Forums before death by AOL, social media and spammers... "We can't have nice things"
|    comp.lang.forth    |    Forth programmers eat a lot of Bratwurst    |    117,927 messages    |
[   << oldest   |   < older   |   list   |   newer >   |   newest >>   ]
|    Message 117,431 of 117,927    |
|    dxf to Hans Bezemer    |
|    Re: Parsing timestamps?    |
|    08 Jul 25 13:17:28    |
      From: dxforth@gmail.com              On 7/07/2025 9:21 pm, Hans Bezemer wrote:       > On 07-07-2025 05:48, dxf wrote:       >> ...       >> He appears to dislike the idea of standard-imposed minimums (e.g. Carter's       >> suggestion of 16) but suggested:       >>       >> a) the user can offload to memory if necessary from       >> fpu hardware;       >>       >> b) an ANS FLOATING and FLOATING EXT wordset includes       >> the necessary hooks to extend the fp stack.       >       > In 4tH, there are two (highlevel) FP-systems - with 6 predetermined       configurations. Configs number 0-2 don't have an FP stack, they use the       datastack. 3-5 have a separate FP stack - and double the precision. The       standard FP stacksize is 16, you can        extend it by defining a constant before including the FP libs.              Given the ANS minimum of 6 and recognizing that merely displaying an fp number       can       consume several positions I added another 5 as headroom. I organized it such       that       if the interpreter or ?STACK was invoked, anything more than 6 would report the       overflow:               DX-Forth 4.60 2025-06-25               Software floating-point (separate stack)               1e 2e 3e 4e 5e 6e .s 1. 2. 3. 4. 5. 6.000001 |
[   << oldest   |   < older   |   list   |   newer >   |   newest >>   ]
(c) 1994, bbs@darkrealms.ca