dvdauthor 0.6.11_0 doesn't build; missing iconv.h inclusion

[ryandesign@…]

Rytis Sileika reported on macports-users that dvdauthor does not build:

$ sudo port install dvdauthor
Password:
--->  Fetching dvdauthor
--->  Attempting to fetch dvdauthor-0.6.11.tar.gz from http://downloads.sourceforge.net/dvdauthor
--->  Verifying checksum(s) for dvdauthor
--->  Extracting dvdauthor
--->  Configuring dvdauthor
--->  Building dvdauthor with target all
Error: Target org.macports.build returned: shell command " cd "/opt/local/var/macports/build/_Users_rschmidt_macports_dports_multimedia_dvdauthor/work/dvdauthor-0.6.11" && make all " returned error 2
Command output: if /usr/bin/gcc-4.0 -DHAVE_CONFIG_H -I. -I. -I.  -DSYSCONFDIR="\"/opt/local/etc\"" -I/opt/local/include/libxml2 -I/opt/local/include -I/opt/local/include -I/opt/local/include/freetype2 -I/opt/local/include  -I/opt/local/include -Wall -O2 -MT dvdpgc.o -MD -MP -MF ".deps/dvdpgc.Tpo" -c -o dvdpgc.o dvdpgc.c; \
then mv -f ".deps/dvdpgc.Tpo" ".deps/dvdpgc.Po"; else rm -f ".deps/dvdpgc.Tpo"; exit 1; fi
if /usr/bin/gcc-4.0 -DHAVE_CONFIG_H -I. -I. -I.  -DSYSCONFDIR="\"/opt/local/etc\"" -I/opt/local/include/libxml2 -I/opt/local/include -I/opt/local/include -I/opt/local/include/freetype2 -I/opt/local/include  -I/opt/local/include -Wall -O2 -MT dvdcli.o -MD -MP -MF ".deps/dvdcli.Tpo" -c -o dvdcli.o dvdcli.c; \
then mv -f ".deps/dvdcli.Tpo" ".deps/dvdcli.Po"; else rm -f ".deps/dvdcli.Tpo"; exit 1; fi
if /usr/bin/gcc-4.0 -DHAVE_CONFIG_H -I. -I. -I.  -DSYSCONFDIR="\"/opt/local/etc\"" -I/opt/local/include/libxml2 -I/opt/local/include -I/opt/local/include -I/opt/local/include/freetype2 -I/opt/local/include  -I/opt/local/include -Wall -O2 -MT readxml.o -MD -MP -MF ".deps/readxml.Tpo" -c -o readxml.o readxml.c; \
then mv -f ".deps/readxml.Tpo" ".deps/readxml.Po"; else rm -f ".deps/readxml.Tpo"; exit 1; fi
readxml.c: In function 'readxml':
readxml.c:78: warning: pointer targets in passing argument 1 of 'strcmp' differ in signedness
readxml.c:89: warning: pointer targets in initialization differ in signedness
readxml.c:89: warning: pointer targets in initialization differ in signedness
readxml.c:144: warning: pointer targets in initialization differ in signedness
readxml.c: At top level:
readxml.c:193: error: parse error before 'get_conv'
readxml.c:194: warning: return type defaults to 'int'
readxml.c: In function 'get_conv':
readxml.c:195: error: nested functions are disabled, use -fnested-functions to re-enable
readxml.c:195: error: syntax error before 'ic'
readxml.c:197: error: 'ic' undeclared (first use in this function)
readxml.c:197: error: (Each undeclared identifier is reported only once
readxml.c:197: error: for each function it appears in.)
readxml.c:197: error: 'iconv_t' undeclared (first use in this function)
readxml.c:212: warning: implicit declaration of function 'iconv_open'
readxml.c: In function 'utf8tolocal':
readxml.c:223: error: 'iconv_t' undeclared (first use in this function)
readxml.c:223: error: parse error before 'c'
readxml.c:230: warning: implicit declaration of function 'iconv'
readxml.c:230: error: 'c' undeclared (first use in this function)
make[2]: *** [readxml.o] Error 1
make[1]: *** [all] Error 2
make: *** [all-recursive] Error 1

Error: Status 1 encountered during processing.
$

[ryandesign@…]

I thought maybe updating the port to the latest version, 0.6.14, might help. But it didn't. So I'll file that as a separate ticket: #13477.

[ryandesign@…]

Rytis found that the problem can be fixed by adding #include "iconv.h" to readxml.c. I can confirm that this allows dvdauthor to build. I made a patch and am attaching it. meissnem, please let me know if I may commit this change or if you have any other comments.

Since this problem appears not to be fixed in the latest version of the software, this fix should also be reported to the developers of the software for inclusion in the next version.

[ryandesign@…]

Maybe #include <iconv.h> would be better. It seems to work with both, but my recollection is that #include "..." is for headers and files provided with a piece of software, while #include <...> is for headers and files provided by other software already on the system.

[meissnem@…]

I prefer #include <...>

Ryan, please go ahead and commit the change -- I have no objections at all.

[ryandesign@…]

Great, committed in r31701.

Can you take this issue up with the authors of the software, or see if they already have a bug filed on it?