3-27-13 – The Onion on Your Belt

Elsinore, I’m having problems with the “Keep me logged in on this computer” checkbox. I’m using Chrome and I cleared my browser & cookie cache recently and ever since I’m having to log in each time I bring up the forums, regardless of whether I check that box or not. Could you ask Zeke if he’s got any ideas when he’s got a couple of minutes? I’ve got Fiddler installed so I can get him header traces if he wants. I can still get in, so it’s not critical, but it’s kind of a pain to login every time.

Works fine:
Chrome X.X (?)
Windows 7

I’ll bet you’d find that it stops working if you clear the cookie cache. I’m running a similar setup and all was well for many moons until I did that recently.

Sorry about that. Again. Not Fark’s fault this time. I really need to get the scraper more automated so it’s not subject to the memory faults of its creator. Did I mention I’m turning 50 in a couple of weeks? 😉

If you want a hand with the coding (it is C# with some sort of SQL backend right?) I’ll be happy to lend a hand. Who do you host with? I had to get a little creative with my ‘automated’ jobs at VeloBase.com since I host with GoDaddy and don’t have any way of scheduling jobs outside of my web application on their servers. Message me and I can give you details on a few of my solutions if you want.

Somebody supposedly named Eeek! sponsored me for a month of T-Fark. I tried to say thank you, but when I went to look at the profile for an email address, there apparently was no such user (tastes like chicken).

If you’re reading this Eeek!, thank you!

If you know who Eeek! is, please say thank you to them for me.

Eeek! is one of the Mods, she covers for Gwinny when Gwinny is offline.

^he…he covers for Gwinny (as I do on occasion…both of us will be next week).

But weird on the login thing. Zeke’s been out of town, but is back, though catching up on some scout stuff tonight. I’ll mention it to him. Nothing under the hood has changed recently, though….when did this start? Any info is good, so header traces, whatever, thanks 🙂

I’m seeing the same thing on my desktop at home. Win 8/chrome & firefox. Haven’t tried IE. Always logs me out but login info is saved. My laptop running same config stays logged in.

If you want a hand with the coding (it is C# with some sort of SQL backend right?) I’ll be happy to lend a hand. Who do you host with? I had to get a little creative with my ‘automated’ jobs at VeloBase.com since I host with GoDaddy and don’t have any way of scheduling jobs outside of my web application on their servers. Message me and I can give you details on a few of my solutions if you want.

Thanks for the offer, Barracuda. It is C# running against a couple of SQL Server DBs, and the site is hosted on GoDaddy.

Let me give you an idea of the current process and future plan and then you can see if any of the issues you’ve already solved might apply here.

First, there are two components that run on my desktop. Right now, they’re both just WPF applications; the future plan is to make them Windows services so that they’re always running and auto-start on reboot. But since they’re just apps now, I start them manually sometime on Wed.

The first component is the Statistics Librarian, or scraper. Its UI looks like this. I manually enter the Active Contest information and then click either Create or Watch. “Create” is used when I know the link ID (i.e., if I’m starting the scraper after the link’s been posted here). “Watch” is used otherwise and polls Fark’s topic/farktography page looking for a new contest; when it finds one, it uses the link ID from that. Note that I don’t parse the contest number, theme or description; that’s manually entered.

Once a link ID is known and the contest opens (midnight EST), the scraper wakes up every 15 minutes and scrapes both the contest and voting result pages. It downloads any new entries found, creates & saves thumbnails locally, and saves all of this info to the local FarkStats database. Every 15 minutes is a new snapshot and a new set of records.

The second component is called DataPush and it’s responsible for summarizing the detailed info in FarkStats, calculating entrant statistics, and combining it all in such a way as to be efficiently displayed on the website. It reads from FarkStats, writes the condensed data to a local BiblioStats database which has a schema created just to support the website, and uploads the changed records and thumbnails to the website through a web service. The website also has a BiblioStats database with an identical schema, so this is my version of SQL replication. DataPush uses Entity Framework so it can detect only those records that have changed.

I created two components for this because I didn’t want problems pushing the data to the web to interfere with the process of scraping and capturing the data. DataPush waits on a system event which is triggered by the scraper after each successful scrape cycle just before sleeping.

The website itself uses read-only access to the web BiblioStats database and caches almost all displayed data using backing SQL dependencies to invalidate the cache when the web service receives new data.

The future plan is to 1) enhance the FarkStats schema so that I can enter multiple future contests in advance; and 2) move the non-visual logic in DataPush and the Librarian to Windows services which run non-stop in the background and are timed to wakeup and grab the next contest every Wed night. That will take care of any forgetfulness on my part, as long as there are sufficient future contests lined up. (I know I could have the scraper read the contest info from the page, but for now, I prefer formatting it myself. If I ever get time to enhance it to support PS contests, I’ll have to move to that method for sure.)

^he…he covers for Gwinny (as I do on occasion…both of us will be next week).

But weird on the login thing. Zeke’s been out of town, but is back, though catching up on some scout stuff tonight. I’ll mention it to him. Nothing under the hood has changed recently, though….when did this start? Any info is good, so header traces, whatever, thanks 🙂

Eeek! is a HE? For years I thought he was a she. And I am not sure why. Of course a lot of people think I am she as well. Apparently Ennuipoet is not the most masculine of monikers, that I and occasionally like to dress as Marilyn Monroe.

/NTTATWT

You mean you’re not a girl?