Pissed.

Not like you mean over in US though.

No, not angry.

You know what I mean.

BTW, save me having to write a new thread.

Where's Trent and the Supreme Loser?

Jeez, where has everyone gone? I guess it's nice weather now and no one wants to sit inside and computer.

I'm in a position where I don't have the time to sit in front of the computer as much as I want.

Went to a concert tonight at... I guess you would call it a pub. It was basically the Pigeon's Nest from Animal Crossing but with alcohol instead of pigeon milk.

... That's how to be off-topic in Off Topic.

Got bored with WoW a few days ago. Will probably not continue playing it. I only started playing to escape from RL problems anyway but it turns out problems do not always go away when they are ignored (you would be surprised how often it works, despite people saying it does not ).

So... blah.

Shouldn't that access grant a return of a boost::variant as per your definition?

Not with that signature, as far as I know...your asking for a variant return type....I'm unsure there is a technique granting what you want, but I am too a n00b to know for sure.

Anywhere I, would've followed you (even on Twitter)

Armando, I'm sorry that I couldn't get to you before you started feeling this way about your C++ skills. I'll swallow my pride, you're the one that I teach, please don't say goodbye

Josh

Well, that would get close (as I initially tought) to what Armando's asking but no cigar, he is askinf to implement a method that automatically detects the type and returns the corresponding item.

I sense come C++11ness in there, as well as something far beyond my reach

By the way I could use some teaching as well, are you holding classes?

You might be able to use a template to get the dynamic return value. My C++ is a little bit rusty.

auto x = row.get<int>("column"); // x is implicitly an int.

That's probably the closest you'll get. That said, I really think that SQL is the best tool for the job, and you should keep your C++ API scoped to just be capable of handing off SQL requests and getting at the results as efficiently as possible (so don't bother trying to create a query language in C++). Let SQL do what SQL was designed to do. Use C++ to make the request and consume the data in as few keystrokes as possible.

I'd appreciate it if you publish the result. fetchmail + mutt is becoming unmanageable given my current situation behind a Microsoft proxy. It worked fine when a server-side mutt became my primary MUA, but since I most often don't have access to my server during work hours I am stuck managing multiple copies of my mailbox and trying to manually keep them in sync.

I haven't had the time to try to set up my own UNIX-friendly proxy, but I suspect that I'd get in trouble if I got caught anyway. At least if I could access G-mail via IMAP then I'd be able to get the best of both worlds. In theory. I think that the way that G-mail does IMAP is actually not very good... I digress. The alternative would be setting up my own E-mail service, which would be a lot of work and would result in less reliable service.

Hear, hear!

I think that's the very problem OfflineIMAP was designed to fix. Admittedly, I've mostly been configuring the Mutt side of things, but an ideal set-up would have a one-for-one offline backup on the main machine, with smaller, filtered versions on portable devices. OfflineIMAP synchronizes everything to Gmail and even includes a dedicated Gmail protocol, as apparently Google's IMAP isn't 100% pure .

Mutt can handle IMAP out of the box, though, and Mutt configuration alone has kept me tied up for a bit. Here are a few observations for anyone interested in emailing like a boss:

• The mail_check variable can slow down Mutt when set too low. To be honest, I'd rather disable automatic checking altogether--though this will become moot when using OfflineIMAP.

• HTML email is not supported out of the box. Set up a reader (w3m, lynx, etc.) and add an entry to mailcap. This does not require a helper script, as one source suggests. The bare command suffices.

• Straight from the manual (concerning MIME types in .mailcap): "Entries with the copiousoutput tag should always be specified as the last one per type."

I use template structs/classes to get "Variant" return values:

template<>
struct stack<int> {
  static int get(LuaGlueBase *, lua_State *s, int idx)
  {
    return luaL_checkint(s, idx);
  }
  
  static void put(LuaGlueBase *, lua_State *s, int v)
  {
    lua_pushinteger(s, v);
  }
};

And to use such a thing:

   int foo = stack<int>::get(glue, glue->state(), 0);