back to article Epoch-alypse now: BBC iPlayer flaunts 2038 cutoff date, gives infrastructure game away

Feeling old yet? Let the Reg ruin your day for you. We are now substantially closer to the 2038 problem (5,849 days) than it has been since the Year 2000 problem (yep, 8,049 days since Y2K). Why do we mention it? Well, thanks to keen-eyed Reg reader Calum Morrison, we've spotted a bit of the former, and a hint of what lies …

  1. Disgusted Of Tunbridge Wells Silver badge
    Facepalm

    Strange that they wouldn't use NULL to mean never expires, rather than setting a timestamp arbitrarily far into the future.

    1. Korev Silver badge

      I'm sure "normal people" would get a bit confused by that... It's nothing that an if statement couldn't fix though...

    2. Hans Neeson-Bumpsadese

      I think they've got something like that - I'm sure I've seen programmes with "available for more than a year" which I took as a message that a developer had put in to cover a case where there is no explicit expiry date

      1. Disgusted Of Tunbridge Wells Silver badge

        My guess further down the thread is that they were setting an unsigned int to -1. On 32-bit that would set it to 2038.

        And because (unint32's -1 != unint64's -1), it wouldn't show the "never expires" text.

        1. Dan 55 Silver badge

          The available until date is set to 18th of January 2038. So someone, for whatever reason, deliberately set it to one day earlier than 0x7FFFFFFF, either at 00:00:00, 03:14:07, or 23:59:59 (pick your poison).

          Maybe the new guy wasn't aware of the value to get it to say "for more than a year", which might be -1.

          1. Fred Daggy Bronze badge

            Not sure I am a fan of packing two meanings in to a single data field. Perhaps a bool or flag for "Never expires". Then another field (which might be the aforementioned 2^32 datetime field, which indicates the expiry. Date of expiry only checked if the bool indicates it is required.

            Or is my programming style way out of date?

            1. The Basis of everything is...
              Joke

              Far too sensible. Go and sit in the corner until you're ready to rejoin society.

            2. Dazed and Confused

              Next you will be wanting the Itanium processors NaM flag to indicate the difference between a zero and a "Not a Number" like the difference between a NULL string and a string of length zero.

            3. LybsterRoy Bronze badge

              You should see some of the arguments I've had with people (mainly Delphi programmers) on using a NULL as a flag when an actual real flag would work so much better.

              I actually had a conversation with someone who claimed to be the person who wrote the system that's always used to justify NULL <> empty space.

            4. Dan 55 Silver badge
              Trollface

              Or is my programming style way out of date?

              Just think of the options when there is no expiry time, in increasing degrees of inventiveness (or f'd-up-ness, call it what you will):

              1. If you set bit 31, you've got 31 flags in bits 0-30 for your own use

              2. If you set bit 31, you could read bits 0-30 as bitfields

              3. You can store values from -1 to -2147483648, with different negative values or ranges of negative values having different meanings

              Or for maximum f'd-up-ness:

              4. You could solve the Y2K38 problem by interpreting -1 to -2147483648 as a positive number with a new epoch starting at 2038-01-19 03:14:08 and use custom time calculation and display functions. Don't forget the base value is -1 which should be interpreted as 0!

              All this kind of innovative inventive programming was done on 80s home computers to save a byte in RAM somewhere, your programming style is more modern and profligate with memory.

            5. but what do I know?
    3. Zippy´s Sausage Factory
      Coat

      I wonder if NULL actually does mean never expires, and right now one BBC BOFH is showing this article to another BBC BOFH and demanding they cough up the £20? I mean - a bet's a bet, after all.

      1. Jellied Eel Silver badge

        Or one lawyer is waving it at another lawyer, whilst the judge looks bored.

        I'm guessing under the hood, expiry dates get complicated. So there may be a streaming window where a file is PPV, then free to air, then unavailable on iPlayer but available via Bbc Worldwide's front-end. Then whether there are time-limited licenses to use music, clips, or copyright expiring in general.

        Which can all get rather complicated & why some old content isn't legally available online because streaming rights weren't a thing when the works were created. Apparently the Bbc is struggling to refit the TARDIS (Temporal Administration & Rights Discovery and Implementation System) so it can register historical copyrights in favor of the Bbc. Hence why it's demanding an RPI+ licence fee increase, because time travel isn't cheap.

    4. Anonymous Cowerd

      NULL should never have a meaning other than "I don't know".

      If you think otherwise, I don't want you coding for me.

      1. LybsterRoy Bronze badge

        Re: NULL should never have a meaning other than "I don't know".

        Downvoted because I'm a human being and I have yet to be able to differentiate between a printed empty string and a printed NULL unless someone has written extra code to print NULL.

      2. Anonymous Coward
        Anonymous Coward

        Re: NULL should never have a meaning other than "I don't know".

        It means no value, not I don’t know.

        The reason maybe you don’t know, or it could be it isn’t relevant in the specific circumstance.

    5. LybsterRoy Bronze badge

      Or be kind to humans and have a boolean flag expires/does not expire

      I hate having to guess that seeing nothing means something!

