Rah for The Gate

[ciphergoth]

Lovely meal with my Mum and my lovers at The Gate last night. The food was fabulous just as everybody said - the potato cakes from the mezze starter particularly stand out - and the company sparkling and lovely as you would expect. It was particularly good to catch up with news on the rest of my family. Today, a cloudless sky and glorious sunshine; even on this commuter train, the clothes and the mood of the people around me say "summer". Date with spikeylady tonight.

I got a question posted to "Ask Slashdot". Sadly all the answers are really dim; no-one seems to have even read the question properly, let alone given it any thought. A colleague at work describes it as "I read the first sentence of your question, assumed you were stupid, and answered it". It's a shame; considered properly, I think it's an interesting problem. Maybe it'll help me shake my Slashdot habit. (NB if you're interested, please answer it there, not here!)

Sadly it looks like the chance to work on an interesting Wiki-like project at work may pass me by, but it seems that the guy who is doing it is interested in my ideas. I have many ideas on how a Wiki should work, but there's only very limited implementation time. Still, the guy doing the work is very bright and very productive, so some cool stuff will probably result...

Comments

[pavlos.livejournal.com]

Serves you right for posting Slashdot questions. Before reading your clarification, I prepared this:

----

I don't understand your goal exactly. Is it:

A. To use the libraries for locating and parsing the ID3 tag. You trust them to consume data efficiently, and the problem is just how to encapsulate file access over HTTP.

B. You can use your own code to locate and fetch the ID3 tag, and you wish to use the libraries to parse it. For this you could embed the tag in a fake local MP3 file. Does that need to be the same size as the original, and if so do you need some kind of sparse file support.

C. You need to write new code to locate and parse the ID3 tag while consuming data efficiently (the libraries read too eagerly). you would like to make your code part of the libraries, but architectural considerations make this hard.

for A and B (with sparse files) you need some sort of loopback device, to implement HTTP encapsulation or sparse files. Both of these sound like they could have general use. Is the problem that they require root access?

In that case could a socket (on the filesystem) be an alternative? I don't know if they can be seekable and if your libraries require that. But if you trust the libraries to read once and only seek forward, you could probably implement the appropriate sockets.

----

Pavlos