GNOME.org
 

Re: [xml] misc warnings compiling libxml2-2.6.22 (cvs snapshot dloaded 2005/12/09) with HP aC++/ANSI C B3910B A.06.05 [Jul 25 2005]

On Fri, Dec 09, 2005 at 04:57:06PM -0800, Rick Jones wrote:

Out of curiousity I downloaded the cvs snapshot on 2005-12-09 which 
expanded to 2.6.22, and tried compiling it on an HP-UX 11.23 Itanium (ia64) 
system with HP aC++/ANSI C B3910B A.06.05 [Jul 25 2005].

A few pseudo-random observations/questions...

This one may have been overly picky on the part of the compiler:

cc -DHAVE_CONFIG_H -I. -I. -I. -I./include -I./include -D_REENTRANT -g 
-Wp,-H30000 -c parserInternals.c  +Z -DPIC -o .libs/parserInternals.lo
"parserInternals.c", line 1883: warning #2068-D: integer conversion 
resulted in a change of sign
          return (-1);


  right, added a cast, as I prefer to not change this API ... this is 
unfortunate.


cc -DHAVE_CONFIG_H -I. -I. -I. -I./include -I./include -D_REENTRANT -g 
-Wp,-H30000 -c HTMLparser.c +Z -DPIC -o .libs/HTMLparser.lo
"HTMLparser.c", line 4782: warning #2550-D: variable "oldname" was set but 
never used
                const xmlChar *name, *oldname;


  Yup, cosmetic, but cleaned up :-)



Some stuff that probably wants a cast and some unreachable code:

cc -DHAVE_CONFIG_H -I. -I. -I. -I./include -I./include -D_REENTRANT -g 
-Wp,-H30000 -c xpath.c  +Z -
DPIC -o .libs/xpath.lo
"xpath.c", line 2915: warning #2513-D: a value of type "void *" cannot be 
assigned to an entity of
type "xmlXPathFunction"
      XML_CAST_FPTR(ret) = xmlHashLookup2(ctxt->funcHash, name, ns_uri);
                         ^

"xpath.c", line 10483: warning #2513-D: a value of type "void *" cannot be 
assigned to an entity of
 type "xmlXPathFunction"
                      XML_CAST_FPTR(func) = op->cache;


  I still don't have a good solutions for those. The previous version of
XML_CAST_FPTR() which was attempting a convoluted cast broke register aliasing
on gcc 4 on s390, and all platforms supported so far have no actual problems
storing function pointers in data pointers, gcc also shows up a warning,
I just don't have a nice solution yet at this point.


"xpath.c", line 11095: warning #2128-D: loop is not reachable from 
preceding code
      do {
      ^



  hum ... strange, there is a goto next_node below which allows to get
back in that do loop. So it is reachable from following code, not a very
nice piece of code, but I would rather be prudent there.


next:

cc -DHAVE_CONFIG_H -I. -I. -I. -I./include -I./include -D_REENTRANT -g 
-Wp,-H30000 -c xmlschemas.c +Z -DPIC -o .libs/xmlschemas.lo
"xmlschemas.c", line 2500: warning #2550-D: variable "is_not" was set but 
never used
      int i, is_not;
             ^


   okay


"xmlschemas.c", line 7992: warning #2550-D: variable "ns" was set but never 
used
            xmlNsPtr ns;
                     ^


  okay


"xmlschemas.c", line 13235: warning #2111-D: statement is unreachable
      return (0);
      ^


  okay


"xmlschemas.c", line 13279: warning #2111-D: statement is unreachable
      return (NULL);
      ^


 okay


"xmlschemas.c", line 17781: warning #2111-D: statement is unreachable
      return (0);
      ^


  okay


This one caught my eye:

cc -DHAVE_CONFIG_H -I. -I. -I. -I./include -I./include -D_REENTRANT -g 
-Wp,-H30000 -c xmlschemastypes.c  +Z -DPIC -o .libs/xmlschemastypes.lo
"xmlschemastypes.c", line 1699: warning #2186-D: pointless comparison of 
unsigned integer with zero
      if ((ret != 0) || (*cur != 0) || !VALID_DATETIME((&(dt->value.date))))
                                        ^

"xmlschemastypes.c", line 1699: warning #2186-D: pointless comparison of 
unsigned integer with zero
      if ((ret != 0) || (*cur != 0) || !VALID_DATETIME((&(dt->value.date))))
                                        ^


  The macros expands to quite a lot of code. Maybe running on the preprocessor
output would help pinpoint what exactly is at stake there. I don't really 
understand the message too ... unless something like u < 0 I don't see the
point being raised.


This one brings-up another point:

cc -DHAVE_CONFIG_H -I. -I. -I. -I./include -I./include -D_REENTRANT -g 
-Wp,-H30000 -c xmlmodule.c+Z -DPIC -o .libs/xmlmodule.lo
"xmlmodule.c", line 284: warning #2167-D: argument of type "void **" is 
incompatible with parameter
 of type "shl_t *"
      rc = shl_findsym(&handle, name, TYPE_UNDEFINED, symbol);
                       ^


  data vs. function pointers conversions.


