notmuch/lib/gen-version-script.sh


# we go through a bit of work to get the unmangled names of the
# typeinfo symbols because of
# http://sourceware.org/bugzilla/show_bug.cgi?id=10326

if [ $# -lt 2 ]; then
    echo Usage: $0 header obj1 obj2 obj3
    exit 1;
fi

HEADER=$1
shift

printf '{\nglobal:\n'
nm  $* | awk '$1 ~ "^[0-9a-fA-F][0-9a-fA-F]*$" && $3 ~ "Xapian.*Error" {print $3}' | sort | uniq | \
while read sym; do
    demangled=$(c++filt $sym)
    case $demangled in
	typeinfo*) 
	    printf "\t$sym;\n"
	    ;;
	*)
	    ;;
    esac
done
nm $* | awk '$1 ~ "^[0-9a-fA-F][0-9a-fA-F]*$" && $2 == "T" && $3 ~ "^(getline|getdelim|canonicalize_file_name)$" {print $3 ";"}'
sed  -n 's/^[[:space:]]*\(notmuch_[a-z_]*\)[[:space:]]*(.*/ \1;/p' $HEADER
printf "local: *;\n};\n"
libnotmuch: export Xapian typeinfo symbols The lack of such exporting seems to cause problems catching exceptions, as suggested by http://gcc.gnu.org/wiki/Visibility This manifested in the symbol-hiding test failing when notmuch was compile with gcc 4.4.5. On i386, this further manifested as notmuch new failing to run (crashing with an uncaught exception on first run). 2011-07-16 20:59:44 +02:00
			`# we go through a bit of work to get the unmangled names of the`
			`# typeinfo symbols because of`
			`# http://sourceware.org/bugzilla/show_bug.cgi?id=10326`

			`if [ $# -lt 2 ]; then`
			`echo Usage: $0 header obj1 obj2 obj3`
			`exit 1;`
			`fi`

			`HEADER=$1`
			`shift`

			`printf '{\nglobal:\n'`
lib/gen-version-script.sh: replace --defined argument to nm with awk OpenBSD nm apparently doesn't support --defined. The awk condition is based on the assumption that all defined symbols have some hex number in the first column. Thanks to Matthias Guedemann reporting the problem, and an earlier version of this patch. 2011-09-05 04:50:36 +02:00			`nm $* \| awk '$1 ~ "^[0-9a-fA-F][0-9a-fA-F]$" && $3 ~ "Xapian.Error" {print $3}' \| sort \| uniq \| \`
libnotmuch: export Xapian typeinfo symbols The lack of such exporting seems to cause problems catching exceptions, as suggested by http://gcc.gnu.org/wiki/Visibility This manifested in the symbol-hiding test failing when notmuch was compile with gcc 4.4.5. On i386, this further manifested as notmuch new failing to run (crashing with an uncaught exception on first run). 2011-07-16 20:59:44 +02:00			`while read sym; do`
			`demangled=$(c++filt $sym)`
			`case $demangled in`
			`typeinfo*)`
			`printf "\t$sym;\n"`
			`;;`
			`*)`
			`;;`
			`esac`
			`done`
build: add canonicalize_file_name to symbols exported from libnotmuch.so This is needed for our compat version of canonicalize_file_name to be used. 2014-04-09 13:24:02 +02:00			`nm $* \| awk '$1 ~ "^[0-9a-fA-F][0-9a-fA-F]*$" && $2 == "T" && $3 ~ "^(getline\|getdelim\|canonicalize_file_name)$" {print $3 ";"}'`
re-enable notmuch.sym generation using POSIX sed Unfortunately Robin Green's patch 52e4dedf9aa was lost when I created gen-version-script.sh. This merges his changes manually into that script. It turns out tabs seem not needed in version script files, so I simplified a bit and removed the printf. Thanks to Alexander Botero-Lowry for help and testing. 2011-08-21 21:45:00 +02:00			`sed -n 's/^[[:space:]]\(notmuch_[a-z_]\)[[:space:]](./ \1;/p' $HEADER`
libnotmuch: export Xapian typeinfo symbols The lack of such exporting seems to cause problems catching exceptions, as suggested by http://gcc.gnu.org/wiki/Visibility This manifested in the symbol-hiding test failing when notmuch was compile with gcc 4.4.5. On i386, this further manifested as notmuch new failing to run (crashing with an uncaught exception on first run). 2011-07-16 20:59:44 +02:00			`printf "local: *;\n};\n"`