mdbtools-dev

Re: [mdb-dev] problem exporting data from a mdb file

[Brian B. <bri...@gm...>]

Howard,

Is it possible for you to try something more recent than 0.5
(preferably CVS) as I believe this is a corrected issue.

Brian


2005/8/22, how...@ko... <how...@ko...>:
> When I export data from a mdb file the first column of data is null.  I a=
m
> using the 0.5 version on Solaris.  I have also tried the same scenario on=
 a
> Linux system and the export worked correctly.  It seems to be a solaris
> only problem.    Any help would be appreciated.  Thanks.
>=20
> Howard
>=20
>=20
>=20
> -------------------------------------------------------
> SF.Net email is Sponsored by the Better Software Conference & EXPO
> September 19-22, 2005 * San Francisco, CA * Development Lifecycle Practic=
es
> Agile & Plan-Driven Development * Managing Projects & Teams * Testing & Q=
A
> Security * Process Improvement & Measurement * http://www.sqe.com/bsce5sf
> _______________________________________________
> mdbtools-dev mailing list
> mdb...@li...
> https://lists.sourceforge.net/lists/listinfo/mdbtools-dev
>

[mdb-dev] problem exporting data from a mdb file

[<how...@ko...>]

When I export data from a mdb file the first column of data is null.  I am
using the 0.5 version on Solaris.  I have also tried the same scenario on a
Linux system and the export worked correctly.  It seems to be a solaris
only problem.    Any help would be appreciated.  Thanks.

Howard

Re: [mdb-dev] Building mdbtools-0.5 on Solaris 2.9

[Jeff S. <why...@ya...>]

--- "Richard A. O'Keefe" <ok...@cs...> wrote:
> I want to build mdbtools-0.5 on a SPARC/Solaris 2.9 box.

FWIW, I believe at one time I got mdbtools working on a Sparc/Solaris
(2.)8 machine.

> I'm doing research on some medical data, and some important background
> information has just been made available to me, but only as a .mdb file.
> (It's no use me asking the supplier for something portable, they're only
> a government department and have everything but a clue.)

I work for a government contractor in the USA, so I understand (note that
my work on mdbtools is done on my own time).

> 1.  I only want to build and install the command-line tools.
>     Whyever do I need glib for _that_?

Glib provides stuff such as data structures and other core functionality
that is not provided in standard C.  Note that it does not contain any
graphical functionality.  For the graphical application we use Gtk+,
which does build on top of Glib, to provide graphical functionality.

> 2.  I downloaded and installed glib-2.7.7.
>     Now this is glib 2 point 7 point 7, and the .pc file that's
>     created for it rightly says so.
> 
>     BUT glib-2.7.7 installs everything *else* as "2.0", not 2.7.7.
> 
>     When I do the configure step for mdbtools, it asks pkg-config,
>     and pkg-config says that the version is 2.7.7, but then the
>     configure script screeches in horror that glib is *really* 2.0.
> 
>     Configure is wrong, glib really *really* is 2.7.7, it's just *called* 2.0.
> 
>     i backed out everything I could, reinstalled, and hacked the .pc
>     file to say that the version was 2.0.  That didn't work.  I tried
>     2.0.0, but *that* didn't work.

I believe your Glib install was correct, but the checks in mdbtools 0.5
must not have worked correctly.

> 3.  When I run gmake check during glib installation, some of the tests
>     pass, and some, concerned IIRC with threads, fail.  Since I only
>     want to run the command line tools, does this matter at all?

mdbtools does not utilize Glib's threading features, so it should not matter.

> So, given that I don't particularly want to install a complete Linux
> environment on my Solaris box, and just want to run the command line
> tools, how do I install mdbtools-0.5?

As both 0.5 and 0.6pre1 are so old, I must suggest building from what is in
mdbtools' cvs repository.  That will require that the autotools triad be
installed if it is not already (automake, autoconf, and libtool).

 -- Jeff Smith

__________________________________________________
Do You Yahoo!?
Tired of spam?  Yahoo! Mail has the best spam protection around 
http://mail.yahoo.com 

[mdb-dev] mdb-schema: type recognition?

[Markus N. <ne...@it...>]

Hi,

I am (still) trying to get some MDB files with meteo
data into PostgreSQL8. Doing so, I observed that without
modifications the resulting DB became 25 times larger
than the original.

I realized, that mdb-schema didn't write varchar(length) but
char(length). With a sed modification:

export MDBFILE=StazioniAutomatiche.mdb
export MDB_ICONV=utf-8
export PGDB=meteodata
mdb-schema -S $MDBFILE postgres  | sed 's+Char+varchar+g' | psql $PGDB

and a subsequent mdb-export I reduced the resulting DB
size to 8 times larger (still too much for me).