  2. TheProf Silver badge
    Facepalm

    A fix for this

    The Tories are on this and come 2038 nobody will have to worry about accessing anything biased on the BBC.

    1. hoola Silver badge

      Re: A fix for this

      Or more possibly, anything what so ever on the BBC......

    2. codejunky Silver badge

      Re: A fix for this

      @TheProf

      That does depend on the tores getting elected next time but the BBC seems more than a little worried that without the TV tax they wont be around. Not surprised they are worried based on their content.

      1. TheProf Silver badge

        Re: A fix for this

        The BBC's TV content may not suit you or myself but I belive there are one or two other people who rely on the BBC for local news and dancing and sewing programmes. As much as 'we' dislike them talent shows and soap operas are the current staple of British TV.

        For a corporation that has seen it's income gradually reduced by the culture wrecking party it's still managing to turn out a few watchable shows. It's not perfect by any stretch of the imagination but it is bloody popular.

        Imagine that future where every terrestrial TV channel is showing programmes interrupted every 10 minutes for adverts for food delivery services and on-line bingo companies.

        1. codejunky Silver badge

          Re: A fix for this

          @TheProf

          "The BBC's TV content may not suit you or myself but I belive there are one or two other people who rely on the BBC for local news and dancing and sewing programmes"

          Sounds like a lot of room for slimming the number of channels then.

          "It's not perfect by any stretch of the imagination but it is bloody popular."

          Fantastic so it should have no fear of relying on its popularity to fund it instead of everyone else who has no interest. Shouldnt it?

          "Imagine that future where every terrestrial TV channel is showing programmes interrupted every 10 minutes for adverts for food delivery services and on-line bingo companies."

          Yeah it sucks. The only reason to watch those channels is because there is something worth watching. Which seems to be how the majority get along. Or paying a subscription to watch streaming services without adverts, which also seems to be popular. So if the BBC is so 'bloody popular' it could raise revenue direct from those who want it. Do you think that would be enough? (I dont and they dont seem to believe that either).

          1. Rattus
            IT Angle

            Re: A fix for this - You are not required to buy a TV licance

            Given that I am assuming you only stream content, simply don't by a TV and use a monitor instead. no need for a licence...

            Now I agree a lot of the content on TV is missable - I for one do not watch. However I do pay for a licence because I listen to the radio, and I want a news service (OK I would like them to verify stories rather better than they currently do - but AFICT the beeb is still leading the way)

            1. DialTone

              Re: A fix for this - You are not required to buy a TV licance

              Or DO buy a TV and don't use it to watch live TV - also no need for a licence. It's a fallacy that you need a licence for owning a TV - the licence is required only if you watch or record TV as it's being broadcast, or use the iPlayer service.

              I've been licence-free for some time now but I still own and use several TV's as I do watch Netflix occasionally, use my gaming console connected to the TV, watch YouTube etc, and with prices starting at ~£300 for a decent-sized TV versus many multiples of that for a comparable monitor it's a no-brainer to get a TV instead of a monitor!

              I do hear what you're saying about radio - I personally don't listen to BBC stations, however you're not required to have a licence to listen to radio (including BBC stations), nor to access BBC websites (including the news).

              I agree that this may seem unfair on its face, and there's many legitimate arguments in support of funding the BBC, but I guess that's something that will need to be resolved moving forward, especially as they move towards the end of their Royal Charter. I've no doubt we'll all end up paying for it in some way eventually.

            2. codejunky Silver badge

              Re: A fix for this - You are not required to buy a TV licance

              @Rattus

              "Given that I am assuming you only stream content, simply don't by a TV and use a monitor instead. no need for a licence..."

              Which is what I do. Previously I have had sky and virgin TV which was worth paying for but I grudgingly had to prop up the BBC even though it was no use to me. So in the end I figured I wouldnt waste my money paying an extra company for literally nothing!

              "Now I agree a lot of the content on TV is missable"

              And that applies to all services. We are all different with our own tastes. As you say you like the radio, I dont. Neither is right or wrong just different tastes. That is why the BBC fears having to be funded by people who actually like their content, because they will lose a lot of money very quickly.

          2. Peter Gathercole Silver badge

            Re: A fix for this @codejunky

            It's funny how so much of the other channel providers on broadcast TV carry material originally produced for or by the BBC...

            But the statement yesterday from Nadine Dorries was just a smack in the face for the BBC. I mean, saying that holding the license fee down to was to benefit the poor and the especially the elderly, when this government or one of it's immediate predecessors was responsible for dropping the over 70's free TV license, forcing the BBC to pick it up, and then abolishing it completely.

            I wonder how many of the people she's talking about actually do already pay a lot more than the license fee for other subscription TV services (OK, I know there are many people who do really struggle day-to-day, and they deserve a lot more than just keeping the license fee down, but that's not all of them).

            And then suggesting that there may be a Government fund to pay for certain important content and radio! This just flies in the face of the editorial independence of the BBC, and risks just making it a government mouthpiece.

            The existing license, though enforced by the law, is actually not government money, and has given the BBC the ability to claim to not be funded by the government. It's not perfect by any means especially in this day and age, but nobody has yet come up with a real viable alternative.

