acroread won't start: "/usr/bin/acroread must be executed from the startup script."

[Peter Cordes]

peter@tesla:~$ acroread
/usr/bin/acroread must be executed from the startup script.
peter@tesla:~$ echo $?
1

 This is 9.1.0-7jaunty2+medibuntu1 on AMD64 Jaunty (with some packages from jaunty-proposed). I compile my own kernel from source from Ubuntu's jaunty git repo, but with my own config. I don't _think_ I left out anything important, like an IPC mechanism that acroread wants...

I haven't used acroread for a long time, and this is the first time I've tried since upgrading from v8.something. I don't remember ever seeing this prob. before, though.

ltrace output reveals that acroread prints the error message and exits right after a stat of "/opt/Adobe/Reader9/Reader". The system call succeeds, but __xstat returns -1. Weird.

and from
$ ltrace -f -S -o acro.trace /bin/bash /usr/bin/acroread
... startup stuff ...
12596 _ZNSsC1Ev(0x9789ae0, 0xffd1eb18, 0x8b6aa52, 1, 0x970ab5c) = 0x97895e4
12596 __cxa_atexit(0x8078700, 0, 0x970c204, 0x9789ae0, 0xffd1eb18) = 0
12596 getenv("GTK_IM_MODULE") = NULL
12596 XInitThreads(0xf656aff4, 0xffd1eba8, 0xf6422775, 1, 0xffd1ebd4) = 1
12596 pthread_key_create(0x9796960, 0, 0x8c5b114, 0xffd1eb48, 0x8563a4c) = 0
12596 pthread_key_create(0x9796968, 0, 0x8c5b114, 0xffd1eb48, 0x8563a4c) = 0
12596 pthread_key_create(0x9796964, 0, 0x8c5b114, 0xffd1eb48, 0x8563a4c) = 0
12596 malloc(88) = 0x0a8e3780
12596 pthread_mutex_init(0xa8e3780, 0, 0xa8e3780, 0xffd1eb3c, 0x8a099ee) = 0
12596 pthread_cond_init(0xa8e379c, 0, 0xa8e3780, 0, 0xa8e3780) = 0
12596 pthread_getspecific(0, 0x8c5b114, 0x8b6a1a8, 0xffd1eb48, 0x8563a51) = 0
12596 malloc(9160) = 0x0a8e37e0
12596 memset(0x0a8e37e0, '\000', 9160) = 0x0a8e37e0
12596 pthread_setspecific(0, 0xa8e37e0, 0xa8e37e0, 0, 9160) = 0
12596 pthread_self(0xa8e37e0, 0xa8e37e0, 0xffd1eb28, 0x8a09ab4, 0xa8e3780) = 0xf55a0720
12596 pthread_mutex_lock(0xa8e3780, 0xa8e37e0, 0xa8e37e0, 0xffd1eb28, 0x8a09ab4) = 0
12596 pthread_mutex_unlock(0xa8e3780, 0xa8e3780, 0xa8e3780, 0xa8e37e0, 0x40000001) = 0
12596 pthread_mutex_lock(0xa8e3780, 0xa8e37e0, 0xffd1eb28, 0x8a09b21, 0xa8e3780) = 0
12596 pthread_mutex_unlock(0xa8e3780, 0xa8e3780, 0xa8e37e0, 0x40000001, 0xffd1eb3c) = 0
12596 _Znwj(24, 0xa8e3780, 0xa8e37e0, 0x40000001, 0xffd1eb3c) = 0xa8e5bb0
12596 memcpy(0x0a8e5bb4, "", 16) = 0x0a8e5bb4
12596 pthread_getspecific(2, 0xa8e37e0, 0xa8e37e0, 0xa8e37e0, 0xffd1eb3c) = 0
12596 malloc(172) = 0x0a8e5bd0
12596 memset(0x0a8e5bd0, '\000', 172) = 0x0a8e5bd0
12596 pthread_setspecific(2, 0xa8e5bd0, 0xa8e5bd0, 0, 172) = 0
12596 _Znwj(28, 0xa8e37e0, 0xa8e37e0, 0xa8e37e0, 0xffd1eb3c) = 0xa8e5c80
12596 pthread_self(0xffd1eb20, 0xa8e5bd0, 0xffd1eb2c, 0x8a09943, 0x97969a0) = 0xf55a0720
12596 pthread_mutex_lock(0x97969a0, 0xffd1eb20, 0xa8e5bd0, 0xffd1eb2c, 0x8a09943) = 0
12596 pthread_mutex_unlock(0x97969a0, 0x97969a0, 0x97969a0, 0xa8e37e0, 0xa8e37e0) = 0
12596 pthread_mutex_lock(0x97969a0, 0xffd1eb20, 0xffd1eb2c, 0x8a09957, 0x97969a0) = 0
12596 pthread_mutex_unlock(0x97969a0, 0x97969a0, 0x97969a0, 0x97969a0, 0xa8e37e0) = 0
12596 pthread_getspecific(2, 0xffd1eb3c, 0x8a09bd6, 0x8c5b114, 0x8b6a1a8) = 0xa8e5bd0
12596 pthread_getspecific(0, 0xf656aff4, 0xffd1eba8, 0xf6422775, 1) = 0xa8e37e0
12596 _Znwj(172, 0xa8e37e0, 0xffd1eb48, 0x8563a8d, 0) = 0xa8e5ca0
12596 memcpy(0x0a8e5ca4, "", 16) = 0x0a8e5ca4
12596 pthread_getspecific(0, 0xa8e5ca0, 0xa8e5ca0, 0x8b6a1a8, 0xffd1eaf4) = 0xa8e37e0
12596 _Znwj(16, 0xa8e5ca0, 0xa8e5ca0, 0xffd1eae8, 0x8a0a67b) = 0xa8e5d50
 ... repeated x 40 or so
12596 _Znwj(16, 0xa8e5ca0, 0xa8e5ca0, 0xffd1eae8, 0x8a0a67b) = 0xa8e6038

