use this with outlook so that people who access a machine by its different IP’s on different networks can all avoid being told ‘certificate name does not match the server’.

]]> Comment on Converting an STL vector iterator to a raw pointer by Philip Bliss http://artins.org/ben/converting-an-stl-vector-iterator-to-a-raw-pointer#comment-1985 Philip Bliss Thu, 26 Feb 2009 14:03:05 +0000 http://artins.org/ben/converting-an-stl-vector-iterator-to-a-raw-pointer#comment-1985 <p>After doing slightly more research on this matter, I realized that v.front() is undefined for empty vectors--gcc seems to give a null pointer in that case, but the syntax &v.at(0) is preferable as it gives a defined result for empty vectors--an out<em>of</em>range exception, which is useful for debugging.</p> <p>As for the one-past-the-end pointer, I suppose &v.at(v.size() - 1) + 1 would work, although it looks quite inelegant.</p> After doing slightly more research on this matter, I realized that v.front() is undefined for empty vectors–gcc seems to give a null pointer in that case, but the syntax &v.at(0) is preferable as it gives a defined result for empty vectors–an outofrange exception, which is useful for debugging.

As for the one-past-the-end pointer, I suppose &v.at(v.size() - 1) + 1 would work, although it looks quite inelegant.

]]> Comment on Converting an STL vector iterator to a raw pointer by Philip Bliss http://artins.org/ben/converting-an-stl-vector-iterator-to-a-raw-pointer#comment-1984 Philip Bliss Thu, 26 Feb 2009 13:46:49 +0000 http://artins.org/ben/converting-an-stl-vector-iterator-to-a-raw-pointer#comment-1984 <p>I use &v.front() to get a pointer to the first element, which can be used to pass vectors to functions that expect dynamic arrays. The nice thing about that is that it automatically returns 0 for empty vectors, so no check is required.</p> <p>In the case of an empty vector, &v.front() + (it - v.begin()) would generate an invalid pointer, probably giving a segmentation fault, which is OK, since the behaviour of a dynamic array here would be to silently provide access to an incorrect memory address, possibly leading to memory corruption.</p> <p>For the pointer-past-the-end-of-the-vector, &v.back() + 1 seems like a good idea.</p> I use &v.front() to get a pointer to the first element, which can be used to pass vectors to functions that expect dynamic arrays. The nice thing about that is that it automatically returns 0 for empty vectors, so no check is required.

In the case of an empty vector, &v.front() + (it - v.begin()) would generate an invalid pointer, probably giving a segmentation fault, which is OK, since the behaviour of a dynamic array here would be to silently provide access to an incorrect memory address, possibly leading to memory corruption.

For the pointer-past-the-end-of-the-vector, &v.back() + 1 seems like a good idea.

]]> Comment on Converting an STL vector iterator to a raw pointer by Ben Artin http://artins.org/ben/converting-an-stl-vector-iterator-to-a-raw-pointer#comment-1918 Ben Artin Tue, 03 Feb 2009 17:18:58 +0000 http://artins.org/ben/converting-an-stl-vector-iterator-to-a-raw-pointer#comment-1918 <p>You should keep in mind that your code copies the buffer at pPointer into a vector. If you change the vector, memory at pPointer will not change, and vice versa.</p> You should keep in mind that your code copies the buffer at pPointer into a vector. If you change the vector, memory at pPointer will not change, and vice versa.

]]> Comment on Converting an STL vector iterator to a raw pointer by Eyal Itskovits http://artins.org/ben/converting-an-stl-vector-iterator-to-a-raw-pointer#comment-1917 Eyal Itskovits Tue, 03 Feb 2009 14:51:40 +0000 http://artins.org/ben/converting-an-stl-vector-iterator-to-a-raw-pointer#comment-1917 <p>I've came accross a oposite problem in which I had to convert from a raw pointer to a STL iterator.</p> <p>However, because you can USE raw pointer as random access iterator, I could do the following - </p> <p>BYTE *pPointer; std::vector myVec(pPointer, pPointer + bufferSize);</p> <p>And myVec.begin() will represent the pointer.</p> <p>Eyal.</p> I’ve came accross a oposite problem in which I had to convert from a raw pointer to a STL iterator.