            Many UK governments have disliked the BBC for being independent and either too left wing, or too right wing, depending on their political bent, but this one appears to be the one actually trying to break it!

            Another step in the demise of UK democracy.

            1. codejunky Silver badge

              Re: A fix for this @codejunky

              @Peter Gathercole

              "It's funny how so much of the other channel providers on broadcast TV carry material originally produced for or by the BBC..."

              While I hate to say it 'the good old days' when they made stuff to be watched. Its interesting to watch (mostly comedians) complaining that their highly popular content (even now) could not be made in this day and age. Even having parts edited out or having warnings of triggering content.

              "I wonder how many of the people she's talking about actually do already pay a lot more than the license fee for other subscription TV services"

              But is that wrong? People paying for something they want but forced by law to pay the competitor too? It would still be the same problem for low income people with an android phone having to pay apple or buying a car but having to pay a british leyland tax.

              "And then suggesting that there may be a Government fund to pay for certain important content and radio! This just flies in the face of the editorial independence of the BBC, and risks just making it a government mouthpiece."

              To be honest its probably not so much an issue if its branded the governments mouthpiece. Then people know not to take it seriously. But unfortunately the BBC decided on its 'agenda' and people took it seriously as its honesty fell off.

              My issue is that the BBC is political. They have been caught stacking the audience in QT, misrepresenting the MMCC co2 debate and altering news articles without pointing out they made such drastic edits that the original story was completely incorrect. Instead of reporting it turned to trying to influence

              1. Peter Gathercole Silver badge

                Re: A fix for this @codejunky

                I'm not talking about ancient comedy shows. I'm talking about the documentaries, especially the natural history ones, the costume dramas, QI, old Top Gear (although this is long in the tooth now) and so many other shows (I hesitate to mention Stricly, but it is popular). I would have mentioned Dr Who, but....

                The Reith lectures. The BBC Proms, the more I think, the more there is.

                Many of these shows would never be commissioned by other channels, so losing them would be a loss to many people who claim to never to watch the BBC.

                And like it or not, BBC news provides a better (IMHO) view of what is going on than Sky or the ITV news. And not just at a national level. I don't know whether you've ever listened to what is laughingly called news on the commercial radio stations.

                Nothing's perfect. Maybe more oversight is needed, and not just from the BBC Trust who are just like the inmates running the prison.

                1. codejunky Silver badge

                  Re: A fix for this @codejunky

                  @Peter Gathercole

                  "I would have mentioned Dr Who, but...."

                  Yeah, I wasnt a fan of Dr Who (just not my taste) but watching it be gutted by 'the message' was stupid.

                  "Many of these shows would never be commissioned by other channels, so losing them would be a loss to many people who claim to never to watch the BBC."

                  I dont know how that works. Someone who doesnt watch the BBC wont have a loss if they didnt watch them. And talent like the original top gear has gone to a company who wants to attract viewers over following 'the message'.

                  "And like it or not, BBC news provides a better (IMHO) view of what is going on than Sky or the ITV news."

                  Does it? I honestly dont know because I read the news not watch it. When I visit family and they have the news on (usually BBC sometimes ITV) I find myself calling bull or stupidity so much its nuts. I usually end up having to fill out the large portions of missing information for them.

                  "I don't know whether you've ever listened to what is laughingly called news on the commercial radio stations."

                  I really dont. I prefer to read the news so any glimpses of news on TV or radio is only when visiting family.

          3. SpamuelBeckett

            Re: A fix for this

            "Or paying a subscription to watch streaming services without adverts"

            Bless you for thinking that.

            Amazon Prime shoves adverts for other Prime stuff you have to subscribe or buy to in order to watch.

            Disney+ also advertise movies that need you to pay a fee too.

            Now TV also shows normal adverts for any sort of items.

            Netflix is the best really for no adverts, but thats not going to last forever, especially if people are already stating "without adverts" when quantifiably that is nonsense.

            1. codejunky Silver badge

              Re: A fix for this

              @SpamuelBeckett

              "Amazon Prime shoves adverts for other Prime stuff you have to subscribe or buy to in order to watch."

              Sounds like the BBC where they advertise their own stuff in place of commercials. And of course you already pay the BBC because you want to watch other channels.

              "Netflix is the best really for no adverts"

              You can guess which one I am using. I know other people who use the other streaming services but I dont.

              I guess if these other services manage to attract viewers with commercials the BBC should find it easy to attract viewers with their 'quality' content? That is the test the BBC knows it will fail.

              1. Tom 7 Silver badge

                Re: A fix for this

                "Netflix is the best really for no adverts" I must be using it wrong. I get screenfulls of adverts for shit I'm never going to watch before I get anywhere near what I'm searching for,

                1. codejunky Silver badge

                  Re: A fix for this

                  @Tom 7

                  "I get screenfulls of adverts for shit I'm never going to watch before I get anywhere near what I'm searching for,"

                  Really? How? I use a tablet and tap my profile then go find what I want to watch. I dont see a single advert only a library of shows. I dont recall any issues watching on my PC either but aint done that in a while

          4. DuncanLarge Silver badge

            Re: A fix for this

            > Or paying a subscription to watch streaming services without adverts

            That's not always true. Paying a subscription does not stop you from being forced to watch the adverts.