**** here's the interesting part ******

12596 getenv("ACRO_ARG0") = "/usr/bin/acroread"
12596 getenv("ACRO_INSTALL_DIR") = "/opt/Adobe/Reader9/Reader"
12596 getenv("ACRO_CONFIG") = "intellinux"
12596 __xstat(3, "/opt/Adobe/Reader9/Reader", 0xffd1e948 <unfinished ...>
12596 SYS_stat64(0xffd20d42, 0xffd1e4bc, 0xf656aff4, 0xffd20df7, 0x8c3eed8) = 0
12596 <... __xstat resumed> ) = -1
12596 fprintf(0xf656b560, "%s must be executed from the sta"..., "/usr/bin/acroread" <unfinished ...>
12596 SYS_write(2, "/usr/bin/acroread must be execut"..., 60) = 60
12596 <... fprintf resumed> ) = 60
12596 exit(1 <unfinished ...>
12596 _ZNSsD1Ev(0x9789ae0, 1, 0xf643ae0b, 0x8c400c6, 1) = 0x97895e4
... destructors ...
12596 SYS_exit_group(1 <no return ...>
12596 +++ exited (status 1) +++

 So maybe there's something wrong with the env vars set by the startup script, and the binary doesn't think it's been run from the script? It bails right after checking a path built from ACRO_* env vars. That directory does exist.

  Wait, maybe there's a bug in ia32-libs, since the system call stat64 succeeded (return 0), but the __xstat returned -1. That doesn't look right, if ltrace is correct. Or maybe xstat is checking something in the struct stat.

ProblemType: Bug
Architecture: amd64
CrashDB: medibuntu
DistroRelease: Ubuntu 9.04
Package: acroread 9.1.0-7jaunty2+medibuntu1
ProcEnviron:
 SHELL=/bin/bash
 PATH=(custom, user)
 LANG=en_CA.utf8
 LANGUAGE=en_CA:en
SourcePackage: acroread
Uname: Linux 2.6.28-14-tesla x86_64
UnreportableReason: This is not a genuine Ubuntu package

[Lionel Le Folgoc (mrpouit)]

This package is now provided in "Canonical partner" repository, even for amd64. Could you retry with the 9.1.2-3jaunty1 version?

[xteejx (xteejx)]

We are closing this bug report because it lacks the information we need to investigate the problem, as described in the previous comments. Please reopen it if you can give us the missing information, and don't hesitate to submit bug reports in the future. To reopen the bug report you can click on the current status, under the Status column, and change the Status back to "New". Thanks again!

Note: The version in Maverick works fine.

[Darko Veberic (darko-veberic-kit)]

this bug resurfaced on a fresh install of ubuntu 13.10 amd64 (just a few days old, updated). the acroread was installed with AdbeRdr9.5.5-1_i486linux_enu.bin but i also tried dpkg -i AdbeRdr9.5.5-1_i386linux_enu.deb and apt-get -f install with the same results...

[Darko Veberic (darko-veberic-kit)]

the strace ends with:

22768 stat64("/opt/Adobe/Reader9/Reader", {st_mode=S_IFDIR|0755, st_size=4096, ...}) = 0
22768 write(2, "/usr/bin/acroread must be execut"..., 60) = 60
22768 exit_group(1) = ?
22768 +++ exited with 1 +++

so it seems someone is not happy about something in this directory /opt/Adobe/Reader9/Reader ...

[Darko Veberic (darko-veberic-kit)]

i tried all three methods of installing acroread on ubuntu 13.10 amd64: from AdbeRdr9.5.5-1_i486linux_enu.bin, from AdbeRdr9.5.5-1_i386linux_enu.deb, and enabling the raring partner repo. in all cases i got the library dependencies fixed, so this is not a problem. the message "/usr/bin/acroread must be executed from the startup script." still appears, no matter what.

i have modified the /usr/bin/acroread startup script so that it ran full verbosity strace and later also ltrace. the outputs of both runs are attached. what was not visible in strace appeared in ltrace: the __xstat() call on /opt/Adobe/Reader9/Reader/intellinux fails for unknown reason. i have checked all directory and file permissions in /opt/Adobe and should be fine. i also completely disabled apparmor, to no avail. the acroread refuses to start in saucy!

i would not care much for this non-open source program, but, sorry to say, this is the only program which prints pdf files correctly, with all the necessary options that one needs. evince, okular etc, they all don't get the printer margins right, do not have shrink/fit to printable area etc and have erratic behavior for long/short edge binding, not to speak about the multiple-pages-per-sheet.

please, fix the installation of acroread on ubuntu 13.10 and place it back to canonical partners...?

[Darko Veberic (darko-veberic-kit)]

see comment below and in this file the line near the end:

__xstat(3, "/opt/Adobe/Reader9/Reader/intellinux", 0xffdac1e8) = -1

[Launchpad Janitor (janitor)]

Status changed to 'Confirmed' because the bug affects multiple users.