Just a sample of the Echomail archive
Cooperative anarchy at its finest, still active today. Darkrealms is the Zone 1 Hub.
|    BINKD    |    Support for the Internet BinKD mailer    |    8,958 messages    |
[   << oldest   |   < older   |   list   |   newer >   |   newest >>   ]
|    Message 6,800 of 8,958    |
|    Dan Cross to Wilfred van Velzen    |
|    Re: -64 and -46 option missing in 101    |
|    23 Apr 20 07:30:55    |
      TID: Mystic BBS 1.12 A46       MSGID: 3:770/100 07736998       REPLY: 2:280/464 5ea07abe       TZUTC: 1200       On 22 Apr 2020 at 07:10p, Wilfred van Velzen pondered and said...                Wv> On 2020-04-22 12:54:00, you wrote to me:        Wv> DC> Sure. Which part of it?        Wv>         Wv> Why you don't trust binkd.              Ah. It's a lot of code, it's exquisitely complex,       and it's written in a weakly-typed, unsafe language.              The first time I tried to run it, I ran into a       memory corruption issue that had lain dormant for       two decades. Another memory-out-of-bounds bug       showed up when it was compiled with IPv6 support.       A filesystem permissions problem caused an       infinite loop with no diagnostic. On my system,       it mysteriously crashes *inside the Perl       interpreter* when parsing nodelists; I some       sort of memory corruption issue in binkd that       manifests itself inside the Perl shared object.              Finally, I'm not at all sure how resilient it       is to avoiding data corruption in the event of       failure (either binkd or the system).              At that point, I realized I was faced with a       decision: do I double down on binkd and attempt       to fix it, or do I write my own replacement?              The former involves ensuring compatibility with       all sorts of legacy environments that I'm neither       equipped nor interested in supporting (OS/2 and       Win9x with ancient compilers, for example: even       if I wanted to support those -- and let me be       clear, I do not -- I don't have a development       environment for them and while I could set one       up, I'm just not interested) and that require       workarounds for lacking features available in       modern standards. The whole suite has very       little in the way of unit-level tests, let alone       integration style tests.              And how much effort do I really want to put into       it, anyway?              The other option means I can make a clean break       from the past, write in a type- and memory-safe       language (I chose Go), write as many tests as I       like, and take advantage of built-in support for       modern testing practices etc. Further, I'm       beholden to no one but myself to maintain       compatibility. Since this is all just for fun       anyway, I chose that route.              --- Mystic BBS v1.12 A46 2020/04/20 (Windows/32)        * Origin: Agency BBS | Dunedin, New Zealand | agency.bbs.nz (3:770/100)       SEEN-BY: 1/123 57/0 90/1 103/705 120/340 601 153/250 154/10 203/0       SEEN-BY: 220/70 221/0 226/17 30 227/114 229/101 200 426 1014 240/5832       SEEN-BY: 249/109 307 317 267/800 280/464 5003 5555 288/100 292/854       SEEN-BY: 292/8125 310/31 317/3 340/1000 342/200 396/45 423/120 712/848       SEEN-BY: 770/0 1 100 340 772/0 1 210 220 230 500 2452/250       PATH: 770/100 1 280/464 229/426           |
[   << oldest   |   < older   |   list   |   newer >   |   newest >>   ]
(c) 1994, bbs@darkrealms.ca