            Example: Babylon 5 on Amazon Prime. Adverts automatically inserted at the most annoying moments. Broadcast TV at least makes it clear where they go and when they are coming but on Amazon Prime you are there watching the action unfold as the Vorlons finally attack the Shadows and boom 3 adverts. The SAME 3 adverts you see, over and over and over.

            Unskippble, even if you have ALREADY watched them because you got distracted and had to wind backwards to re-watch that action.

        2. Cederic Silver badge

          Re: A fix for this

          I do like the Great British Sewing Bee.

          If only the rest of their output was as balanced and politically neutral. The culture wreckers _are_ the BBC.

          1. DuncanLarge Silver badge

            Re: A fix for this

            I have always wondered about the connection of Bee's to sewing and spelling.

        3. LybsterRoy Bronze badge

          Re: A fix for this

          --For a corporation that has seen it's income gradually reduced by the culture wrecking party --

          Or as a corporation that has increasingly spent its money on producing rubbish and grossly overpaying some people to write/speak commentary that is meaningless.....

          I started watching "Around the World in 80 Days". In its favour it did keep the books title.

      2. gerdesj Silver badge
        Paris Hilton

        Re: A fix for this

        "tores"

        Is that a bull or a footie manager?

    3. Anonymous Coward
      Anonymous Coward

      The Tories fix for this

      is easy.

      The BBC will simply cease to exist the day after their charter runs out. Boris and pals hate the Beeb. Starmer and pals also hate the Beeb.

      For both to hate the Beeb sorta tells me that the beeb is not doing a bad job.

      How about instead of replacing the beeb, replace both the Tories and Labour parties.

      1. Anonymous Coward
        Anonymous Coward

        Re: The Tories fix for this

        "For both to hate the Beeb sorta tells me that the beeb is not doing a bad job."

        This is known as the balance fallacy, or the fallacy of false equivalence. It assumes that all criticism of the BBC is equally valid or invalid, whether it comes from Labour or the Tories.

        1. SundogUK Silver badge

          Re: The Tories fix for this

          The Tories hate the BBC because they are too left wing; Labour hates the BBC because they are not left wing enough.

          1. Tom 7 Silver badge

            Re: The Tories fix for this

            Why are people so stupid as to fall for that?>

      2. LybsterRoy Bronze badge

        Re: The Tories fix for this

        I upvoted the concept of replacing both the Tories and Labour parties.

        However, the beeb is not doing a brilliant job either in its programming or its news coverage.

    4. JimboSmith Silver badge

      Re: A fix for this

      The problem is that the commercial channels don’t want the BBC taking adverts or a subscription.I worked for a commercial broadcaster who didn't like the license fee. However they also didn't want the BBC taking advertising or sponsorship. Neither did the idea of encryption/subscription go down any better. People apparently might find the idea of two monthly charges unwelcoming and ditch the commercial one.

  3. Brewster's Angle Grinder Silver badge

    I'm glad you eventually get round to pointing out it's not a uniquely unix problem. Most (all?) implementations of C uses the "POSIX epoch" of Jan 1st 1970 so any C compiler with a 32 bit signed int for time_t is in the shit. Although I've just checked the spec and neither the width nor the epoch are mandated in modern C standard. (I've not looked backwards to see if it ever was.)

    The real problem is not application but filesystems and file formats which have used a 32 bit signed time_t in the format. That's more of a unix problem.

    Anyway, it looks like this will pay for my retirement!

    1. MacroRodent Silver badge

      so long

      Officially you should use the typedef name "time_t", but historically the calls and documents used "long" in old Unix, and many programmers have lazily continued using it for time without the typedef name. On 32-bit Linux and most other 32-bit systems, "long" is 32-bits. On 64-bit Linux, "long" grew to 64-bits, but 64-bit Windows still keeps it at 32-bits for backward-compatibility. So a Windows program that at some point uses bare "long" for time will have trouble, but one on Linux will not!

      1. Anonymous Coward
        Anonymous Coward

        Re: so long

        Problem is everyone thought that 2038 was a long time in the future and anything after that was a long long time away

        1. John Brown (no body) Silver badge

          Re: so long

          I remember as a kid buying the first issue of 2000AD and thinking the future was still a long way away!

          (which I'd kept it now, apparently it would have had some value!)

      2. Brewster's Angle Grinder Silver badge

        Re: so long

        I've done some digging: the original C Programming Language First Edition (1978) - i.e K&R C - doesn't define time_t and associated functions. They seem to arrive in ANSI C (C89) fully spec'd.

        But, yeah, back in the day I used to treat time_t as long. And all DOS C compilers I used adhered to Unix standards. However there are probably some embedded systems where that wasn't true.