Then I installed the ODBC driver of PostgreSQL8 on a MS-Windows
machine and transferred one of the tables via OBDC.
Interestingly, the type definitions are quite different:

# MDB-TOOLS to PostgreSQL8:
meteodata=# \d s7003
                  Table "public.s7003"
    Column     |            Type             | Modifiers
---------------+-----------------------------+-----------
 dataunix      | bigint                      |
 canale        | smallint                    |
 codice_misura | integer                     |
 campo4        | character varying(510)      |
 campo5        | character varying(510)      |
 campo6        | bigint                      |
 data          | timestamp without time zone |
 campo8        | bigint                      |
 campo9        | bigint                      |
 misura        | double precision            |
 campo11       | bigint                      |


#From MS-ACCESS via ODBC to PostgreSQL8:
meteodataodbc=# \d s7003
                 Table "public.s7003"
    Column     |           Type           | Modifiers
---------------+--------------------------+-----------
 DataUnix      | integer                  |
 Canale        | smallint                 |
 Codice Misura | smallint                 |
 Campo4        | character varying(254)   |
 Campo5        | character varying(254)   |
 Campo6        | integer                  |
 Data          | timestamp with time zone |
 Campo8        | integer                  |
 Campo9        | integer                  |
 Misura        | double precision         |
 Campo11       | integer                  |

The differences are:
- integer vs bigint
- smallint vx integer
- character varying(254) vs character varying(510) (without sed it
  were char(510) in mdbtools)
- timestamp with time zone vs timestamp without time zone

Is there a chance to better recognize the data types in
the MDB-tools (assuming that ODBC is telling the truth)?

Best regards

 Markus

Re: [mdb-dev] Problem with mdbtools, Fedora Core 4 and libreadline.so

[Sam M. <pa...@gm...>]

You could force it and sym link .4 to .5; or you could rebuild the
package, or use the CVS (which is the best option as it has the most
features)

On 8/20/05, Alan Moss <am...@sm...> wrote:
> Hello,
>=20
> In trying to load the mdbtools-0.5-1.i386.rpm onto my Fedora Core 4 syste=
m
> I get the message that libreadline.so.4 is needed by mdbtools-0.5-1.i386.
> However libreadline.so.5 is already loaded on my system.  Do you have any
> suggestions on how to resolve this conflict?
>=20
> I sent this previously but the response was that I am not a member.
> However I am a member, and I subscribed as am...@sm....  Hopefully th=
e
> reply-to address on this email has been corrected to reflect the proper
> email address.
>=20
> Thank you,
> Alan Moss
>=20
>=20
>=20
>=20
>=20
>=20
> -------------------------------------------------------
> SF.Net email is Sponsored by the Better Software Conference & EXPO
> September 19-22, 2005 * San Francisco, CA * Development Lifecycle Practic=
es
> Agile & Plan-Driven Development * Managing Projects & Teams * Testing & Q=
A
> Security * Process Improvement & Measurement * http://www.sqe.com/bsce5sf
> _______________________________________________
> mdbtools-dev mailing list
> mdb...@li...
> https://lists.sourceforge.net/lists/listinfo/mdbtools-dev
>

[mdb-dev] Building mdbtools-0.5 on Solaris 2.9

[Richard A. O'K. <ok...@cs...>]

I want to build mdbtools-0.5 on a SPARC/Solaris 2.9 box.
I'm doing research on some medical data, and some important background
information has just been made available to me, but only as a .mdb file.
(It's no use me asking the supplier for something portable, they're only
a government department and have everything but a clue.)
I have *no* interest in working with this data in Access form,
I just want to get it out as quickly as I can.

I was delighted to find mdbtools, but I have had a rather stressful
and markedly unsuccessful couple of days trying to install it.  (And
I have been installing software from the net onto Unix boxes since
1982 and in particular am reasonably familiar with configure/gmake.)

1.  I only want to build and install the command-line tools.
    Whyever do I need glib for _that_?

2.  I downloaded and installed glib-2.7.7.
    Now this is glib 2 point 7 point 7, and the .pc file that's
    created for it rightly says so.

    BUT glib-2.7.7 installs everything *else* as "2.0", not 2.7.7.

    When I do the configure step for mdbtools, it asks pkg-config,
    and pkg-config says that the version is 2.7.7, but then the
    configure script screeches in horror that glib is *really* 2.0.

