[xine-bugs] [ xine-Bugs-1114677 ] hstrerror on solaris is in libresolv.so

SourceForge Headquarters 225 Broadway Suite 1600 San Diego, CA 92101 +1 (858) 454-5900

Bugs item #1114677, was opened at 2005-02-02 13:17
Message generated for change (Comment added) made by valtri
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=109655&aid=1114677&group_id=9655

Category: portability problem
Group: None
>Status: Closed
>Resolution: Fixed
Priority: 5
Submitted By: Niki W. Waibel (nikiwaibel)
Assigned to: Nobody/Anonymous (nobody)
Summary: hstrerror on solaris is in libresolv.so

Initial Comment:
the hstrerror check in configure.ac always fails on solaris, because it 
does not use -lresolve to do the check.

maybe an
AC_CHECK_LIB (LIBRARY, FUNCTION, [ACTION-IF-FOUND],
          [ACTION-IF-NOT-FOUND], [OTHER-LIBRARIES])
test should be done as well if the function is not found in the std c lib.

----------------------------------------------------------------------

>Comment By: Frantisek Dvorak (valtri)
Date: 2005-02-03 08:24

Message:
Logged In: YES 
user_id=543511

The patch applied,  I'm closing the bug.

Thanks you,
  Frantisek

----------------------------------------------------------------------

Comment By: Niki W. Waibel (nikiwaibel)
Date: 2005-02-03 04:25

Message:
Logged In: YES 
user_id=1951

118363: xine-lib_cvs20050202.solaris.nww.patch
is fine. no problems on all mentioned architectures.

(patch includes patch from bug id # 1115001 and 
has beeb applied and checked against the cvs code 20050202)

----------------------------------------------------------------------

Comment By: Niki W. Waibel (nikiwaibel)
Date: 2005-02-03 03:11

Message:
Logged In: YES 
user_id=1951

PRIXMAX appears in
./src/demuxers/demux_film.c
./src/demuxers/demux_flv.c

and i guess that PRIiMAX can be changed to PRIdMAX in
./po/cs.po
as well.

so here is the new sol patch including 118357: solaris_musepack.diff from 
bug # 1115001.

----------------------------------------------------------------------

Comment By: Niki W. Waibel (nikiwaibel)
Date: 2005-02-03 01:59

Message:
Logged In: YES 
user_id=1951

the current patch (118352: solaris.diff) does not have the s/PRIXMAX/
PRIxMAX/g changes...

----------------------------------------------------------------------

Comment By: Frantisek Dvorak (valtri)
Date: 2005-02-03 01:06

Message:
Logged In: YES 
user_id=543511

Hi Niki,

many thanks for testing and exploring on Solaris! I'll apply
your ideas and hstrerror patch. 

Can you please test the new attached patch? If it will work,
I'll commit it.

Thank you,
  Frantisek

----------------------------------------------------------------------

Comment By: Niki W. Waibel (nikiwaibel)
Date: 2005-02-02 20:10

Message:
Logged In: YES 
user_id=1951

there is also a problem with PRIXMAX on solaris. there is only 
PRIxMAX. the following solved this issue:
===
find xine-lib_cvs20040202 -type f -exec grep -q PRIXMAX {} \;         -exec sed -i -e "s/PRIXMAX/PRIxMAX/g" {} \; -print
===

----------------------------------------------------------------------

Comment By: Niki W. Waibel (nikiwaibel)
Date: 2005-02-02 19:48

Message:
Logged In: YES 
user_id=1951

oops. to change all PRIiMAX to PRIdMAX you need:
===
find . -type f -exec grep -q PRIiMAX {} \;         -exec sed -i -e "s/PRIiMAX/PRIdMAX/g" {} \; -print
===
sorry.

----------------------------------------------------------------------

Comment By: Niki W. Waibel (nikiwaibel)
Date: 2005-02-02 19:34

Message:
Logged In: YES 
user_id=1951

independed to the hstrerror problem there is the PRIiMAX problem, also 
mentioned in bug #1062987. my solution for that is:
===
         find xine-lib_cvs20040202 -type f -exec grep -q PRIiMAX {} \;                 -exec sed -i -e "s/PRIiMAX/PRIdMAX/" {} \; -print
===
i guess that PRIdMAX is "more" standard then PRIiMAX.

i also guess that you could do a check in configure.ac easily, but i think 
it is easier to change it to PRIdMAX.

somehow i do not like the idea of
#ifdef (sun)
#define ...
#endif
of the patch in bug #1062987...

----------------------------------------------------------------------

Comment By: Niki W. Waibel (nikiwaibel)
Date: 2005-02-02 19:13

Message:
Logged In: YES 
user_id=1951

the bug looks similar, but the patch mentioned does NOT work!!!

i still have:
/* Define to 1 if you have 'hstrerror' in <netdb.h> */
/* #undef HAVE_HSTRERROR */
in config.h, which is WRONG.

here is a snip from config.log:
===
configure:42359: gcc -o conftest -Wall -O2 -I/misc/sparc-sun-solaris2.9/
include -I/usr/X11R6/include -I/usr/openwin/include -I/misc/sparc-sun-
solaris2.9/include/ffmpeg -I/misc/sparc-sun-solaris2.9/include/postproc -I/
misc/sparc-sun-solaris2.9/include -I/usr/X11R6/include -I/usr/openwin/include 
-L/misc/sparc-sun-solaris2.9/lib -R/misc/sparc-sun-solaris2.9/lib -L/misc/
sparc-sun-solaris2.9/lib -L/usr/X11R6/lib64 -L/usr/X11R6/lib -L/usr/openwin/
lib conftest.c -lposix4  >&5
conftest.c: In function `main':
conftest.c:4: warning: control reaches end of non-void function
/var/tmp//ccAPUlf5.o(.text+0x8): In function `main':
conftest.c: undefined reference to `hstrerror'
collect2: ld returned 1 exit status
configure:42365: $? = 1
configure: failed program was:
| #include <netdb.h>
| int main(void) {
|   hstrerror(0);
| }
===
if you want to RUN the test program, you need to compile it
with -lresolve on solaris (tested on sol8-sparc/sol9-sparc/sol9-x86).

the patch from bug #1062987:
needs to be EXTENDED by sthg like:
===
AC_CHECK_LIB([resolv], [hstrerror], [ac_cv_function_system_hstrerror="yes"]
)
===

i've attached the patch with works with the current cvs code.

----------------------------------------------------------------------

Comment By: Michael Roitzsch (mroi)
Date: 2005-02-02 13:48

Message:
Logged In: YES 
user_id=552060

Could you check bug #1062987, please? Especially try the 
attached patch. Maybe this bug here is even a duplicate of 
#1062987. 

Michael 

----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=109655&aid=1114677&group_id=9655