        1. MacroRodent Silver badge

          Re: so long

          Yes, the library described in the original K&R is quite limited. Most implementations of C on non-Unix systems added some simulations of Unix calls for time, raw file file access, etc. The ANSI C spec specified a more complete library, with some calls inherited from Unix, and some that were commonly implemented in various C libraries. The "Rationale" section of the original ANSI C spec, where they describe their choices in making the standard, is fascinating reading for anyone interested in the C language.

    2. Tom 7 Silver badge

      All data types everntually expand to fill the disk space available.

      Sorry cloud space.

  4. Alan J. Wylie

    VMS got it right

    VMS (since 1977) has stored time as 100ns clock ticks since 17 November 1858 (the start of the Reduced Julian Day (an astronomical timescale, the "reduced" variant was introduced by the Smithsonian Astrophysical Observatory in 1957 to record the orbit of Sputnik). It will run out of bits in the year 31,086.

    1. Evil Auditor

      Re: VMS got it right

      ...and about 29,000 years from now, someone will have discovered an ancient calculation machine and will have figured out how it worked. And they will see that its clock stops in the year 31,086. And some of them will start a cult that believes the end is near for an ancient civilisation allowed their calender to run until then...

      1. tip pc Silver badge

        Re: VMS got it right

        What’s the clock used on VGER?

    2. Phil O'Sophical Silver badge

      Re: VMS got it right

      VMS got most things right.

      There is, though, a well-known bug filed against VMS for a related issue. The standard message display only permits 4-digit years, so even if the clock is fine until 31,086 there will be a display error when it ticks over on Dec 31st 9999. Last time I saw that bug in the DEC tracking system it was in an 'accepted' state, with a note that it will be fixed "in a future major architecture". Sadly unlikely now...

  5. Anonymous Coward
    Anonymous Coward

    >with a snapshot that implies Old Auntie might be using a 32-bit Linux in iPlayer

    This would be among the least likely explanations, given this is a value encoded in a database somewhere. Most likely on a technical basis is someone is using MAXINT for "never expires", which is always going to land at that epoch however big your epoch may be allowed to grow.

    Or perhaps someone in the iPlayer team has a mischievous streak and set the displayed maximum there deliberately as a nod to their fellow technologists.

    Given the iPlayer volume slider goes to 11, I know which of the two I'd prefer to be true.

    1. Disgusted Of Tunbridge Wells Silver badge
      Holmes

      Ah,

      My guess based on your post, they previously used MAX_INT. They updated to 64-bit and 2038 no longer matches MAX_INT.

      1. Anonymous Coward
        Anonymous Coward

        Close. The point is that MAX_INT is almost always going to be 2,147,483,647, regardless of whether you're using C/C++ on a 32 or 64bit linux, writing some enterprise glue code in Java or storing that value into an Integer field in MySQL.

        2,147,483,647 always lands you on that magic moment in 2038, regardless of whether your epoch is represented as a 32 or 64 bit number. The real question is what your system does with epoch 2,147,483,648, not what it does with the predecessor.

        1. Disgusted Of Tunbridge Wells Silver badge

          Well, if not MAX_INT, then -1 on an unsigned int.

          If that int became 64-bit it would no longer match the date in 2038.

          1. This post has been deleted by its author

        2. msobkow Silver badge

          MAX_INT is going to be the maximum value for the integer size of the machine, whether than be 32, 64, 128, or some bizarre number of bits, not the maximum value of a 32-bit integer ala older boxes or Java.

          1. Arthur the cat Silver badge

            Yes. I learnt to program on a machine with MAX_INT = 8388607(*)

            (*) Well, it would have been if there had been a C compiler for it. Dennis Ritchie had only just started inventing C back then.

            1. dajames Silver badge

              Yes. I learnt to program on a machine with MAX_INT = 8388607(*)

              So did I ... happy days!

              (*) Well, it would have been if there had been a C compiler for it.

              I sometimes wonder how C would have looked on an ICL 1900. The fact that a character was only 6 bits (and that the character set wasn't ASCII) would have been interesting ... as would taking the address of a character when it was stored in a quarter of a 24-bit machine word.

          2. Anonymous Coward
            Anonymous Coward

            This is half-true in some historical contexts, but I picked my words more carefully than that. INT_MAX is defined to be the same value on both 32 and 64 bit linux, to ensure interoperability. You'll find this is the case in almost all other contexts too.

            This is allowed because the C standard itself only sets a minimum bound for each type size. It doesn't tell you to align them to your maximum word length.

    2. Mike 16 Silver badge

      Goes to 11?

      Being geographically undesirable to iPlayer, I was not aware of this. I do know that many Atari arcade games starting with Marble Madness have a volume setting (operator only) that goes to 11.

      One wonders which was the chicken and which the egg. Since Marble Madness and Spinal Tap were both released in 1984, perhaps the good folks of MiniTrue, er, BBC, were busy at the time.

  6. Atomic Duetto

    Baa-ram-ewe!

    Not compliant these days of course, needs a number

  7. Anonymous Coward
    Anonymous Coward

    Erm... "Palin's Round The World In 80 Days" on BBC4, or "Round The World In 80 Days" with David Tennant on BBC1?