    Configure is wrong, glib really *really* is 2.7.7, it's just *called* 2.0.

    i backed out everything I could, reinstalled, and hacked the .pc
    file to say that the version was 2.0.  That didn't work.  I tried
    2.0.0, but *that* didn't work.

3.  When I run gmake check during glib installation, some of the tests
    pass, and some, concerned IIRC with threads, fail.  Since I only
    want to run the command line tools, does this matter at all?

4.  Surprising as it may seem, just because I'm in a Computer Science
    department and have worked as a software developer and teach
    software engineering and want to install this on the machine on my
    desktop, which is not used by anyone else, &c &c, it does *NOT*
    follow that I have root access.  It's University policy that if a
    machine is connected to the campus network, *only* a specially
    approved sysadmin may have root access to it.  A prebuilt binary
    that had to go in /usr/local would be of very little use to me.
    Our sysadmin has created a /users/local/ which I *do* have write
    access to, and if/when I get mdbtools running, that's where the
    stuff will have to go.

So, given that I don't particularly want to install a complete Linux
environment on my Solaris box, and just want to run the command line
tools, how do I install mdbtools-0.5?

[mdb-dev] Problem with mdbtools, Fedora Core 4 and libreadline.so

[Alan M. <am...@sm...>]

Hello,

In trying to load the mdbtools-0.5-1.i386.rpm onto my Fedora Core 4 system
I get the message that libreadline.so.4 is needed by mdbtools-0.5-1.i386. 
However libreadline.so.5 is already loaded on my system.  Do you have any
suggestions on how to resolve this conflict?

I sent this previously but the response was that I am not a member. 
However I am a member, and I subscribed as am...@sm....  Hopefully the
reply-to address on this email has been corrected to reflect the proper
email address.

Thank you,
Alan Moss

[mdb-dev] Problem with mdbtools, Fedora Core 4 and libreadline.so

[Alan M. <am...@sm...>]

Hello,

In trying to load the mdbtools-0.5-1.i386.rpm onto my Fedora Core 4 system
I get the message that libreadline.so.4 is needed by mdbtools-0.5-1.i386. 
However libreadline.so.5 is already loaded on my system.  Do you have any
suggestions on how to resolve this conflict?

Thank you,
Alan Moss

[mdb-dev] patch to not #include <config.h>

[Rene E. <re...@de...>]

forwarded 323256 mdb...@li...
tag 323256 + patch
thanks

[ For mdbtools-dev: Background is
 http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=3D323256, therefore pleas
 let 32...@bu... and me (re...@de...) in the loop ]

The following patch fixes the build of mdbtools and libmdbtools itself
to not need #include <config.h> because not every project using
libmdbtools has that and #include'ing it is bad anyway:

Index: configure.in
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D
RCS file: /cvsroot/mdbtools/mdbtools/configure.in,v
retrieving revision 1.36
diff -u -u -r1.36 configure.in
--- configure.in        29 Mar 2005 03:40:48 -0000      1.36
+++ configure.in        16 Aug 2005 20:03:41 -0000
@@ -21,6 +21,12 @@

 AM_ICONV

+# add -DHAVE_INCONV to CFLAGS. We should not use #include <config.h> in
+# public headers
+if test "$am_cv_func_iconv" =3D "yes"; then
+       CFLAGS=3D"$CFLAGS -DHAVE_ICONV"
+fi
+
 dnl no optional stuff by default
 OPTDIRS=3D""

Index: include/mdbtools.h
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D
RCS file: /cvsroot/mdbtools/mdbtools/include/mdbtools.h,v
retrieving revision 1.77
diff -u -u -r1.77 mdbtools.h
--- include/mdbtools.h  5 Aug 2005 15:52:16 -0000       1.77
+++ include/mdbtools.h  16 Aug 2005 20:03:41 -0000
@@ -32,7 +32,6 @@
 #include <ctype.h>
 #include <string.h>
 #include <glib.h>
-#include <config.h>

 #ifdef HAVE_ICONV
 #include <iconv.h>

This patch simply removed the need for #including config.h by just specifyi=
ng
-DHAVE_ICONV on the mdbtools build since that is what seems to be why confi=
g.h
is included currently...

While doing this I found another problem. If someone should build on a syst=
em
where there may be a iconv available or may not or if the application using
libmdbtools does not use autoconf and therefore does not set HAVE_ICONV bad
things may happen because of this from mdbtools.h:

typedef struct {
[...]
        MdbStatistics *stats;
#ifdef HAVE_ICONV
        iconv_t iconv_in;
        iconv_t iconv_out;
#endif
} MdbHandle;

Build without HAVE_ICONV miss two struct member; this is a ABI change.

