[sword-devel] imp2vs segfault

[Greg Hellings]

On Tue, Sep 6, 2011 at 11:01 PM, Chris Little <chrislit at crosswire.org> wrote:
> The utils tend to choke when you throw data at them in a radically wrong
> versification system. Specifically, they'll choke if you give them texts
> with extra chapters. Joel only has 3 chapters in the KJV versification,
> which you're using, so you get a segfault when you feed imp2vs a Bible with
> four chapters in Joel.

Well that explains that phenomenon.

>
> There is a bug somewhere, but fixing the bug still isn't going to make this
> the correct versification.

An error message would at least be nice. "Hey, not sure what to do
with this verse."

>
> One of the Catholic versifications is probably the closest we have. I'm not
> sure about the "endlessly repeating messages of linking different verses
> together." Are they literally endless and repeating the same linkage? That's
> not a bug that I know of in any of the tools. (If it's the same verse
> repeating, which verse is it? And have you checked that there isn't some
> sort of error in the verse numbering in/around that verse?)

I would just like to reiterate - the files are freely available for
anyone to take a look at and test out.  This way anyone can see what
is going on and hopefully help come to an informed decision about what
is going on.

http://dl.thehellings.com/SL_FR_BI_PDV.thml.lorem
http://dl.thehellings.com/SL_PG_BI_TPC.thml.lorem

These are the two files in question.  I have now uploaded a copy of
the output of this command
$ imp2vs SL_FR_BI_PDV.thml.lorem -4 -o SL_FR_BI_PDV -v Catholic 2>&1 >
SL_FR_BI_PDV.thml.lorem.err
to the above location as well for you to check out.  It is a 95MB file
that resulted from allowing the above command to run for about 20 or
30 seconds before doing a keyboard break.  I'm looking at about 2.25
million lines of output there which start out perfectly normal
from file: Ge 1:1
adding entry: Genesis 1:1 length 240/240
from file: Ge 1:2
adding entry: Genesis 1:2 length 149/149
from file: Ge 1:3
adding entry: Genesis 1:3 length 136/136
...

And then turn into
adding entry: Amos 9:15 length 2159/2159
from file: Ob 1
adding entry: Obadiah 1:1 length 823/823
linking entry: Obadiah 1:2 to Obadiah 1:1
linking entry: Obadiah 1:3 to Obadiah 1:1
linking entry: Obadiah 1:4 to Obadiah 1:1
<snip>
linking entry: Obadiah 1:21 to Obadiah 1:1
from file: Ob 2
adding entry: Obadiah 2:1 length 492/492
linking entry: [ Module Heading ] to Obadiah 2:1
linking entry: [ Testament 1 Heading ] to Obadiah 2:1
linking entry: Genesis 0:0 to Obadiah 2:1
linking entry: Genesis 1:0 to Obadiah 2:1
linking entry: Genesis 1:1 to Obadiah 2:1
linking entry: Genesis 1:2 to Obadiah 2:1
linking entry: Genesis 1:3 to Obadiah 2:1
<snip>
linking entry: Jonah 1:14 to Obadiah 2:1
linking entry: Jonah 1:15 to Obadiah 2:1
linking entry: Jonah 1:16 to Obadiah 2:1
from file: Ob 3
adding entry: Obadiah 3:1 length 465/465
linking entry: [ Module Heading ] to Obadiah 3:1
linking entry: [ Testament 1 Heading ] to Obadiah 3:1
linking entry: Genesis 0:0 to Obadiah 3:1
linking entry: Genesis 1:0 to Obadiah 3:1
linking entry: Genesis 1:1 to Obadiah 3:1
linking entry: Genesis 1:2 to Obadiah 3:1
linking entry: Genesis 1:3 to Obadiah 3:1
<snip>
linking entry: Jonah 2:8 to Obadiah 3:1
linking entry: Jonah 2:9 to Obadiah 3:1
linking entry: Jonah 2:10 to Obadiah 3:1
linking entry: Jonah 2:11 to Obadiah 3:1
from file: Ob 4
adding entry: Obadiah 4:1 length 438/438
linking entry: [ Module Heading ] to Obadiah 4:1
linking entry: [ Testament 1 Heading ] to Obadiah 4:1
linking entry: Genesis 0:0 to Obadiah 4:1
linking entry: Genesis 1:0 to Obadiah 4:1
linking entry: Genesis 1:1 to Obadiah 4:1

And so on. I don't see what's so special about this point in the
module, but maybe someone else can?  Again, I encourage you all to
take a look at the files yourself and see if you all can see something
I'm missing.

--Greg


>
> There is also a French versification system that we know of but haven't
> implemented yet, which might be a better fit.
>
> --Chris
>
>
> On 9/6/2011 10:30 AM, Greg Hellings wrote:
>>
>> Wouldn't know.  I don't go near Windows without hip waders (or the
>> promise of a paycheck, when it comes to my day job) on if I can help
>> it. Plus, I doubt that imp2vs.exe is as up to date as the version I'm
>> using in Linux.  You're more than welcome to give it a try.  I used
>> the command
>>
>> imp2vs<file>  -4 -o .
>>
>> to do the import.  I also tried adding '-v Catholic' in, since these
>> are French and Portugese modules, but that results in endlessly
>> repeating messages of linking different verses together.
>>
>> --Greg
>>
>> On Tue, Sep 6, 2011 at 12:24 PM, David Haslam<dfhmch at googlemail.com>
>>  wrote:
>>>
>>> Does the same segmentation fault occur when using imp2vs.exe in Windows?
>>>
>>> David
>>>
>>> --
>>> View this message in context:
>>> http://sword-dev.350566.n4.nabble.com/imp2vs-segfault-tp3793577p3794046.html
>>> Sent from the SWORD Dev mailing list archive at Nabble.com.
>>>
>>> _______________________________________________
>>> sword-devel mailing list: sword-devel at crosswire.org
>>> http://www.crosswire.org/mailman/listinfo/sword-devel
>>> Instructions to unsubscribe/change your settings at above page
>>>
>>
>> _______________________________________________
>> sword-devel mailing list: sword-devel at crosswire.org
>> http://www.crosswire.org/mailman/listinfo/sword-devel
>> Instructions to unsubscribe/change your settings at above page
>
>
> _______________________________________________
> sword-devel mailing list: sword-devel at crosswire.org
> http://www.crosswire.org/mailman/listinfo/sword-devel
> Instructions to unsubscribe/change your settings at above page
>