    (not sure why they are broadcasting both series at the same time!)

    Cisco have got round the y2038 problem by redefining 'day zero' for the RTC... some default to 1/1/2010, others to 1/1/2013, when the NVRAM battery goes flat (depending on IOS version)

    1. Androgynous Cupboard Silver badge

      Or "Around the world in 80 gardens", which is also on. Almost sat down to watch that with the kids by mistake the other day, which I imagine would have gone down like a bucket of cold sick.

    2. Anonymous Coward
      Anonymous Coward

      (not sure why they are broadcasting both series at the same time!)

      Probably someone realised that the plot & characters of the Tennant-inhabited series was entirely unlike the actual book -- and so that someone then thought the Palin one might serve as a more accurate representation :-)

      Indeed, I re-read the Verne original very recently -- in fact, just after thinking "I don't remember any of that" during Tennant Ep.1).

      Arguably (IMO), a fair fraction of the book is not necessarily well matched to the demands of modern TV, so a significant degree of "updating" was probably to be expected. It is a shame, though, that so little of the original was allowed to survive, and even that is mutilated. Mind you, perhaps Eps 7&8, which I haven't seen yet, are less off-piste?

      1. DuncanLarge Silver badge

        Re: (not sure why they are broadcasting both series at the same time!)

        > Indeed, I re-read the Verne original very recently -- in fact, just after thinking "I don't remember any of that" during Tennant Ep.1).

        Pretty typical for adaptions of the book.

        As a kid I remember learning of the existence of the story by watching an annoying cartoon of it with a Lion playing Fogg

    3. the spectacularly refined chap Silver badge

      Cisco have got round the y2038 problem by redefining 'day zero' for the RTC... some default to 1/1/2010, others to 1/1/2013, when the NVRAM battery goes flat (depending on IOS version)

      The RTC has absolutely nothing to do with internal date format used by the software on top. The RTC format is determined by the actual chip used, generally it will be some representation of time in a wall clock-style format, i.e. separate fields for year, month ... minutes, seconds. Sometimes those are separate ints for each field, on some chips they are in BCD form.

      Regardless it gets converted to some other representation after being read. That can be on use but most systems use it to set a timer at boot and ignore it after that, since most RTC chips don't offer better than one second resolution.

      The default "bad info" date is purely a style choice when an invalid CMOS checksum is detected. It does get updated from time to time. For many years virtually all PC compatibles would default to 1981 but that started changing around the millennium, a date that long ago simply screams how old the code base is.

      1. that one in the corner

        > The RTC format is determined by the actual chip used, generally it will be some representation of time in a wall clock-style format, i.e. separate fields for year, month ... minutes, seconds.

        True.

        > Regardless it gets converted to some other representation after being read.

        Oh, if only that were universally true. Having worked on an embedded system that is still in use, timestamping every 40ms using the RTC format plus an extra centisecs field. Then they continue using that format to do temporal arithmetic and comparisons against user-defined durations. Some very weird times were reported from that lot (not least data collected in 1720, 1820, 1920 ...).

        Only dealing with shifting the epoch to push back the 2038 End Of Time would have been a pipedream.

    4. dajames Silver badge

      ... or "Round The World In 80 Days" with David Tennant on BBC1

      The Tennant thing is actually "Around the World ..." (which is why I initially failed to find it under 'R' in the list of recordings on my PVR).

  8. Gomez Adams

    What about 2028?

    Have Unisys now totally expunged this issue from their OS1100/2200 systems?

  9. ShadowSystems
    Joke

    Do I feel old?

    Hell yes I feel old. I helped erect Stonehenge Version 0.75Alpha as a test run before we erected the full V1.0. Due to budget overruns & manglement cockups, I'm still waiting for the damned thing to finish POST.

    Wake me when it beeps...

    1. Phil O'Sophical Silver badge

      Re: Do I feel old?

      Didn't it get reset during the last total eclipse?

      1. J.G.Harston Silver badge

        Re: Do I feel old?

        Takes ages to move it back/forward an hour twice a year.

    2. David 132 Silver badge
      Coat

      Re: Do I feel old?

      I'm still waiting for the damned thing to finish POST.

      Well, it's only a 75 Megalith architecture, so pretty slow by modern standards. At least the horizontal cross-stones are using Linux on x86, or as it was known at the time, "lintel".

    3. Tom 7 Silver badge

      Re: Do I feel old?

      There's your mistake - for Stonehenge you'd need to COLUMN. POST was dropped with Woodhenge.

  10. gerryg

    You've all missed the the wonderful irony.

    See back issues of the Register for more details, for example:

    https://www.theregister.com/2007/11/06/highfield_tactics/

    Back in the days of the introduction of iPlayer, only available for Windows and then Apple but no Linux during a presentation at IET, Savoy Place, Linux users were described as wanting to steal content as various side steps were taken to watch the output.

    Long time ago but the word "thieves" might have been used.

    That iPlayer is on Linux architecture would be just too good to be true

  11. JBowler

    How old is Michael Palin?

    It's curious, I was a great fan of his until around 2011 then he disappeared, I don't know why.

