Additionally, this monstrosity is built on top of Wordpress. The site is not even a blog. Why do people insist on doing that? (Yes, I know the answer is "when the only tool you have is a hammer, every problem looks like a nail", but still... have they no professional pride???)
Posts made by Monomelodies
RE: I don't think that means what you think it means...
I don't think that means what you think it means...
Representative line from a project I just took over:
var map = ""; //please be global
Oh, this is only the tip of the iceberg... I got called in on this because:
1. "searching is slow". Well, with a non-indexed database that stores numeric values (e.g. longitude/latitudes) as varchars that doesn't surprise me.
2. The site crashes Safari on iOS - after some digging, I found it was displaying around 1800 markers on a Google map. Twice. I can imagine Safari not being too happy about that.
Comment by the client upon reporting this: "I can't imagine [name-of-original-coder] didn't know all this..."
So. Anybody else doing anything fun this Saturday night?
RE: Assault WTF
I think I have to sort-of side with Ronald here. Molestation and assault are - practically as wel as legally - two different things.
However, it goes without saying that this does NOT make Joe's behaviour in any way acceptable. No is no, and really, if you're someone's boss you ought to know better in the first place (but really ANY man should know better).
And yes, getting gang-raped on a bus is objectively worse than getting fingered during a drunken party, but it can be traumatic all the same. No matter how much you think a woman is leading you on, if she says "no" she means "no". Period. There simply are no excuses for either molestion, assault or rape.
Identity crisis thanks to Google
A little background: for the past 6 weeks or so I (Marijn) have been contracting on-site for a rush job, which involved me hauling my 4 kilogram weighing laptop halfway across the country every day. Tiring, but well payed.
For this project I had a scrum master (before you start: yes, that's TRWTF(tm)) called Servan. Nice and competent guy. On general things were going smoothly. Last week he added some meetings to my Google calendar, mainly involving the finishing up of the project. Fine.
Today I'm back at my own office and want to put something else in my calendar. I'm using my office desktop, laptop is having a rest at home. I open Google Calendar, and after a minute or so I realise ALL the stuff I'd put in there is GONE. WTF. Then I look more closely and my jaw hits the desk.
Google, really? How is this even possible? It's a different machine with a different session, and this man hasn't even been NEAR it. Also, I'm sure this isn't HIS calendar either, he's bogged down with appointments too. Anybody have ANY idea what's going on here???
For bonus points, I tried Gmail and Google+ too. Sure enough, I'm still signed in as Servan (I can tell by the secondary email address it shows). I've no idea if he even uses those (Google+ probably not, Gmail he might but no idea). I'm actually a bit shocked - this means I'm not going to ever give people access to my calendar ever again...
RE: An open MySQL bug now has a birthday cake
I have to wonder what a Postgres user considers "high-traffic". A few hundred queries /sec?
Well, "high" is always subjective of course, but this particular one could handle in the tens of thousands /sec and still chug away happily (this was on a ~50gig database). Then again, this was around 2005/2006 when MySQL was still unusable for anything with that kind of load (the platform started out on MySQL, but we got bored of the weekly restores due to data corruption pretty quickly). I've been told MySQL can be faster if you're mostly doing lots of reads, but I haven't benchmarked it personally since PostgreSQL reads fast enough for our purposes and there was no need to switch back.
Well, as I said, this was 2008. And it wasn't just high-traffic (in fact, the traffic wasn't that high at all) but it was high-volume. Several TB of data and with hundreds of gigs of turnover per-day. Deleting and adding 200 gigs per-day apparently causes Postgres to shit itself.
Can't comment on that, that does sound hefty. My first thought there was "should you even be using an RMDBS for that at all".
Conversely, if your use-case is "I need to support 500k queries /sec across 100 DB shards containing half a petabyte of data" then Postgres isn't going to look so hot. And while I'm loathe to call MySQL good, it can at least be configured to handle that kind of load.
Like I said, it's very possible that kind of extremes cause issues. I would consider it an edge case though, not many products would have those kinds of requirements. So to call PostgreSQL crap based on it failing to handle this particular case seems overly harsh, that's all I'm saying.
Meanwhile, look at all the companies who run core services on MySQL: Google, Wikipedia, Twitter, PayPal, Yahoo!, Facebook.. Now I'm not saying MySQL doesn't have a lot of flaws, but do you really think any of those companies could swap out MySQL for Postgres? Like, at all? Even if they dumped a few million into consulting and the move Postgres just plain isn't going to support their needs.
I honestly wouldn't know, since I don't know "their needs". E.g., I have no idea what "core service" Google uses MySQL for (certainly not their search engine ). Having said that, MySQL has come a long way in recent years (shame Oracle seems to be screwing it over...) and it's now mostly perfectly usable, although there are things in PostgreSQL I really miss everytime I setup a project with MySQL. But it's nowhere near as bad as it used to be.
But let's agree to disagree on this one - you obviously found a tool that works for you, and so have I. Everyone's happy!
RE: An open MySQL bug now has a birthday cake
That's interesting. I've been running PostgreSQL on a number of high traffic sites, and the last time I remember autovacuum causing issues is somewhere in 2006. I'd be interested to know your setup, since ours hasn't caused problems for years, literally.
RE: Does this antipattern have a name?
The double check on is_string for the $sync_system is also awesome. But really, what in the name of baby Jesus is this code supposed to do besides completely unnessecarily wrapping "$class = 'lfSyncFOO'; $sync_system = new $class" in an object that's a greater waste of space than your average Eighties act?
(Out of morbid curiosity, what do 'lf' and 'sf' stand for, and why does an lfSomething throw an sfException?)
RE: We dont need no stinkin tables, we got sys.dual
Not a beautiful solution, but you could try, eg:
rownumber() over( order by foo asc ) as row
) [ ]
where row between 10 and 20
order by foo asc
(I have no idea if this will run.)
It's been quite a while, but I think the trick was to do a TOP(x) with ORDER BY [whatever] ASC and then use that as a subquery to the same but only ordered by DESC, and then reverse the resultset before outputting. Or something. Anyway, it was ugly.