User Tag List

Page 1 of 2 1 2 LastLast
Results 1 to 10 of 16

Thread: [2.50.2] Negative Max Float Error?

  1. #1
    The Timelord
    QDB Manager
    ZC Developer

    Join Date
    Oct 2006
    Location
    Prydon Academy
    Posts
    1,396
    Mentioned
    112 Post(s)
    Tagged
    1 Thread(s)
    vBActivity - Stats
    Points
    4,760
    Level
    21
    vBActivity - Bars
    Lv. Percent
    68.7%

    Sick Bug [2.50.2] Negative Max Float Error?

    Apparently, the negative max constant is -021473.0001

    I checked, and negative max int is -214747, negative max float is -214747.9999. This only affects constants.

    This apparently is true in 2.50.2, 2.50.1, and 2.50.0.

  2. #2
    Here lies mero. Died by his own dumbassitude.
    Join Date
    May 2011
    Posts
    929
    Mentioned
    102 Post(s)
    Tagged
    2 Thread(s)
    vBActivity - Stats
    Points
    5,527
    Level
    23
    vBActivity - Bars
    Lv. Percent
    13.96%
    Not a bug, just the way things work.

  3. #3
    The Timelord
    QDB Manager
    ZC Developer

    Join Date
    Oct 2006
    Location
    Prydon Academy
    Posts
    1,396
    Mentioned
    112 Post(s)
    Tagged
    1 Thread(s)
    vBActivity - Stats
    Points
    4,760
    Level
    21
    vBActivity - Bars
    Lv. Percent
    68.7%
    Quote Originally Posted by Tamamo View Post
    Not a bug, just the way things work.
    Well, one way or another, it's something that deserves correction, as max signed type constant should always match max signed type variable.

    These values sum 236221.0, which doesn't seem to indicate anything.. Signed, or unsigned, long or fixed, that value doesn't make any sense.

    21473.0001 is derived, how?
    @Gleeok : Confirmation requested.

  4. #4
    The Time-Loop Continues ZC Developer
    Gleeok's Avatar
    Join Date
    Apr 2007
    Posts
    4,815
    Mentioned
    259 Post(s)
    Tagged
    10 Thread(s)
    vBActivity - Stats
    Points
    12,933
    Level
    33
    vBActivity - Bars
    Lv. Percent
    23.44%
    Probably a bug.
    This post contains the official Gleeok seal of approval. Look for these and other posts in an area near you.

  5. #5
    The Timelord
    QDB Manager
    ZC Developer

    Join Date
    Oct 2006
    Location
    Prydon Academy
    Posts
    1,396
    Mentioned
    112 Post(s)
    Tagged
    1 Thread(s)
    vBActivity - Stats
    Points
    4,760
    Level
    21
    vBActivity - Bars
    Lv. Percent
    68.7%
    Quote Originally Posted by Gleeok View Post
    Probably a bug.
    that's what I'm feeling here. The value to handle them should be a signed long, and thus identical in both directions. Technically, it could extend in one direction by .0001, which might have been the intent, and how we got this screwy value.

    I have no idea why Tamamo flagged it 'not a bug', but then again, I can't find this in the [ ... ] material.

    I think the value is simply input improperly, but I'm having a fun time trying to find where this is handled... I swear, this shyte is going to drive me doubly out of my flipping mind. I say just make it 214747.9999 in both directions, and be done...if I can ever locate it. Then again, waiting a week, would be better.

  6. #6
    Here lies mero. Died by his own dumbassitude.
    Join Date
    May 2011
    Posts
    929
    Mentioned
    102 Post(s)
    Tagged
    2 Thread(s)
    vBActivity - Stats
    Points
    5,527
    Level
    23
    vBActivity - Bars
    Lv. Percent
    13.96%
    The reason I didn't think it was a bug is because they use a custom type (fix) don't they? If not then...
    From what saffith told me all numbers in zscript are the same thing.


  7. #7
    The Timelord
    QDB Manager
    ZC Developer

    Join Date
    Oct 2006
    Location
    Prydon Academy
    Posts
    1,396
    Mentioned
    112 Post(s)
    Tagged
    1 Thread(s)
    vBActivity - Stats
    Points
    4,760
    Level
    21
    vBActivity - Bars
    Lv. Percent
    68.7%
    I'm pretty sure those are just for Allegro things, using the Fix class. I don't even see the fix.h file anywhere, or know where the defs, and related functions are being loaded. I skimmed to find that in the past, but it isn't among my files, so ...

    Either way, a fixed width integer can be 32b signed. If the values aren't 32b signed longs, they're 32b wide signed fixed ints. I'm pretty sure they are signed longs though... I'll probably wait for the 'promised day' to do anything more about this, but I did again glance through the lexer, and now need more alcohol. We should probably unflag this topic for now.

    For what it's worth, the lexer also has a bug that allows compilation of a script with a ? token i it. It gives a lexical error warning, but compiles it anyhow; which, as was reported to me, can be a problem if a user hits ? instead of / in a calculation. It's because the lexer defines ? and ^ as %token SINGLECHAR, and is set up to report a lexical error for them, but not to halt compilation. :/

    That's more of an oversight than a true bug, but it feels as if it'll be mad easy to correct.

  8. #8
    Here lies mero. Died by his own dumbassitude.
    Join Date
    May 2011
    Posts
    929
    Mentioned
    102 Post(s)
    Tagged
    2 Thread(s)
    vBActivity - Stats
    Points
    5,527
    Level
    23
    vBActivity - Bars
    Lv. Percent
    13.96%
    it seems to be 18 bits but doesn't use all of them, weird.
    The fraction is separate of course and is truncated to 62 bit.

    Definitely a bug.

  9. #9
    The Timelord
    QDB Manager
    ZC Developer

    Join Date
    Oct 2006
    Location
    Prydon Academy
    Posts
    1,396
    Mentioned
    112 Post(s)
    Tagged
    1 Thread(s)
    vBActivity - Stats
    Points
    4,760
    Level
    21
    vBActivity - Bars
    Lv. Percent
    68.7%
    Quote Originally Posted by Tamamo View Post
    it seems to be 18 bits but doesn't use all of them, weird.
    The fraction is separate of course and is truncated to 62 bit.

    Definitely a bug.
    Please send me a PM with the filenames, and line number (the latter, if possible) for where you saw it, so that I can find it readily.

  10. #10
    Here lies mero. Died by his own dumbassitude.
    Join Date
    May 2011
    Posts
    929
    Mentioned
    102 Post(s)
    Tagged
    2 Thread(s)
    vBActivity - Stats
    Points
    5,527
    Level
    23
    vBActivity - Bars
    Lv. Percent
    13.96%
    I didn't, I took the numbers you gave us, and plugged them into my brain which outputs binary anyways.
    WHere are you @Saffith , if your there get your ass in here.

Thread Information

Users Browsing this Thread

There are currently 3 users browsing this thread. (0 members and 3 guests)

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •  
About us
Armageddon Games is a game development group founded in 1997. We are extremely passionate about our work and our inspirations are mostly drawn from games of the 8-bit and 16-bit era.
Social