    1. David 132 Silver badge

      Re: How old is Michael Palin?

      He's 78, if IMDB is to be believed (yeah, I could have gone to Wikipedia but CBA).

      And if you've never watched his late 70s show Ripping Yarns, you've missed a treat.

  12. martinusher Silver badge

    People fuss so....

    You do not need to rebuild your entire software and hardware infrastructure to convert an 'unsigned long' into an 'unsigned long long'. I haven't looked recently but all the embedded code that I've used made the switch ages ago, I forget when. If you've got legacy stuff that's stuck with an unsigned long then the time's going to be a bit off but its hardly the end of the world but these sorts of systems are only interested in relative time and the resulting glitch is only momentary. ("Speaking from experience")

    (BTW -- GPS has been stuck with a 32 bit time counter. The people who designed it were smart enough to have a mechanism for periodically resetting the epoch.)

    1. pklausner

      Re: People fuss so....

      32 bit? I'm only aware of a the infamous 10 bit week counter which occasionally breaks software trying to guesstimate the epoch from their build date or whatever the programmer fancied.

      https://en.wikipedia.org/wiki/GPS_week_number_rollover

  13. Anonymous Coward
    Anonymous Coward

    Big Black Clock

    After a recent reburb’, at least Big Ben should be okay in 2038+

    1. Tom 7 Silver badge

      Re: Big Black Clock

      I wouldnt bet on it - are they not refurbishing HOP? I can easily see that going badly.

      1. David 132 Silver badge
        Coat

        Re: Big Black Clock

        I heard they’re not using unmodified open source for that project - they’ve got their own Fawkes.

  14. Anonymous Coward
    Anonymous Coward

    This is just one...

    ...of many, many bugs and security problems hiding in old code that will need to be fixed in the next 20 years.

    It turns out that we're more used to computers being broken these days than we were in 1999...

    1. John G Imrie

      Re: This is just one...

      But where will we be able to get old pennies from in order to correctly balance the weights?

  15. Captain Boing

    UTC?

    unixtime, epochtime whatever you want to call it is not necessarily UTC - it's a representation of the system clock in seconds so it will go on whatever the system clock is set to.

    It can be UTC but I bet most such affected systems in Oz will stumble before those in Europe - because of the time difference. saying that it all goes belly up at xx:xx:xx UTC means the clocks in Oz must have some ephemeral immunity for that half day... probably more effort than the fix

    jus' sayin'

    1. Anonymous Coward
      Anonymous Coward

      Re: UTC?

      so it will go on whatever the system clock is set to.

      Which, in the case of UNIX and Linux, is defined as UTC.

  16. staringatclouds

    Y2K minor point

    "Y2K was caused by programmers saving space by storing years as two digits"

    No it wasn't, if it were a software problem a patch would have solved it

    It was caused by the hardware storing dates as 2 digits & because it was the hardware this meant replacing the clock chips which usually meant replacing the entire device

    Y2K was primarily a hardware problem, which is why it was such a mare to deal with

    1. Paul Crawford Silver badge

      Re: Y2K minor point

      Hardware 2-digit is the easy bit, just replace a lot of PCs, etc.

      Nope, it was all of those perky files and databases using 2 digits for year that were bothersome and various temporary fixes as well as proper fixes had to be done. Just look at the TLE (two line elements) that are still in use, the assumption there is 00-56 is 2000-2056 and 57-99 is 1957-1999

      Why 1957? Think Sputnik

    2. Anonymous Coward
      Anonymous Coward

      Nope...

      What we call the Y2K problem actually started showing up in the mid-70's in software that did 25-year forecasting. It wasn't a hardware problem then - it was the date representation in software and database/filestore.

      Hardware was only one iteration of the Y2K problem. It also existed at the OS level, the software level and the database level, and all combinations thereof. My favourite combination was the one where where the OS and software companies both had Y2K compliant versions, but the two versions weren't compatible with each other.

      By the time it got called the Y2K problem it was a "mare of an issue" because the scope was huge, the timescales rapidly diminishing and Management baulking because it was going to be expensive to deal with (the bulk of the work was finding out where you had the problems - the fixes were relatively easy).

      I had a front-seat view of this as I was not only a programmer at the time, but I was also working alongside the Y2K remediation team for our company - which meant that I had a view of all the issues for our company and all our customers.

      There were many and various software patches for OS and commercial software. It is true that hardware was harder to deal with (and yes, there were firmware patches as well), but it was never primarily a hardware problem. Hardware was just one of the problems.

    3. Twanky Silver badge

      Re: Y2K minor point

      "Y2K was caused by programmers saving space by storing years as two digits"

      If year (numbers) were stored as two digits (two bytes) then we'd be OK from the year 32766 BC to the year 32767 (assuming they didn't use unsigned 'word' values). The problem is they used a two character representation which could then be fairly easily converted to a single byte number for arithmetic. Seems daft now but simply concatenating strings to represent dates seemed a great idea.

      How were the pension companies (for example) dealing with their calculations for men born post 1935 who would be expected to retire in 2000 or later (age 65, in the UK)?

