Loading ...
Sorry, an error occurred while loading the content.
 

vim60ad make test on hpux does a segv

Expand Messages
  • Mark Waggoner
    I built vim 60ad with several patches that have been posted on 6.0ad. Patches included are: colors patch from Bram syn_id_2attr() patch from Bram fold.c patch
    Message 1 of 2 , May 2, 2001
      I built vim 60ad with several patches that have been posted on 6.0ad.
      Patches included are:
      colors patch from Bram
      syn_id_2attr() patch from Bram
      fold.c patch from Bram
      Modified virtualedit patches from Chase Tingley
      (I sent this patch to the list earlier)

      make test on hpux results in a segv.
      make test on linux did not segv.

      It appears that the failure occurs during test16. Here's the output
      from:

      make test |& tee OUT

      cd testdir; make -f Makefile gui VIMPROG=../vim
      echo Test results: >test.log
      rm -f test16.failed test.ok
      cp test16.ok test.ok
      ../vim -u unix.vim -U NONE --noplugin -s dotest.in test16.in
      [?25I[?1h=[?25h[?25h"test16.in" 10 lines, 139 charactersTests for resetting "secure" flag after GUI has started.

      STARTTEST
      :set exrc secure
      :gui -f
      :.,$w! test.out
      :qa!
      ENDTESTjust some text
      ~
      ~
      ~
      ~
      ~
      ~
      ~
      ~
      ~
      ~
      ~
      ~
      ~
      ~
      ~
      ~
      ~
      ~
      ~
      ~]2;test16.in (~/src/vim60ad/src/hpux10.20/testdir) - VIM]1;test16.in[?25h[?25I:[?25hset cp[?25I[?25h[?25I:[?25hmap dotest /^STARTTEST^^Mj:set ff=unix cpo-=A^^M:.,/ENDTEST/-1w! Xdotest^^M:set ff& cpo+=A^^Mnj0:so! Xdotest^^Md
      otest[?25ITests for resetting "secure" flag after GUI has started.[?25h[?25I/^STARTTEST[?25h
      [?25I:set ff=unix cpo-=A[?25h[?25I:.,/ENDTEST/-1w! Xdotest"Xdotest" 4 lines, 46 characters written[?25h[?25I:set ff& cpo+=A[?25h[?25I/ENDTEST[?25h
      [?25I:so! Xdotest[?25h[?25I:[?25hset exrc secure[?25I[?25h[?25I:[?25hgui -f[?25I[?25h]2;phpc0157:/fs5/c/waggoner/src/vim60ad/src/hpux10.20/]1;phpc0157:/fs5/c/waggoner/src/vim60ad/src/hpux10.20/[?1l>|30;0MVim: Caught deadly signal SEGV
      Vim: Finished.
      |30;0M*** Termination signal 139

      Stop.
      *** Error exit code 1

      Stop.



      Here's gdb backtrace:

      (gdb) where
      #0 0xc01411e0 in kill ()
      #1 0xf59c0 in may_core_dump ()
      #2 0xf596c in mch_windexit ()
      #3 0xabb48 in getout ()
      #4 0xc96b0 in preserve_exit ()
      #5 0xf3cd0 in deathtrap ()
      #6 <signal handler called>
      #7 0xc009fc80 in ltostr ()
      #8 0xc00a2940 in memorymap ()
      #9 0xc00a0964 in free ()
      #10 0xcb7e4 in vim_free ()
      #11 0x7239c in clear_var ()
      #12 0x6ac58 in ex_let ()
      #13 0x82324 in do_one_cmd ()
      #14 0x80b7c in do_cmdline ()
      #15 0x7fc64 in do_source ()
      #16 0x1341ec in gui_init ()
      #17 0x123fd0 in set_termname ()
      #18 0x124938 in termcapinit ()
      #19 0x133e8c in gui_start ()
      #20 0x138d74 in ex_gui ()
      #21 0x82324 in do_one_cmd ()
      #22 0x80b7c in do_cmdline ()
      #23 0xdbedc in nv_colon ()
      #24 0xd69a8 in normal_cmd ()
      #25 0xab960 in main_loop ()
      #26 0xab5d8 in main ()
      (gdb)


      I'm not sure what to look at - if there is more I can track down, please
      let me know.

      --------------------------------------------------------------------------
      Mark Waggoner waggoner@... (503) 712-3335
      Like "Constantinople" and "Timbuktu"
    • Mark Waggoner
      It appears that this was due to the use of expandpath(), which has been noted to be a problem, in menu.vim. The following patch fixes it: ... ***************
      Message 2 of 2 , May 2, 2001
        It appears that this was due to the use of expandpath(), which has been
        noted to be a problem, in menu.vim.

        The following patch fixes it:

        *** runtime/menu.vim.orig Wed May 2 12:05:19 2001
        --- runtime/menu.vim Wed May 2 12:05:48 2001
        ***************
        *** 275,282 ****
        endfun

        " Setup the Edit.Color Scheme submenu
        ! if exists("*expandpath")
        ! let s:n = expandpath("colors/*.vim", &runtimepath)
        let s:idx = 100
        while strlen(s:n) > 0
        let s:i = stridx(s:n, "\n")
        --- 275,282 ----
        endfun

        " Setup the Edit.Color Scheme submenu
        ! if exists("*globpath")
        ! let s:n = globpath(&runtimepath,"colors/*.vim")
        let s:idx = 100
        while strlen(s:n) > 0
        let s:i = stridx(s:n, "\n")
        ***************
        *** 300,306 ****
        " Setup the Edit.Keymap submenu
        if has("keymap")
        amenu 20.460.90 &Edit.Keymap.None :set keymap=<CR>
        ! let s:n = expandpath("keymap/*.vim", &runtimepath)
        let s:idx = 100
        while strlen(s:n) > 0
        let s:i = stridx(s:n, "\n")
        --- 300,306 ----
        " Setup the Edit.Keymap submenu
        if has("keymap")
        amenu 20.460.90 &Edit.Keymap.None :set keymap=<CR>
        ! let s:n = globpath(&runtimepath, "keymap/*.vim")
        let s:idx = 100
        while strlen(s:n) > 0
        let s:i = stridx(s:n, "\n")

        On Wed, 2 May 2001, Mark Waggoner wrote:

        > Date: Wed, 2 May 2001 10:07:55 -0700 (PDT)
        > From: Mark Waggoner <waggoner@...>
        > To: Vim Developer List <vim-dev@...>
        > Subject: vim60ad make test on hpux does a segv
        >
        >
        > I built vim 60ad with several patches that have been posted on 6.0ad.
        > Patches included are:
        > colors patch from Bram
        > syn_id_2attr() patch from Bram
        > fold.c patch from Bram
        > Modified virtualedit patches from Chase Tingley
        > (I sent this patch to the list earlier)
        >
        > make test on hpux results in a segv.
        > make test on linux did not segv.
        >

        blah blah blah...

        --------------------------------------------------------------------------
        Mark Waggoner waggoner@... (503) 712-3335
        Haven't you heard about the BIG HUNGRY BEAR?
      Your message has been successfully submitted and would be delivered to recipients shortly.