Re: faxgetty exited on signal 6

Ryan Noll <[email protected]> Sat, 2 Apr 2016 17:29:14 -0700
Newsgroups gmane.comp.telephony.fax.hylafax.user
Message-ID <CAGzpdJAxi_2QPOfCSrh94pwFwkM9V4PYfb30zOJ4UOPz1JgA4g@mail.gmail.com>
On Tue, Mar 29, 2016 at 7:05 PM, Ryan Noll <[email protected]> wrote:
> On Tue, Mar 29, 2016 at 5:02 PM, Lee Howard <[email protected]> wrote:
>> On 03/29/2016 04:45 PM, Ryan Noll wrote:
>>>
>>> The assertion that failed when running faxgetty interactively is
>>>
>>> Assertion failed: (x == lastx), function _TIFFFax3fillruns, file
>>> tif_fax3.c, line 452.
>>> Abort
>>>
>>> Now that was completely different then where I thought the abort would
>>> occur. This would lead me to think that the problem may lay in the
>>> version of libtiff that is installed. The version I have installed is
>>> 4.04. I vaguely remember seeing some previous posts where HylaFax was
>>> not compatible with some versions of libtiff. Did I read that
>>> correctly? Does HylaFax+ have this same incompatibility?
>>
>>
>> Every time libtiff advances a step HylaFAX has to react to it.  So, in the
>> meantime there will be a libtiff version with which HylaFAX doesn't work.
>> However, I don't know that to be the case with HylaFAX+ 5.5.8 or HylaFAX.org
>> 6.0.6.  That said, I haven't used .org's releases since v4.2.5... so I don't
>> really know.
>>
>> Yes, that assert message is coming from libtiff.  So, it would seem to be a
>> bug in libtiff... or possibly in how HylaFAX is using libtiff.
>>
>> I would still recommend trying HylaFAX+.  If that fails then consider
>> stepping-back through libtiff versions (you'll have to rebuild+reinstall
>> HylaFAX after each) to see if the problem was introduced into libtiff
>> relatively recently (like, since the v4.0 release... so step-back to the
>> last v3).
>>
>> Thanks,
>>
>> Lee.
>
> I will try HylaFax+. I hope to report back success, but I'll need to
> test the build steps in a VM before I try it on a real box. (It has
> been a while since I've installed something outside the ports tree.) I
> hope to have something reported back because this may help someone
> else.
>
> Thank you for the assistance.
>
> Ryan

Hello,

I was able to install and configure HylaFax+ 5.5.8 without that much
difficulty on FreeBSD 10.2. I can report that the 100% CPU utilization
issue reported in Bug 166071 does does not occur in my setup, no
special flags were used to compile. The only minor difficulty that I
ran into is that gcc needs to be installed (I installed via ports,
lang/gcc) before HylaFax+'s configure is run.

To make life a little easier on myself, I used the HylaFax in ports to
compile the necessary dependencies. A list of the ports installed is
listed below, some installed programs are due to personal preference
and were build dependencies (a * denotes the HylaFax port
dependencies, ** denotes probably a good idea to install beforehand):

*  afm-1.0_1
autoconf-2.69
autoconf-wrapper-20131203
automake-1.15_1
automake-wrapper-20131203
binutils-2.25.1_1,1
bison-2.7.1,1
ca_root_nss-3.22.2
cups-client-2.0.3_2
cups-image-2.0.3_2
dialog4ports-0.1.5_2
expat-2.1.0_3
freetype2-2.6.3
gcc-4.8.5_2
gettext-runtime-0.19.7
gettext-tools-0.19.7
**  ghostscript9-nox11-9.06_10
gmake-4.1_2
gmake-lite-4.1_1
gmp-5.1.3_3
gnutls-3.4.10
gsfonts-8.11_6
help2man-1.43.3_1
indexinfo-0.2.4
jbig2dec-0.12
*  jbigkit-2.1_1
jpeg-turbo-1.4.2
lcms2-2.7_2
libevent2-2.0.22_1
libffi-3.2.1
libgcrypt-1.6.5_1
libgpg-error-1.21
libidn-1.31
libpaper-1.1.24.4
libtasn1-4.7
libtool-2.4.6
m4-1.4.17_1,1
mDNSResponder-576.30.4
mpc-1.0.3
mpfr-3.1.3_1
nasm-2.11.08_1,1
nettle-3.2
p11-kit-0.23.2
p5-Locale-gettext-1.06
perl5-5.20.3_8
pkg-1.6.2
pkgconf-0.9.12_1
png-1.6.21
svgalib-1.4.3_7
*  tiff-4.0.6_1
tmux-2.1_1
trousers-tddl-0.3.10_7
vim-lite-7.4.1556
wget-1.16.3_1

Please note that the versions installed are current as to my installation.

Configuration was straightforward and pretty much was exactly the same
as "regular" HylaFax. The setup is working pretty well using
FreeSwitch as the virtual modem, and I am able to fax over a FiOS line
and a BasicTalk line. As far as reliability numbers, I have not run
anything conclusive as I am just sending one page type tests to the
"Fax Toy" number and limited tests using "Fax Zero".

Note: I am using FreeSwitch to communicate to Grandstream GXW4108 (FXO
gateway) and Grandstream GXW4008 (FXS gateway) ATAs. To get faxes
working, it appears that setting the T.38 FAX Settings to Passthrough,
9600, and disabled ECM (4108: Settings, Channels Settings page and
4008: Profile 1).

The firmware version for the GXW-4008 is Program 1.0.14.1, Bootloader
1.0.0.18, Core 1.0.14.1, Base 1.0.14.1, Extra 1.0.14.1, CPE 1.0.1.48.
The firmware version for the GXW-4108 is Program 1.4.1.5, Loader
1.1.3.4, Boot 1.1.3.2.

The settings in autoload-configs/spandsp.conf.xml are (abbreviated)
the following:

    <fax-settings>
        <param name="use-ecm" value="true"/>
        <param name="verbose" value="true"/>
        <param name="disable-v17" value="true"/>
        <param name="enable-t38" value="true"/>
        <param name="enable-t38-request" value="true"/>
        <param name="ident" value="1234567890"/>
        <param name="header" value="Name"/>
        <param name="spool-dir" value="/tmp"/>
        <param name="file-prefix" value="faxrx"/>
    </fax-settings>

Note: I did report Bug #166071, but I am going to comment on the bug
and recommend installing HylaFax+.

My apologies for this message being so long, but hopefully it will
help someone else...

Ryan


____________________ HylaFAX(tm) Users Mailing List _______________________
  To subscribe/unsubscribe, click http://lists.hylafax.org/cgi-bin/lsg2.cgi
 On UNIX: mail -s unsubscribe [email protected] < /dev/null
  *To learn about commercial HylaFAX(tm) support, mail [email protected].*