With PA2.0W (64-bit PA) and with IPF/ia64, (either 32 bit or 64) there is 
support for dlopen et al.  In fact, the shl_findsym manpage has this to say:

      Future HP-UX environments may not support these routines and flags or
      may only support a subset of them.  Instead, they will use the SVR4
      dynamic loading API.  Users are encouraged to migrate to the dl*
      family of dynamic linking routines.  See the dlclose(3C), dlerror(3C),
      dlget(3C), dlgetname(3C), dlmodinfo(3C), dlopen(3C), and dlsym(3C) man
      pages for more information.


for some reason while I'm typing the above I am getting a feeling of deja 
vu. anyway...


   Well I think we used that code explicitely because dlopen() wasn't available
on old HP-UX systems...


cc -DHAVE_CONFIG_H -I. -I. -I. -I./include -I./include -D_REENTRANT -g 
-Wp,-H30000 -c schematron.c +Z -DPIC -o .libs/schematron.lo
"schematron.c", line 1000: warning #2177-D: function 
"xmlSchematronLoadInclude" was declared but ne
ver referenced
  xmlSchematronLoadInclude(xmlSchematronParserCtxtPtr ctxt, xmlNodePtr cur)
  ^


  yes unfinished code.


next:

        source='testapi.c' object='testapi.o' libtool=no \
        depfile='.deps/testapi.Po' tmpdepfile='.deps/testapi.TPo' \
        depmode=none /bin/sh ./depcomp \
        cc -DHAVE_CONFIG_H -I. -I. -I. -I./include -I./include  
        -D_REENTRANT -g -Wp,-H30000 -c `test -f 'testapi.c' || echo 
    './'`testapi.c
"testapi.c", line 286: warning #2068-D: integer conversion resulted in a 
change of sign
      if (no == 2) return('ø');
                         ^

"testapi.c", line 299: warning #2068-D: integer conversion resulted in a 
change of sign
      return(-1);
            ^

"testapi.c", line 311: warning #2068-D: integer conversion resulted in a 
change of sign
      return(-1);


   okay, but anything below this is autogenerated, not something I think I
will fix as it requires fixing the generation script.


and some additional picky stuff:

creating testModule
        source='runtest.c' object='runtest.o' libtool=no \
        depfile='.deps/runtest.Po' tmpdepfile='.deps/runtest.TPo' \
        depmode=none /bin/sh ./depcomp \
        cc -DHAVE_CONFIG_H -I. -I. -I. -I./include -I./include  
        -D_REENTRANT -g -Wp,-H30000 -c `test -f 'runtest.c' || echo 
    './'`runtest.c
"runtest.c", line 3142: warning #2550-D: variable "ret" was set but never 
used
      int res = 0, len, ret;
                        ^

"runtest.c", line 212: warning #2550-D: variable 
"libxmlMemoryAllocatedBase" was set but never used
  static long libxmlMemoryAllocatedBase = 0;
              ^


  okay, I assume it was used to look in a debugging session...


"runtest.c", line 3859: warning #2177-D: function "thread_specific_data" 
was declared but never referenced
  thread_specific_data(void *private_data)
  ^


  not used on HP-UX but used for pthreads systems... let's ignore this


moving on:

creating runtest
        source='runsuite.c' object='runsuite.o' libtool=no \
        depfile='.deps/runsuite.Po' tmpdepfile='.deps/runsuite.TPo' \
        depmode=none /bin/sh ./depcomp \
        cc -DHAVE_CONFIG_H -I. -I. -I. -I./include -I./include  
        -D_REENTRANT -g -Wp,-H30000 -c `test -f 'runsuite.c' || echo 
    './'`runsuite.c
"runsuite.c", line 1054: warning #2550-D: variable "res" was set but never 
used
      int res, ret = 0;
          ^


  okay


"runsuite.c", line 96: warning #2550-D: variable 
"libxmlMemoryAllocatedBase" was set but never used
  static long libxmlMemoryAllocatedBase = 0;


  same thing


then:

creating parse3
        source='tree2.c' object='tree2.o' libtool=no \
        depfile='.deps/tree2.Po' tmpdepfile='.deps/tree2.TPo' \
        depmode=none /bin/sh ../../depcomp \
        cc -DHAVE_CONFIG_H -I. -I. -I../.. -I../../include -I../../include 
-I./include  -D_REENTRANT      -g -Wp,-H30000 -c `test -f 'tree2.c' || echo 
'./'`tree2.c
"tree2.c", line 30: warning #2550-D: variable "dtd" was set but never used
      xmlDtdPtr dtd = NULL;       /* DTD pointer */
                ^


  it's for educational purpose, I will ignore this as this may be useful in
code reuse.


and then finally, an outright build error:

creating parse2
No suffix list.
sh: /usr/bin/xsltproc:  not found.
*** Error exit code 127 (ignored)
sh: /usr/bin/xsltproc:  not found.
*** Error exit code 127 (ignored)
sh: /usr/bin/xsltproc:  not found.
*** Error exit code 127 (ignored)
sh: /usr/bin/xsltproc:  not found.
*** Error exit code 127 (ignored)
Making all in example

there is no xsltproc on the system:


  configure should have found this ... strange ...



though I must admit my ability to quickly parse shell stuff is pretty weak.


  and my ability to read generated configure code, is not much better...
I have fixed all okay'ed, and will commit as soon as I verify all regression
tests still pass as expected.

  thanks,

Daniel

-- 
Daniel Veillard      | Red Hat http://redhat.com/
veillard redhat com  | libxml GNOME XML XSLT toolkit  http://xmlsoft.org/
http://veillard.com/ | Rpmfind RPM search engine http://rpmfind.net/
[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]