Source: ncurses
Version: 5.9+20150516-1
Tags: moreinfo
User: [email protected]
Usertags: rebootstrap

Hi,

I've seen ncurses FTBFS three times now and still no clue why. The
symptom is that the build hangs at:

| test -d /tmp/buildd/ncurses-5.9+20150516/obj-test || mkdir 
/tmp/buildd/ncurses-5.9+20150516/obj-test
| cd /tmp/buildd/ncurses-5.9+20150516/obj-test && CFLAGS="-g -O2 
-fstack-protector-strong -Wformat -Werror=format-security" 
/tmp/buildd/ncurses-5.9+20150516/test/configure \
|       --prefix=/usr --build=x86_64-linux-gnu 
--bindir=/usr/lib/ncurses/examples --with-ncursesw 
--with-curses-dir=/tmp/buildd/ncurses-5.9+20150516/obj-wide
| checking build system type... x86_64-pc-linux-gnu
| checking host system type... x86_64-pc-linux-gnu
| Configuring for linux-gnu
| checking whether make sets ${MAKE}... yes
| checking for gcc... gcc
| checking for C compiler default output... a.out
| checking whether the C compiler works... yes
| checking whether we are cross compiling... no
| checking for executable suffix... 
| checking for object suffix... o
| checking whether we are using the GNU C compiler... yes
| checking whether gcc accepts -g... yes
| checking version of gcc... 4.9.2
| checking for gcc option to accept ANSI C... none needed
| checking $CC variable... ok
| checking how to run the C preprocessor... gcc -E
| checking for mawk... mawk
| checking for a BSD compatible install... /usr/bin/install -c
| checking for tdlint... no
| checking for lint... no
| checking for alint... no
| checking for splint... no
| checking for lclint... no
| checking if filesystem supports mixed-case filenames... yes
| checking for exctags... no
| checking for ctags... no
| checking for exetags... no
| checking for etags... no
| checking for ctags... no
| checking for etags... no
| checking if -lm needed for math functions... yes
| configure: checking for gcc __attribute__ directives...
| ... scanf
| ... printf
| ... unused
| ... noreturn
| checking if we must define _GNU_SOURCE... yes
| checking if _XOPEN_SOURCE really is set... yes
| checking for signal global datatype... volatile sig_atomic_t
| checking if you want to use pkg-config... yes
| checking for pkg-config... /usr/bin/pkg-config
| checking if you want to see long compiling messages... yes
| checking for ncurses wrap-prefix... _nc_
| checking if you want to check for wide-character functions... yes
| checking for specific curses-directory... 
/tmp/buildd/ncurses-5.9+20150516/obj-wide
| checking for specified curses library type... ncursesw
| checking for multibyte character support... yes
| checking pkg-config for ncursesw... no
| Looking for ncursesw-config
| checking for ncursesw-config... no
| checking for ncursesw6-config... no
| checking for ncursesw5-config... no

At this point, attaching the main configure shell reveals that it
continuously calls stat:

| stat("/tmp/buildd/ncurses-5.9+20150516/obj-wide/include/ncursesw", 
0x7ffd1feb0220) = -1 ENOENT (No such file or directory)

The generated configure scrip contains a snippet that roughly resembles:

| while test $cf_add_incdir != /usr/include; do
|   if test -d $cf_add_incdir; then
|        ...
|   # no else branch
|   fi
| done

I've seen this in a native amd64 reproducible build on
jenkins.debian.net:

https://jenkins.debian.net/job/reproducible_builder_alpha/46528/console
(the above excerpt is from this log)

I've also seen this in a cross build for x32 and a cross build for i386.

https://jenkins.debian.net/job/rebootstrap_i386_gcc49_nobiarch/139/console

In the x32 build (no log available), I created the directory being
stat()ed. At that point, the build continued and finished successfully.
Both cross builds set DEB_BUILD_OPTIONS="nocheck parallel=1".

Notably, the reproducible jenkins job was able to build this version of
ncurses earlier, so it seems that something else is the cause. The
buildds had no problems building it either.

This is all I know.

If the bug does not receive sufficient detail to be diagnosed, it should
be closed within one month of inactivity.

Helmut


-- 
To UNSUBSCRIBE, email to [email protected]
with a subject of "unsubscribe". Trouble? Contact [email protected]

Reply via email to