home *** CD-ROM | disk | FTP | other *** search
- In article <ROLF.93Apr22144052@green.mathematik.uni-stuttgart.de>, rolf@green.mathematik.uni-stuttgart.de (Rolf Schreiber) writes:
- > Hi,
- >
- > I like lemacs (19.4) very much, so I tried to install 19.6 on a RS6000
- > under AIX 3.2 .
- >
- > I applied the patches from bobhays@spss.com (Bob Hays), posted here a
- > week or so ago.
- >
- > Additionally I had to change the following line in config.h:
- >
- > #define LD_SWITCH_SITE -L/usr/lpp/X11/Xamples/lib/
-
- You didn't make links into /usr/lib for some of the X11 stuff that IBM
- kindly doesn't place in the default location (NOTE: I think the 3.2.2
- install here did this, so I wonder why you didn't get the links).
-
- > and to insert a
- >
- > #undef HAVE_TIMEVAL
- >
- > line in s-aix3-2.h (I don't know what I'm doing here, but it works for me).
-
- Yep, now that the time stuff is turned on, you need to tell config how
- to do time work (there are at least two major ways with lots of little
- variances - see editfns.c for some ideas:-). You did the right thing.
-
- > Elsewhere there were the following errors:
- > cc -O -Q -D_BSD -Demacs -I./lwlib -c editfns.c
- > "./systime.h", line 206.0: 1506-236 (W) Macro name EMACS_GET_TZ_OFFSET has been redefined.
- > "editfns.c", line 663.15: 1506-007 (S) Struct or union timeval is undefined.
- > "editfns.c", line 663.62: 1506-007 (S) Struct or union timezone is undefined.
- > "editfns.c", line 663.94: 1506-045 (S) Undeclared identifier t.
- > "editfns.c", line 663.99: 1506-045 (S) Undeclared identifier dummy.
- > The error code from the last failed command is 1.
- >
- > Then it compiled without any problems.
-
- Huh? Then it compiled without any problems? After what change?
-
- Anyway, I found a diff that may have missed the set I posted, in systime.h:
-
- *** systime.h Tue Apr 13 14:17:25 1993
- --- systime.h-dist Tue Apr 13 14:16:46 1993
- ***************
- *** 35,41 ****
-
- /* AIX needs both <sys/time.h> and <time.h>. */
- #ifdef _AIX
- - #include <sys/time.h>
- #include <time.h>
- #endif
-
- --- 35,40 ----
-
- This forces sys/time.h and time.h to be included; I wish IBM would fix
- their header files to work right:-). However, I'd like Jamie to check
- this patch and make sure it is needed - the time stuff is really
- complex (IMHO) due to all the branch paths for #ifdefs.
-
- > Now I've got this problems:
- >
- > - gnus still does not work (it didn't either under 19.4). Starting
- > gnus hangs up emacs totally.
-
- There was a fix for this posted by another kind netter (the fix was
- originally posted in the lucid-bugs list, I think) which is:
-
- *** process.c Mon Apr 19 16:16:54 1993
- --- process.c-dist Mon Apr 19 16:14:55 1993
- ***************
- *** 1541,1553 ****
- #endif
- #endif
-
- - #ifdef AIX
- - {
- - int one = 1;
- - ioctl(inch, FIONBIO, &one);
- - }
- - #endif
- -
- XPROCESS (proc)->childp = host;
- XPROCESS (proc)->command_channel_p = Qnil;
- XPROCESS (proc)->buffer = buffer;
- --- 1541,1546 ----
-
- Sorry that I didn't apply this when I first built 19.6, but I wanted
- to see if the problem had been fixed with the 19.6 release (and GNUs
- wasn't a priority here, so I tried it a few days after banging on our
- 19.6 build).
-
- > - query-replace works well, but query-replace-regexp inserts a "/" for
- > every character it would have to substitute. Example:
- > M-x query-replace-regexp abc RET xyzzyx RET
- > makes a 0//////def out of 0abcdef
- > The same problem was with 19.4
- >
- > - When I mark some Files with m or with d for deletion, there shows up
- > a "/"-sign in front of this line, but the files aren't marked
- > (hitting x results in (No deletions requested))
- >
- > - In shell-mode appears a "/" instead of a newline-character when I
- > hit return. e.g: hitting Return five times results in a line:
- > $ /$ /$ /$ /$
-
- These I haven't seen yet - I'll check them out here and see what I
- get.
-
- > I hope, that you can help me with some of that problems.
-
- Hopefully. Have fun! - Bob
-
- --
- ===============================================================================
- The opinions expressed above are those of Bob Hays and not SPSS, Inc.
- bobhays@spss.com Phone: (312) 329-3529 | "I do not remember the anal probe"
- Fax: (312) 329-3657 | - Passion Fish
-
-