Well, I can work around on OpenOffice.org by just specifying -DHAVE_ICONV
when building the mdbtools-using stuff but that bug remains and should be
fixed in the future, but this is not a regression since this bug exists too
if the projects' config.h you #include doesn't #define HAVE_ICONV anyway..

Gr=FC=DFe/Regards,

Rene
--=20
 .''`.  Rene Engelhard -- Debian GNU/Linux Developer
 : :' : http://www.debian.org | http://people.debian.org/~rene/
 `. `'  re...@de... | GnuPG-Key ID: 248AEB73
   `-   Fingerprint: 41FA F208 28D4 7CA5 19BB  7AD9 F859 90B0 248A EB73

Re: [mdb-dev] mdb-export and PostgreSQL

[Markus N. <ne...@it...>]

On Tue, Aug 09, 2005 at 06:12:14PM -0400, Brian Bruns wrote:
> Hi,
> 
> The problem is that mdb-export was designed to export to CSV format in
> which case " is the proper quoting character.  Somewhere along the
> line somebody contributed the -I option.
> 
> There is a -R option to set the row delimiter.  So you could say something like:
> 
> mdb-export -I -R ';
> ' Northwind.mdb Shippers
> 
> and get the desired results, although I find this sort of a kludge
> since mdb-export doesn't use the delimiter as a mask, everything is
> taken literaly, so the preferable version:
> 
> mdb-export -I -R ';\n' Northwind.mdb Shippers
> 
> doesn't work as one would hope....that should be a simple fix, I'll
> see if I can get to it this week.
> 
> Third option is to use good old sed (which you alluded to) like so:
> 
> mdb-export -I Northwind.mdb Shippers | sed -e 's/$/;/'
> 
> which will produce the desired results.
> 
> As for changing the quoting character, it should be configurable as
> well.  Default should be " (double quote) when doing CSV style export
> and ' (single quote) when using the -I switch.  I'll see if I can
> knock this out something this week too.

This would be cool. For now I used

 mdb-export -S -I StazioniAutomatiche.mdb Misure | sed 's+$+;+g'| tr '"' "'" | psql meteo

(see the nice 'tr' trick  :-).

Markus  

> Brian
> 
> 2005/8/9, Markus Neteler <ne...@it...>:
> > Hi,
> > 
> > I am currently trying to get MDB data into a PostgreSQL database.
> > Running
> > 
> > mdb-export -S -I StazioniAutomatiche.mdb Misure
> > ...
> > INSERT INTO Misure (Id_Misura, Descrizione_Misura, UM) VALUES (236,"Tensione minima giornaliera batteria","V")
> > ...
> > 
> > To make PostgreSQL happy, a semicolon is needed at the end of each
> > line (also reported:
> >  http://sourceforge.net/mailarchive/forum.php?thread_id=7234522&forum_id=5183
> > ). I have patched src/util/mdb-export.c accordingly (line 185), but I feel
> > that this should go somehow into CVS.
> > 
> > A second problem is that all " (double) quotes need to be ' (single) quotes,
> > otherwise PostgreSQL won't insert. Not sure where to fix this one in
> > the C code.
> > 
> > I have a shell workaround using tr and sed, but would prefer a fix in
> > the source code (to avoid future traps).
> > 
> > Thanks
> > 
> >  Markus
> > 
> > 
> > 
> > -------------------------------------------------------
> > SF.Net email is Sponsored by the Better Software Conference & EXPO
> > September 19-22, 2005 * San Francisco, CA * Development Lifecycle Practices
> > Agile & Plan-Driven Development * Managing Projects & Teams * Testing & QA
> > Security * Process Improvement & Measurement * http://www.sqe.com/bsce5sf
> > _______________________________________________
> > mdbtools-dev mailing list
> > mdb...@li...
> > https://lists.sourceforge.net/lists/listinfo/mdbtools-dev
> >

-- 
Markus Neteler     <neteler itc it>       http://mpa.itc.it
ITC-irst -  Centro per la Ricerca Scientifica e Tecnologica
MPBA - Predictive Models for Biol. & Environ. Data Analysis
Via Sommarive, 18        -       38050 Povo (Trento), Italy

Re: [mdb-dev] mdb-export and PostgreSQL

[Brian B. <bri...@gm...>]

Hi,

The problem is that mdb-export was designed to export to CSV format in
which case " is the proper quoting character.  Somewhere along the
line somebody contributed the -I option.