      1. martinusher Silver badge

        Re: Y2K minor point

        Its not really 'two bytes' but two columns in a punch card corresponding to two decimal digits. (Think EBCDIC rather than binary.)

    4. staringatclouds

      Re: Y2K minor point

      Thanks for the downvotes guys, the year I spent sorting out hardware issues must have been a total waste

  17. Kristian Walsh

    There seems to be a lot of confusion about what C considers to be the size of an int on 64-bit machines.

    Basically, the value of INT_MAX (for that is its real name) is whatever is defined in ‹limits.h› for your compiler; it does not necessarily have anything to do with the machine architecture.

    On my 64-bit Linux with a unchanged clang installation, I have 64-bit time_t, 64-bit size_t, but 32-bit int.

    $ uname -m

    x86_64

    $ cat test.c

    #include <limits.h>

    #include <time.h>

    #include <stdio.h>

    int main(int argc, char* argv[])

    {

    printf("INT_MAX = %d, sizeof time_t = %ld\n", INT_MAX, sizeof(time_t));

    return 0;

    }

    $ gcc test.c

    $ ./a.out

    INT_MAX = 2147483647, sizeof time_t = 8

    32-bit int is a de-facto default in C; after all, int is the type you use when you don’t care about exactly how big it is. If size of the variable is important, stdint.h has the type-names you should be using instead.

    If you need 64-bit ints, you probably know you do already, and are using uint64_t/int64_t…or intptr_t, ptrdiff_t, size_t and the other types that are guaranteed to match your address-space.

    All of this, is, however, completely irrelevant: there’s almost zero chance that the iPlayer backend was written in C. The software is most likely written in Java, which explicitly defines int as 32 bits. The fact that the developers used an ad-hoc approach to date/times rather than a proper library also points to Java, which before 1.8 (in 2014) had a rats-nest of badly-concieved date-time functions in the standard library. (Yes there were good third-party packages, particularly jodatime which became java.time, but devs have to realise that the built-ins are crap and go looking for a better library… before they start using the built-in functions extensively in code)

    I suspect the code does indeed have a defined constant of Integer.MAX_VALUE somewhere to mean “no expiry”. That’s 2038-01-19 at 03:14:07 but when passed to the presentation code, that would most likely have been moved back to the last full day preceding (thus “January 18., 2038”), because stating “available until xx. Month, year” can be read inclusively, promising availability until the end of that day. Not a big deal on something so far away, but an expiry of this Friday at midnight would generate a lot of annoyed users if presented as “available until [this Friday]”.

  18. bobbear

    I can't imagine the BBC lasting that long, so I doubt it's going to be a problem. As far as I am concerned, the sooner the BBC encrypt their channels and the licence fee is scrapped the better. It won't make one jot of difference to me as I long ago gave up watching their biased news and current affairs output, awful PC dramas, low-life soaps, reality rubbish and don't get me started on what passes for comedy these days and the bloated salaries they pay to the likes of Lineker et al.., but what I object most strongly to is having to pay the BBC tax to be able to legally receive and watch other advertising supported real time satellite and terrestrial 'free to air' content which I am also indirectly paying for in addition.

    If we ever get decent broadband in our area to support streaming as opposed to the current wet string distributed snail pace so-called "Superfast" offering, that's the day I cancel my BBC tax contribution..

  19. nsmhd

    This is on Sky’s system

    Based on the screen shot in the article this is a screen shot from the Sky + app.

    Sky have their own implementation of BBC iPlayer where Sky ingest the BBCs content to their servers and serve it to Sky STBs via a progressive download rather than streaming.

    They use this progressive download system because older Sky + HD STB do not support streaming.

    On SkyQ it is possible to access the native iPlayer app but also use Skys progressive download system.

    This means this could well be something on Skys infrastructure rather than the BBCs.

    It’s also worth pointing out that I’m sure the Sky + app is end of life now in favour of the Sky Go app.

  20. Pantagoon
    Mushroom

    Deluded Fools

    I love how optimistic people are that we will actually get to 2038.

  21. Graham 32

    "seven seconds after pi o'clock"?

    Never heard of it before. The internet took be to the Afrikaans version of Wikipedia which says

    > In Engels word daar van "pi o'clock" gepraat, wat 3:14:15 voorstel.

    I get the idea.

  22. Anonymous Coward
    Anonymous Coward

    I Totally Agree with the Defund The BBC Brigade...

    Why should I pay more for products via advertising budgets, to subsidize people watching garbage on commercial TV stations?!?

  23. ICL1900-G3 Bronze badge

    2038?

    I'll be dead by then, so one thing fewer to worry about.

  24. jollyboyspecial

    Bug?

    "We asked the BBC about iPlayer bug."

    What we're seeing isn't actually a bug as such is it?

  25. CrazyOldCatMan Silver badge

    annoying, smug Mac and Linux users

    *Smiles proudly*

POST COMMENT House rules

Not a member of The Register? Create a new account here.

  • Enter your comment

  • Add an icon

Anonymous cowards cannot choose their icon

Biting the hand that feeds IT © 1998–2022