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

Re: netbeans fails to build on Normal Windows build

Expand Messages
  • Alejandro Lopez-Valencia
    El Tue, 09 Dec 2003 18:10:50 -0000, Mike Williams ... Look at the way it ms managed in Make_cyg.mak and Make_ming.mak, modify and submit a patch to Bram. (In
    Message 1 of 5 , Dec 9, 2003
    • 0 Attachment
      El Tue, 09 Dec 2003 18:10:50 -0000, Mike Williams
      <mike.williams@...> escribió:


      >
      > Perhaps it is a problem in feature.h for Win32 - USE_ICONV is only
      > defined if FEAT_MBYTE is defined, which is dependent on FEAT_BIG
      > being defined, but this should be a NORMAL build.
      >
      > Any suggestions of the right way to go to fix this?
      >

      Look at the way it ms managed in Make_cyg.mak and Make_ming.mak, modify
      and submit a patch to Bram. (In other words, Win32 builds should use
      FEAT_BYTE unless there is a very good reson not to.)

      --
      The limits of my language are the limits of my world.
      Ludwig Wittgenstein
    • Mike Williams
      Hi ... My bad, I was getting confused from trying to work out the difference between when it built ok and when it didn t. The VC make file and feature.h are
      Message 2 of 5 , Dec 10, 2003
      • 0 Attachment
        Hi

        On 9 Dec 2003 at 13:27, Alejandro Lopez-Valencia wrote:

        > El Tue, 09 Dec 2003 18:10:50 -0000, Mike Williams
        > <mike.williams@...> escribió:
        > >
        > > Perhaps it is a problem in feature.h for Win32 - USE_ICONV is only
        > > defined if FEAT_MBYTE is defined, which is dependent on FEAT_BIG
        > > being defined, but this should be a NORMAL build.
        > >
        > > Any suggestions of the right way to go to fix this?
        > >
        >
        > Look at the way it ms managed in Make_cyg.mak and Make_ming.mak, modify
        > and submit a patch to Bram. (In other words, Win32 builds should use
        > FEAT_BYTE unless there is a very good reson not to.)

        My bad, I was getting confused from trying to work out the difference
        between when it built ok and when it didn't. The VC make file and
        feature.h are ok.

        This is actually a bug in netbeans.c on Windows. The check for EINTR
        is against the last Winsock error for which the errno.h EINTR value
        will never be returned. The fix is to redefine EINTR as WSAEINTR
        which seems to be the Winsock equivalent. Just need to undefine
        EINTR in case errno.h has been included in a header.

        Here is the patch that should fix the bug.

        *** netbeans.c.orig Wed Dec 10 09:41:07 2003
        --- netbeans.c Wed Dec 10 09:41:23 2003
        ***************
        *** 32,37 ****
        --- 32,39 ----
        * errno... */
        # define sock_errno WSAGetLastError()
        # define ECONNREFUSED WSAECONNREFUSED
        + # undef EINTR
        + # define EINTR WSAEINTR
        # define sock_write(sd, buf, len) send(sd, buf, len, 0)
        # define sock_read(sd, buf, len) recv(sd, buf, len, 0)
        # define sock_close(sd) closesocket(sd)


        TTFN

        Mike
        --
        A closed mouth gathers no feet.
      • Bram Moolenaar
        ... Thanks, I ll include it. -- From know your smileys : *
        Message 3 of 5 , Dec 24, 2003
        • 0 Attachment
          Mike Williams wrote:

          > On 9 Dec 2003 at 13:27, Alejandro Lopez-Valencia wrote:
          >
          > > El Tue, 09 Dec 2003 18:10:50 -0000, Mike Williams
          > > <mike.williams@...> escribi=F3:
          > > >
          > > > Perhaps it is a problem in feature.h for Win32 - USE_ICONV is only
          > > > defined if FEAT_MBYTE is defined, which is dependent on FEAT_BIG
          > > > being defined, but this should be a NORMAL build.
          > > >
          > > > Any suggestions of the right way to go to fix this?
          > > >
          > >
          > > Look at the way it ms managed in Make_cyg.mak and Make_ming.mak, modify
          > > and submit a patch to Bram. (In other words, Win32 builds should use
          > > FEAT_BYTE unless there is a very good reson not to.)
          >
          > My bad, I was getting confused from trying to work out the difference
          > between when it built ok and when it didn't. The VC make file and
          > feature.h are ok.
          >
          > This is actually a bug in netbeans.c on Windows. The check for EINTR
          > is against the last Winsock error for which the errno.h EINTR value
          > will never be returned. The fix is to redefine EINTR as WSAEINTR
          > which seems to be the Winsock equivalent. Just need to undefine
          > EINTR in case errno.h has been included in a header.
          >
          > Here is the patch that should fix the bug.

          Thanks, I'll include it.

          --
          From "know your smileys":
          *<|:-) Santa Claus (Ho Ho Ho)

          /// Bram Moolenaar -- Bram@... -- http://www.Moolenaar.net \\\
          /// Creator of Vim - Vi IMproved -- http://www.Vim.org \\\
          \\\ Project leader for A-A-P -- http://www.A-A-P.org ///
          \\\ Help AIDS victims, buy here: http://ICCF-Holland.org/click1.html ///
        Your message has been successfully submitted and would be delivered to recipients shortly.