Re: UGLY_DEFS for darwin in configure.in

[John W. Eaton]

On  3-Oct-2007, Thomas Treichl wrote:

| John W. Eaton schrieb:
| > On 23-Sep-2007, Thomas Treichl wrote:
| > 
| > | I've done a lot of tests and I spend quite a lot of time to see if the 
lines
| > | 
| > |    *-*-darwin*)
| > |      UGLY_DEFS=`echo $DEFS | sed 
's,\\",\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\",g'`
| > |    ;;
| > | 
| > | in configure.in of the current Octave 2.9.14 sources are still necessary. 
The 
| > | big problem that I've got is that I actually am not able to reproduce the 
| > | problem. This means that I'm not able to compile Octave with MacOSX's 
original 
| > | 'sed' at all, the problems are much bigger then just setting the 
UGLY_DEFS in a 
| > | different way than on any other *nix platform (The reason may be that 
either 
| > | MacOSX's sed is too much POSIX or too much broken in some way, I actually 
don't 
| > | know).
| > | 
| > | My wish would be that these lines could be completely be removed from the 
| > | ./configure.in script but only if we are sure that there is absolutely no 
other 
| > | MacOSX distribution packaging process broken.
| > 
| > I think the only place this is really used now is the octave-bug
| > script.  The idea is that when someone reports a bug, we'd like to
| > know what macros are defined when Octave was compiled.  Currently,
| > running the octave-bug scripts prints something like this:
| > 
| >   DEFS:
| > 
| >     -DPACKAGE_NAME="" -DPACKAGE_TARNAME="" -DPACKAGE_VERSION=""
| >     -DPACKAGE_STRING="" -DPACKAGE_BUGREPORT="" -DOCTAVE_SOURCE=1
| >     -D_GNU_SOURCE=1 -DSTDC_HEADERS=1 -DHAVE_SYS_TYPES_H=1
| >     -DHAVE_SYS_STAT_H=1 -DHAVE_STDLIB_H=1 -DHAVE_STRING_H=1
| >     -DHAVE_MEMORY_H=1 -DHAVE_STRINGS_H=1 -DHAVE_INTTYPES_H=1 
-DHAVE_STDINT_H=1
| >     ...
| > 
| > All the fancy quoting is needed to try to get these lines passed
| > through Make and the shell so that they appear correctly quoted in the
| > text of the message that the octave-bug script finally sends.
| > 
| > I'd be happy to have these lines replaced by the contents of config.h,
| > but I don't see how to do that reliably.
| > 
| > jwe
| 
| Hallo John,
| 
| I was talking to Marius (HMUG) who finally told me offside the list that the 
| 'commented out' three lines are okay for the HMUG version of Octave. I didn't 
| get any answer from the Fink packagers but I was talking to Ben who is 
running a 
| Fink version of Octave (and I had a look at the patch file that is used 
before 
| Octave is build - the commented out three lines aren't there). The output of 
| Fink looks like
| 
|    ans = ... -DPACKAGE_TARNAME=\\ -DPACKAGE_VERSION=\\ -DPACKAGE_STRING=\\ ...
| 
| instead of
| 
|    ans = ... -DPACKAGE_TARNAME="" -DPACKAGE_VERSION="" -DPACKAGE_STRING="" ...

What do you mean by "output of Fink"?  In any case, the \\ is wrong.
It should be "".

| So my final summary is: Fink's output is different, for HMUG the commented 
out 
| lines are okay and MacPorts as well as Octave.app already use the commented 
out 
| lines. I'm sorry that I don't have better news or infos but I think it's time 
to 
| get the *darwin* specials out of configure.in. If you say ok, should I send a 
| patch for these three lines?

I'm all for getting rid of UGLY_DEFS, but I won't want to do that
unless we can find another way to reliably save the #defines from
config.h in the bug reporting script.

jwe