[elinks-users] renderer.c:1685: error: 'struct cache_entry' has no member named 'cache_mode' when compiling eLinks v0.11.3 in Debian (PowerPC)

Phillip Pi ant at zimage.com
Sun Jul 1 14:49:23 PDT 2007


On Mon, Jul 02, 2007 at 12:25:51AM +0300, Kalle Olavi Niemitalo wrote:
> 
> > I still got errors. http://www.savefile.com/files/856384 for the zip
> > file (included config.h and config.log after running make).
> 
> The build got past renderer.o, however, so I think the workaround
> is valid for bug 936.  The new errors are:
> 
> |       [CC]   src/intl/gettext/loadmsgcat.o
> | loadmsgcat.c: In function '_nl_load_domain':
> | loadmsgcat.c:286: warning: implicit declaration of function 'mmap'
> | loadmsgcat.c:286: error: 'PROT_READ' undeclared (first use in this function)
> | loadmsgcat.c:286: error: (Each undeclared identifier is reported only once
> | loadmsgcat.c:286: error: for each function it appears in.)
> | loadmsgcat.c:287: error: 'MAP_PRIVATE' undeclared (first use in this function)
> | loadmsgcat.c:328: warning: implicit declaration of function 'munmap'
> | make[3]: *** [loadmsgcat.o] Error 1
> 
> There seem to be two bugs here.
> 
> (1) configure is failing to detect munmap():
> 
> | configure:20380: checking for munmap
> | configure:20437: gcc -o conftest -g -O2 -Wall  -I/usr/include  -rdynamic -L/usr/lib conftest.c -lX11 -lssl -lcrypto -ldl  -lz -l
> | /usr/bin/ld: cannot find -lX11
> | collect2: ld returned 1 exit status
> | configure:20443: $? = 1
> 
> An earlier part of configure has got the directories of X from
> imake and added -lX11 to $LIBS, assuming that it would exist
> because the directories are there:
> 
> | configure:19422: result: libraries /usr/lib, headers /usr/include
> 
> We should probably make the configure script explicitly check for
> -lX11 rather than rely on the results of AC_PATH_X.  You may be
> able to work around this bug with "configure --without-x".
> 
> (2) loadmsgcat.c doesn't recover from a missing HAVE_MUNMAP.
> It seems that, if HAVE_MUNMAP is not defined (because the munmap
> function does not exist), then src/intl/gettext/loadmsgcat.c
> should also undefine HAVE_MMAP, and skip all uses of mmap().
> However, the error messages you got point to a section of code
> that is inside #ifdef HAVE_MMAP, and the fact that PROT_READ
> was undefined indicates that <sys/mman.h> was not included.
> Therefore, it seems that loadmsgcat.c first undefined HAVE_MMAP
> because HAVE_MUNMAP was missing, but then something else defined
> HAVE_MMAP again, before loadmsgcat.c checked it.
> 
> Specifically, src/intl/gettext/loadmsgcat.c includes src/elinks.h,
> which includes src/osdep/types.h, which includes config.h, which
> can define HAVE_MMAP again.
> 
> If you work around bug (1) with "configure --without-x", I think
> you can ignore bug (2).  If that doesn't help for some reason,
> try adding "#undef HAVE_MMAP" below the #include directives in
> loadmsgcat.c.

I tried again from scratch and ran configure --without-x. I got errors
again. See http://www.savefile.com/files/856531 for details.


> I'll add these bugs to bugzilla.elinks.cz later this week.

OK. :)
-- 
"Trivial hurts, tiny human accidents," said Firenze, as his hooves 
thudded over the mossy floor. "These are of no more significance than 
the scurryings of ants to the wide universe, and are unaffected by 
planetary movements." --Harry Potter book
  /\___/\
 / /\ /\ \     Phillip (Ant) @ http://antfarm.ma.cx (Personal Web Site)
| |o   o| |         Ant's Quality Foraged Links (AQFL): http://aqfl.net
   \ _ /                 E-mail: philpi at earthlink.net or ant at zimage.com
    ( )



More information about the elinks-users mailing list