Overdue Library Book
-
Saw this today:
-
It looks like yet another programmer forgot to check his return codes before displaying the result.
-
Heh. Looks like a -1 on a datetime format using the 1900 epoch.
Though I laughed even more with a lot of sites (and some BBS's) that used 2-digit dates. Their own "Y2K" bug consisted in 2000 being marked as 100, 2001 as 101 ... so you'd find stuff like Last logged in on 13/02/106.
-
I saw websites LAST YEAR that still had this buggy code displaying the incorrect year on the front page! And one of them wasn't long-dead, as you'd expect. I'm imagining a small company, no one with Javascript or even basic HTML experience, and too cheap to hire a consultant.
-
Heh, obviously a beginner's mistake - programmer must've been a newbie...
-
@danixdefcon5 said:
Though I laughed even more with a lot of sites (and some BBS's) that used 2-digit dates. Their own "Y2K" bug consisted in 2000 being marked as 100, 2001 as 101 ... so you'd find stuff like Last logged in on 13/02/106.
I see a lot of that with sites written in Perl. Does anyone know why on Earth Perl returns the year minus 1900?
-
Despite this one obviously being a programming error, books being overdue by more than 100 years are not unheard of.
-
@operagost said:
I saw websites LAST YEAR that still had this buggy code displaying the incorrect year on the front page! And one of them wasn't long-dead, as you'd expect. I'm imagining a small company, no one with Javascript or even basic HTML experience, and too cheap to hire a consultant.
Yes, because consultants can fix anything and they never introduce WTFs of their own...
-
@Cap'n Steve said:
@danixdefcon5 said:
Though I laughed even more with a lot of sites (and some BBS's) that used 2-digit dates. Their own "Y2K" bug consisted in 2000 being marked as 100, 2001 as 101 ... so you'd find stuff like Last logged in on 13/02/106.
I see a lot of that with sites written in Perl. Does anyone know why on Earth Perl returns the year minus 1900?
I always assumed that it was because C does it this way. According to the C specs, tm_year is defined as the number of years since 1900. Now, what I always hate is forgetting that that month (tm_mon) is represented as 0 through 11, not 1 through 12. Anybody know why this is the case?
-
@WeatherGod said:
Now, what I always hate is forgetting that that month (tm_mon) is represented as 0 through 11, not 1 through 12. Anybody know why this is the case?
Because C is essentially thinly-veiled assembly and programmers can't get away from the goddamn zero-based sequence idiom.
-
@WeatherGod said:
I always assumed that it was because C does it this way. According to the C specs, tm_year is defined as the number of years since 1900. Now, what I always hate is forgetting that that month (tm_mon) is represented as 0 through 11, not 1 through 12. Anybody know why this is the case?
Clearly it's defined as the number of months since January.
-
> Heh. Looks like a -1 on a datetime format using the 1900 epoch.
Well that settles it... <a href="http://xkcd.com/376/">Epoch Fail</a>
-
Nice. Guess I need an HTML refresher...oh well...
-
@WeatherGod said:
Zero-based system. The funny thing is that it only applies to the month, and not to the day of month. I find it kind of funny, as usually the OS's epoch is 1970, though I think that a negative time value would cover dates under 1970.I always assumed that it was because C does it this way. According to the C specs, tm_year is defined as the number of years since 1900. Now, what I always hate is forgetting that that month (tm_mon) is represented as 0 through 11, not 1 through 12. Anybody know why this is the case?
-
@bstorer said:
@WeatherGod said:
I always assumed that it was because C does it this way. According to the C specs, tm_year is defined as the number of years since 1900. Now, what I always hate is forgetting that that month (tm_mon) is represented as 0 through 11, not 1 through 12. Anybody know why this is the case?
Clearly it's defined as the number of months since January.And they couldn't stick with that logic when defining tm_mday (which has range 1 through 31)? I mean, they did it for tm_yday [0, 365], and tm_wday [0, 6]. Everything else in the list is zero-based. For the time parts, that makes sense. For the year, month and day parts, it sort of makes sense, but only because it is explicitly defined that way.
-
@operagost said:
You can find plenty of 'em still going with a google search for "June 19108"I saw websites LAST YEAR that still had this buggy code displaying the incorrect year on the front page!
-
@WeatherGod said:
According to the C specs, tm_year is defined as the number of years since 1900. Now, what I always hate is forgetting that that month (tm_mon) is represented as 0 through 11, not 1 through 12. Anybody know why this is the case?
It's so you can have a 12 element char* array of the month names (array indexes start at 0 of course), without wasting a sizeof(char*) chunk of memory for an unused array element 0. Same for tm_wday and a char* array of weekday names.
-
@danixdefcon5 said:
SyncEdit still does say something like "26 Jun108" instead of the correct date, it is a editor used in Synchronet and Synchronet is still being used today (I sometimes log in to X-BIT to play Word Warp or sometimes for other reasons as well). They know about the Y2K bug.Heh. Looks like a -1 on a datetime format using the 1900 epoch.
Though I laughed even more with a lot of sites (and some BBS's) that used 2-digit dates. Their own "Y2K" bug consisted in 2000 being marked as 100, 2001 as 101 ... so you'd find stuff like Last logged in on 13/02/106.
[code]Y2K Patch ~~~~~~~~~ There is a Y2K bug in all known versions of SyncEdit where the year 2000 is displayed as 100, causing the status bar to wrap and making the program unusable. I've released a group of binary-edited SyncEdit executables (for all located versions) that works around this problem: the year is still displayed as 100, but the status bar no-longer wraps. The file that includes these patched files is SYEDTY2K.ZIP and can be downloaded via Telnet/Zmodem or FTP from vert.synchro.net.[/code]