However, because you can USE raw pointer as random access iterator, I could do the following -

BYTE *pPointer; std::vector myVec(pPointer, pPointer + bufferSize);

And myVec.begin() will represent the pointer.

Eyal.

]]> Comment on How to create a multihomed certificate with OpenSSL by Jon http://artins.org/ben/how-to-create-a-multihomed-certificate-with-openssl#comment-512 Jon Tue, 22 Apr 2008 15:14:31 +0000 http://artins.org/ben/how-to-create-a-multihomed-certificate-with-openssl#comment-512 <p>I do what you are doing and have a guy using Safari who is getting errors. Perhaps Safari tried to fix the potential exploits with the altnames but fixed it badly?</p> <p>I include the web page link above.</p> I do what you are doing and have a guy using Safari who is getting errors. Perhaps Safari tried to fix the potential exploits with the altnames but fixed it badly?

I include the web page link above.

]]> Comment on Converting an STL vector iterator to a raw pointer by Daedalus http://artins.org/ben/converting-an-stl-vector-iterator-to-a-raw-pointer#comment-333 Daedalus Fri, 28 Mar 2008 01:32:40 +0000 http://artins.org/ben/converting-an-stl-vector-iterator-to-a-raw-pointer#comment-333 <p>Scouring the net led me to this page, which perfectly explained what I was looking for and how to acheive what I need. Thanks!</p> Scouring the net led me to this page, which perfectly explained what I was looking for and how to acheive what I need. Thanks!

]]> Comment on How to create a multihomed certificate with OpenSSL by Ben Artin http://artins.org/ben/how-to-create-a-multihomed-certificate-with-openssl#comment-22 Ben Artin Wed, 13 Feb 2008 21:56:35 +0000 http://artins.org/ben/how-to-create-a-multihomed-certificate-with-openssl#comment-22 <p>Thanks, Chris. With my certs, I didn't happen to run into the 64-character limit, but now I know what to do if I need to add more hostnames to my certs.</p> Thanks, Chris. With my certs, I didn’t happen to run into the 64-character limit, but now I know what to do if I need to add more hostnames to my certs.

]]> Comment on How to create a multihomed certificate with OpenSSL by Chris Snyder http://artins.org/ben/how-to-create-a-multihomed-certificate-with-openssl#comment-20 Chris Snyder Wed, 13 Feb 2008 20:34:31 +0000 http://artins.org/ben/how-to-create-a-multihomed-certificate-with-openssl#comment-20 <p>I wonder if you've ever run into problems with a CN that's too long? My openssl choked when I tried to list too many servers, apparently it wants a CN that is 64 characters or less.</p> <p>But it seems that maybe browsers don't care about the CN provided that the hostname is listed in one of the subjectAltNames.. or at least, MSIE7, FF2, and Safari3 don't. In other words, I have a working multihomed certificate with a single CN. </p> <p>Thought it might help others to know this, and maybe save you some work when you create new certificates.</p> I wonder if you’ve ever run into problems with a CN that’s too long? My openssl choked when I tried to list too many servers, apparently it wants a CN that is 64 characters or less.

But it seems that maybe browsers don’t care about the CN provided that the hostname is listed in one of the subjectAltNames.. or at least, MSIE7, FF2, and Safari3 don’t. In other words, I have a working multihomed certificate with a single CN.

Thought it might help others to know this, and maybe save you some work when you create new certificates.

]]> Comment on Brrr, thought the goose by Ben Artin http://artins.org/ben/brrr-thought-the-goose#comment-6 Ben Artin Sun, 30 Dec 2007 16:48:43 +0000 http://artins.org/ben/brrr-thought-the-goose#comment-6 <p>Whoops — fixed. Thanks!</p> Whoops — fixed. Thanks!

]]> Comment on Brrr, thought the goose by Leah http://artins.org/ben/brrr-thought-the-goose#comment-5 Leah Sun, 30 Dec 2007 06:20:49 +0000 http://artins.org/ben/brrr-thought-the-goose#comment-5 <p>It's a lovely picture, but I should point out that the bird standing in the middle of the photo is in fact a Canada Goose. I should know, being from Canada and all. :)</p> It’s a lovely picture, but I should point out that the bird standing in the middle of the photo is in fact a Canada Goose. I should know, being from Canada and all. :)

]]>