can't build release because of undefined constants; can't build git because of missing deps

[themaddoctor]

For the release version:

gcc -DHAVE_CONFIG_H -I. -I../common  -I../include -I../common -D_FILE_OFFSET_BITS=64 -I/usr/include/fuse    -g -O2 -MT vmdkmount.o -MD -MP -MF .deps/vmdkmount.Tpo -c -o vmdkmount.o vmdkmount.c
vmdkmount.c: In function ‘vmdkmount_fuse_open’:
vmdkmount.c:183:13: error: ‘EINVAL’ undeclared (first use in this function)
   result = -EINVAL;
             ^~~~~~
vmdkmount.c:183:13: note: each undeclared identifier is reported only once for each function it appears in
vmdkmount.c:216:13: error: ‘ENOENT’ undeclared (first use in this function)
   result = -ENOENT;
             ^~~~~~
vmdkmount.c:229:13: error: ‘EACCES’ undeclared (first use in this function); did you mean ‘AT_EACCESS’?
   result = -EACCES;
             ^~~~~~
             AT_EACCESS
vmdkmount.c: In function ‘vmdkmount_fuse_read’:
vmdkmount.c:273:13: error: ‘EINVAL’ undeclared (first use in this function)
   result = -EINVAL;
             ^~~~~~
vmdkmount.c:320:13: error: ‘ENOENT’ undeclared (first use in this function)
   result = -ENOENT;
             ^~~~~~
vmdkmount.c:354:13: error: ‘EIO’ undeclared (first use in this function)
   result = -EIO;
             ^~~
vmdkmount.c: In function ‘vmdkmount_fuse_readdir’:
vmdkmount.c:600:13: error: ‘EINVAL’ undeclared (first use in this function)
   result = -EINVAL;
             ^~~~~~
vmdkmount.c:617:13: error: ‘ENOENT’ undeclared (first use in this function)
   result = -ENOENT;
             ^~~~~~
vmdkmount.c:633:13: error: ‘errno’ undeclared (first use in this function); did you mean ‘error’?
   result = -errno;
             ^~~~~
             error
vmdkmount.c:649:13: error: ‘EIO’ undeclared (first use in this function)
   result = -EIO;
             ^~~
vmdkmount.c: In function ‘vmdkmount_fuse_getattr’:
vmdkmount.c:801:30: error: ‘ENOENT’ undeclared (first use in this function)
  int result               = -ENOENT;
                              ^~~~~~
vmdkmount.c:814:13: error: ‘EINVAL’ undeclared (first use in this function)
   result = -EINVAL;
             ^~~~~~
vmdkmount.c:843:12: error: ‘errno’ undeclared (first use in this function); did you mean ‘error’?
   result = errno;
            ^~~~~
            error
vmdkmount.c:896:15: error: ‘EIO’ undeclared (first use in this function)
     result = -EIO;
               ^~~
make[1]: *** [Makefile:712: vmdkmount.o] Error 1
make[1]: Leaving directory '/building/forensics/libvmdk-20170226/vmdktools'
make: *** [Makefile:749: all-recursive] Error 1

For the GIT version, all the dependencies are on my system, but automake and configure both insist on dying unless they are also inside the source tree. (This is common for the libraries on libyal, not just for this one.)

[themaddoctor]

I'm sure the error in the release version is just a missing header, but I'm too tired right now to hunt it down.

[themaddoctor]

Found it: configure is not finding errno.h, even though it is on the system.

[themaddoctor]

But now tests fail:

gcc -DHAVE_CONFIG_H -I. -I../common  -I../include -I../common    -g -O2 -MT vmdk_test_getopt.o -MD -MP -MF .deps/vmdk_test_getopt.Tpo -c -o vmdk_test_getopt.o vmdk_test_getopt.c
vmdk_test_getopt.c:33:10: fatal error: vmdk_test_libcnotify.h: No such file or directory
 #include "vmdk_test_libcnotify.h"
          ^~~~~~~~~~~~~~~~~~~~~~~~
compilation terminated.
make[2]: *** [Makefile:1105: vmdk_test_getopt.o] Error 1
make[2]: Leaving directory '/building/forensics/libvmdk-20170226/tests'
make[1]: *** [Makefile:1391: check-am] Error 2
make[1]: Leaving directory '/building/forensics/libvmdk-20170226/tests'
make: *** [Makefile:749: check-recursive] Error 1

[themaddoctor]

After fixing that, get

# TOTAL: 2
# PASS:  1
# SKIP:  1
# XFAIL: 0
# FAIL:  0
# XPASS: 0
# ERROR: 0

.. contents:: :depth: 2

SKIP: test_vmdkinfo.sh
======================

Test input directory: input not found.
SKIP test_vmdkinfo.sh (exit status: 77)

[joachimmetz]

thx for the report I'll have a look as soon as time permits

[joachimmetz]

Could you attach config.log to determine why errno.h was not detected.

[themaddoctor]

config.log

There is actually no mention of errno.h in the config.log file, nor in config.status

[joachimmetz]

check for errno.h is here: https://github.com/libyal/libvmdk/blob/b9013319ec8376bd81b7bd6348296de9edc6418c/acinclude.m4#L47

which is invoked from here: https://github.com/libyal/libvmdk/blob/b9013319ec8376bd81b7bd6348296de9edc6418c/configure.ac#L130

but no mention of errno in config.log

Where config.log from other system:

configure:42155: checking for errno.h
configure:42155: result: yes
configure:42166: checking whether time.h and sys/time.h may both be included
configure:42193: result: yes
configure:42206: checking for getegid

[joachimmetz]

Does not appears to be an issue with shared versions of dependencies:

Building:
   libcerror support:                         yes
   libcthreads support:                       yes
   libcdata support:                          yes
   libclocale support:                        yes
   libcnotify support:                        yes
   libcsplit support:                         yes
   libuna support:                            yes
   libcfile support:                          yes
   libcpath support:                          yes
   libbfio support:                           yes
   libfcache support:                         yes
   libfdata support:                          yes
   libfvalue support:                         yes
   DEFLATE compression support:               zlib
   FUSE support:                              libfuse

Features:
   Multi-threading support:                   libcthreads
   Wide character type support:               no
   vmdktools are build as static executables: no
   Python (pyvmdk) support:                   no
   Python version 2 (pyvmdk) support:         no
   Python version 3 (pyvmdk) support:         no
   Verbose output:                            no
   Debug output:                              no

[joachimmetz]

@themaddoctor could you indicate if this issue persist with the recent version of libvmdk

I'll release a 20181118 version later today

[joachimmetz]

Closing issue, please reopen if this is still an issue

[themaddoctor]

You closed that pretty fast without waiting for me to try again. Even faster when you consider that the bug report is over 9 months old. But that's cool. I'll figure it out when I update.

[joachimmetz]

It looks like at the date it was reported the issue had already been addressed in HEAD.

[joachimmetz]

You closed that pretty fast without waiting for me to try again.

Not sure if I follow you, on the 18th I indicated that I would release a new version and that the issue can be reopened if the issue persists.