Re: [XForms] HP-UX - Itanium -- Error during installation

[Jens Thoms Toerring]

Hi Donto,

On Tue, Aug 30, 2011 at 02:07:44PM +0000, Donato Lorenzo López wrote:
>   I'm using XForms (xforms-1.0.93sp1) on a server wich has:
>  
>               S.O.: HP-UX B.11.31 U ia64 (a.k.a. HPUX 11i v3 Update8 (March 
> 2011))
>               aCC: HP C/aC++ B3910B A.06.25 [Nov 30 2009]
> 
> Installing the library, I had to "fix" some things. I'll explain the steps I 
> took. I commented the error lines, but I guess there is a better option! :)
> 
> I hope it can serve as some kind of help for you. 

Thank you!

> Notes: My installation will be:
>                - in /usr/local/ folder
>                - under 32bit
> 
> 1.- Configure
>        ./configure CC="aCC -Ae +DD32" CFLAGS="-I/usr/include 
> -I/usr/local/include" LDFLAGS="-L/usr/lib -L/usr/local/lib"
> 
> 2.- I had to modify the Makefile files created by previous step 
> (.../lib/Makefile and .../image/Makefile)
>      I had to remove the "--no-undefined" parameter, because it's not 
> recognized by the linker. If I try to launch "make" I'll obtain:
>           ld: Unrecognized argument: --no-undefined)
> 
>     The change is the same in both Makefiles:
>          ## libforms_la_LDFLAGS
> = -no-undefined -Wl,--no-undefined -version-info 2:0:0
>         libforms_la_LDFLAGS
> = -no-undefined -Wl, -version-info 2:0:0

I will have to find a fix that works a bit differently since
those files are created automatically in the normal process
of creating the library, but yours is a good starting point.

> 3.- I "fixed" this error:
> 
>       error #2070: incomplete type is not allowed
>       obj->spec      = fl_calloc( 1, sizeof *obj->spec );
> 
>       The error is in the files: 
>                                           .../lib/clock.c
>                                           .../fd2ps/load.c
>                                           .../fdesign/fd_fake.c
> 
>      I simply commented this lines... yeahh, I like to make things easier.... 
> :)

Mmm, I don't think that's really a "fix" since it leaves some
pointer in the object structure undefined (pointing to some
random place in memory) that then has a good chance of being
used by the program later on - and that then could lead to a
segmentation fault, or worse, just weird behaviour. I'll take
a look how to fix this differently. Perhaps you allow me to ask
for a test by you (I don't have access to a machine with HP-UX)
when I'm done?

> 4.- I "fixed" another error, in the .../lib/spinner.c file:
> 
> "spinner.c", line 164:
> error #2028: expression must have a constant value
> 
>       char buf[ max_len ];
> 
> "spinner.c",
> line 432: error #2028: expression must have a constant value
> 
>       char buf[ max_len ];
> 
> "spinner.c",
> line 452: error #2028: expression must have a constant value
> 
>       char buf[ max_len ]; 
> 
>     Yeahh, you can imagine my great solution:
>  
>                      [line 164] char buf[ 500 ];
>                      [line 432] char buf[ 500 ];
>                      [line 452] char buf[ 500 ];

That's one way to fix it. I guess I would prefer to use a some-
what smaller buffer, perhaps allocated.

> 5.- I launch "make"
> 
> 6.- I launch "make install"
> 
> 7.- I compile and excute my programs using XForms and, apparently, I got no 
> error because of my "great fixes"... 
> 
> Well, the point is one of the programs is having an unexpected behaviour, but 
> I'll send another mail about that...

Thanks, don't hesitate;-)

> I hope it helps you.

Yes, definitely! I will try to integrate your fixes (but it pro-
bably take until the weekend) and then upload a new pre-release.
I hope you don't mind if I then send you an email and ask you if
you're prepared to check the revised version/

                      Thanks and best regards, Jens
-- 
  \   Jens Thoms Toerring  ________      address@hidden
   \_______________________________      http://toerring.de