• ForHonour v0.8

    From Andrew Leary@21:4/105 to Apam on Tue May 21 10:47:34 2019
    Hello Apam!

    The above version seems to have a similar segfault issue to Galactic Dynasty. From my kernel log:

    [18758770.486659] ForHonour[9769]: segfault at 0 ip 00007fca816000aa sp 00007ffcf52d5aa0 error 4 in libc-2.23.so[7fca81592000+1c0000]

    Regards,

    Andrew

    --- GoldED+/LNX 1.1.5-b20180707
    * Origin: Phoenix BBS * phoenix.bnbbbs.net (21:4/105)
  • From apam@21:1/125 to Andrew Leary on Wed May 22 02:41:33 2019
    Hi

    The above version seems to have a similar segfault issue to Galactic Dynasty.

    I don't think I've found the bug yet, as I can't seem to reproduce it,
    but I pushed some changes to my repository, could you try pulling that
    and see if it's any different?

    If no one else is seeing this, my suspicion is it could be related to permissions, and there was one spot I was assuming a file open succeeded,
    which might cause a null pointer exception like you're seeing. The fixes
    I pushed just log it as a problem - if it still crashes, well back to the drawing board :)

    Andrew

    --- MagickaBBS v0.12alpha (Linux/x86_64)
    * Origin: The Fat Sandwich - sandwich.magickabbs.com:2023 (21:1/125)
  • From Andrew Leary@21:4/105 to apam on Tue May 21 13:20:22 2019
    Hello apam!

    21 May 19 22:41, you wrote to me:

    The above version seems to have a similar segfault issue to
    Galactic Dynasty.

    I don't think I've found the bug yet, as I can't seem to reproduce it,
    but I pushed some changes to my repository, could you try pulling that
    and see if it's any different?

    Sure, I just need your repository's address.

    If no one else is seeing this, my suspicion is it could be related to permissions, and there was one spot I was assuming a file open
    succeeded, which might cause a null pointer exception like you're
    seeing. The fixes I pushed just log it as a problem - if it still
    crashes, well back to the drawing board

    I'll check the ownership and permissions on all of the game files. My BBS has each user with their own unix account, which are all members of the group bbs.

    I generally need to have directories for doors with permissions 0770 and the files with permissions 0660 to ensure that the mbse BBS administration account and all of the users will have permission to read and write the files as needed.

    Andrew

    --- GoldED+/LNX 1.1.5-b20180707
    * Origin: Phoenix BBS * phoenix.bnbbbs.net (21:4/105)
  • From apam@21:1/125 to Andrew Leary on Wed May 22 13:49:52 2019
    Hello apam!

    21 May 19 22:41, you wrote to me:

    The above version seems to have a similar segfault issue to
    Galactic Dynasty.

    I don't think I've found the bug yet, as I can't seem to reproduc
    but I pushed some changes to my repository, could you try pulling
    and see if it's any different?

    Sure, I just need your repository's address.

    it's git://git.magickabbs.com/forhonour-c

    the galactic dynasty one is

    git://git.magickabbs.com/galactic-dynasty

    I'll check the ownership and permissions on all of the game files.
    My BBS has
    each user with their own unix account, which are all members of the
    group bbs.

    My thinking was on the packet files and whatever user you run maintenance
    as.

    It's just a stab in the dark really.

    Andrew

    --- MagickaBBS v0.12alpha (Linux/x86_64)
    * Origin: The Fat Sandwich - sandwich.magickabbs.com:2023 (21:1/125)
  • From Andrew Leary@21:4/105 to apam on Wed May 22 03:58:32 2019
    Hello apam!

    22 May 19 09:49, you wrote to me:

    Sure, I just need your repository's address.

    it's git://git.magickabbs.com/forhonour-c

    the galactic dynasty one is

    git://git.magickabbs.com/galactic-dynasty

    I'll check the ownership and permissions on all of the game files.
    My BBS has each user with their own unix account, which are all
    members of the group bbs.

    My thinking was on the packet files and whatever user you run
    maintenance as.

    The game packets are created as owned by the mbse user when they are received by the mailer. They are created as 0600 when received, so it's possible this is the issue. When the maintenance failed was when it was run from the game script, which executes as the individual BBS user. Maybe I'll remove that from the game script and just run it as a cron job hourly instead.

    It's just a stab in the dark really.

    A good guess, I think. I'll see if I can verify it.

    Andrew


    --- GoldED+/LNX 1.1.5-b20180707
    * Origin: Phoenix BBS * phoenix.bnbbbs.net (21:4/105)