Re: Any hint of 0.9.7 final ?

[I've cc'ed the list as other users may face the same problems or even
know solutions. Please keep this discussion on the list]

Hi,

Stéphane Téletchéa writes:
 > + libtoolize --copy --force
 > Remember to add `AC_PROG_LIBTOOL' to `configure.in'.
 > You should update your `aclocal.m4' by running aclocal.

Is that something that the RPM build system forces you to do? RefDB
does not use libtool. RefDB's own autogen.sh script does not call
libtoolize. If you must do this anyway, it is safe to ignore the error
message.

 > checking for dtdparse... no
 > checking for dtdformat... no
 > checking for sgrep... no
 > checking for tidy... /usr/bin/tidy
 > checking for batik-rasterizer... no

These missing programs contribute to the build failure. More below.

 > checking for shared library run path origin... /bin/sh: ./conf/config.rpath: No such file or directory
 > done

I see this warning on several problems without adverse effect. It
seems to be safe to ignore this.

 > checking for http://www.tei-c.org/Stylesheets/p4/fo/tei.xsl... no
 > checking for XML file http://www.tei-c.org/Stylesheets/p4/fo/tei.xsl... checking for No entry for SYSTEM http://www.tei-c.org/Stylesheets/p4/fo/tei.xsl No entry for URI http://www.tei-c.org/Stylesheets/p4/fo/tei.xsl... no
 > XML file not found

This is strange if you know that the file is installed. Most likely
the TEI stylesheets are not correctly listed in the XML catalog, or
there is only an entry in one of the SGML catalogs. If this is true,
you should file a bug report for the TEI stylesheet package. As a
short-term fix, you could try to pass the location of the stylesheets
to the configure file by means of the --with-tei-xsl option.

 > refdbda.c: In function 'daemon_started_by_inetd':
 > refdbda.c:2494: warning: pointer targets in passing argument 5 of 'getsockopt' differ in signedness

I think I've fixed this by adding a proper test to configure.in

 > refdbdbib.c: In function 'getbib':
 > refdbdbib.c:905: warning: passing argument 2 of 'iconv' from incompatible pointer type
 > refdbdbib.c:1004: warning: passing argument 2 of 'iconv' from incompatible pointer type

There seems to be no portable solution to this problem. Some systems
use char**, others use const char**, and no matter what you use in
your code, it'll generate a warning in half of the systems.

 > refdbc.c: In function 'addref':
 > refdbc.c:2890: warning: call to __builtin___strncpy_chk will always overflow destination buffer
 > refdbc.c: In function 'com_getref':
 > refdbc.c:1820: warning: call to __builtin___strncpy_chk will always overflow destination buffer
 > refdbc.c: In function 'addnote':
 > refdbc.c:3587: warning: call to __builtin___strncpy_chk will always overflow destination buffer
 > refdbc.c: In function 'com_getnote':
 > refdbc.c:4538: warning: call to __builtin___strncpy_chk will always overflow destination buffer
 > refdbc.c: In function 'addlink':
 > refdbc.c:5076: warning: call to __builtin___strncpy_chk will always overflow destination buffer
 > refdbc.c: In function 'com_updatejo':
 > refdbc.c:5399: warning: call to __builtin___strncpy_chk will always overflow destination buffer

I've fixed the incorrect string sizes.

 > nmed2ris.c: In function 'main':
 > nmed2ris.c:182: warning: incompatible implicit declaration of built-in function 'exit'

nmed2ris has essentially been replaced by the med2ris Perl
script years ago. I've removed it instead of attempting to fix it.

 > bib2ris.c: In function 'log_print':
 > bib2ris.c:1887: warning: incompatible implicit declaration of built-in function 'strftime'
 > bib2ris.c:1887: warning: passing argument 4 of 'strftime' makes pointer from integer without a cast

I can't figure out what's wrong with this one. All you need to declare
strftime properly is to include <time.h> which I did. It doesn't help
here. Other files use the same strftime call, include <time.h>, and do
not generate this warning (e.g. refdbdref.c). I'm stymied.

 > ../scripts/statgen.pl docbook > refdb-manual-statustable.xml
 > make[1]: *** No rule to make target `include/doc-html.dsl', needed by `refdb-manual/ch01.html'.  Stop.

I've changed the doc build system from SGML to XML after 0.9.6, so I'm
not surprised to see a few glitches. This one seems to be caused by an
incorrect dependency (dsl files are used no more). I've fixed this.

In any case, make should not attempt to build the docs as they are
shipped pre-built in the tar.gz archives. You should only need to
rebuild them after running maintainer-clean. If you do that, or if the
RPM build process mandates that, you'll need the additional tools that
I mentioned above when looking at your configure output.


I've uploaded a new prerelease (0.9.7-pre6) based on these
changes. Please give it a try and let me know.

http://refdb.sourceforge.net/pre/refdb-latest.tar.gz

regards
Markus
-- 
Markus Hoenicka
[email protected]
(Spam-protected email: replace the quadrupeds with "mhoenicka")
http://www.mhoenicka.de



-------------------------------------------------------
Using Tomcat but need to do more? Need to support web services, security?
Get stuff done quickly with pre-integrated technology to make your job easier
Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
http://sel.as-us.falkag.net/sel?cmd=lnk&kid0709&bid&3057&dat1642