There is a -R option to set the row delimiter.  So you could say something =
like:

mdb-export -I -R ';
' Northwind.mdb Shippers

and get the desired results, although I find this sort of a kludge
since mdb-export doesn't use the delimiter as a mask, everything is
taken literaly, so the preferable version:

mdb-export -I -R ';\n' Northwind.mdb Shippers

doesn't work as one would hope....that should be a simple fix, I'll
see if I can get to it this week.

Third option is to use good old sed (which you alluded to) like so:

mdb-export -I Northwind.mdb Shippers | sed -e 's/$/;/'

which will produce the desired results.

As for changing the quoting character, it should be configurable as
well.  Default should be " (double quote) when doing CSV style export
and ' (single quote) when using the -I switch.  I'll see if I can
knock this out something this week too.

Brian

2005/8/9, Markus Neteler <ne...@it...>:
> Hi,
>=20
> I am currently trying to get MDB data into a PostgreSQL database.
> Running
>=20
> mdb-export -S -I StazioniAutomatiche.mdb Misure
> ...
> INSERT INTO Misure (Id_Misura, Descrizione_Misura, UM) VALUES (236,"Tensi=
one minima giornaliera batteria","V")
> ...
>=20
> To make PostgreSQL happy, a semicolon is needed at the end of each
> line (also reported:
>  http://sourceforge.net/mailarchive/forum.php?thread_id=3D7234522&forum_i=
d=3D5183
> ). I have patched src/util/mdb-export.c accordingly (line 185), but I fee=
l
> that this should go somehow into CVS.
>=20
> A second problem is that all " (double) quotes need to be ' (single) quot=
es,
> otherwise PostgreSQL won't insert. Not sure where to fix this one in
> the C code.
>=20
> I have a shell workaround using tr and sed, but would prefer a fix in
> the source code (to avoid future traps).
>=20
> Thanks
>=20
>  Markus
>=20
>=20
>=20
> -------------------------------------------------------
> SF.Net email is Sponsored by the Better Software Conference & EXPO
> September 19-22, 2005 * San Francisco, CA * Development Lifecycle Practic=
es
> Agile & Plan-Driven Development * Managing Projects & Teams * Testing & Q=
A
> Security * Process Improvement & Measurement * http://www.sqe.com/bsce5sf
> _______________________________________________
> mdbtools-dev mailing list
> mdb...@li...
> https://lists.sourceforge.net/lists/listinfo/mdbtools-dev
>

[mdb-dev] mdb-export and PostgreSQL

[Markus N. <ne...@it...>]

Hi,

I am currently trying to get MDB data into a PostgreSQL database.
Running

mdb-export -S -I StazioniAutomatiche.mdb Misure
...
INSERT INTO Misure (Id_Misura, Descrizione_Misura, UM) VALUES (236,"Tensione minima giornaliera batteria","V")
...

To make PostgreSQL happy, a semicolon is needed at the end of each
line (also reported:
 http://sourceforge.net/mailarchive/forum.php?thread_id=7234522&forum_id=5183
). I have patched src/util/mdb-export.c accordingly (line 185), but I feel
that this should go somehow into CVS.

A second problem is that all " (double) quotes need to be ' (single) quotes,
otherwise PostgreSQL won't insert. Not sure where to fix this one in 
the C code.

I have a shell workaround using tr and sed, but would prefer a fix in
the source code (to avoid future traps).

Thanks

 Markus

Re: [mdb-dev] Re: problems with JET4 and nonASCI characters

[Jeff S. <why...@ya...>]

--- Markus Neteler <ne...@it...> wrote:
> On Fri, Aug 05, 2005 at 12:29:56PM -0400, Brian Bruns wrote:
> > export MDB_ICONV=ISO-8859-1

> Phantastic - problem solved.
> 
> Would you mind to add this to the docs?

It is currently described in the 'HACKING' document, but it
really does need to go somewhere in 'user' documentation.

 -- Jeff Smith


		
____________________________________________________
Start your day with Yahoo! - make it your home page 
http://www.yahoo.com/r/hs 
 

Re: [mdb-dev] Re: problems with JET4 and nonASCI characters

[Brian B. <bri...@gm...>]

Doh!

I thought it was in the user guide, but you're right it's not.  I will
add it very soon.

The unix world will be a easier place once everyone is on utf8.

Brian

2005/8/8, Markus Neteler <ne...@it...>:
> On Fri, Aug 05, 2005 at 12:29:56PM -0400, Brian Bruns wrote:
> > The source/target character sets can be selected using the MDB_ICONV
> > environment variable, otherwise the default is UTF-8 (this all applies
> > to jet4, access 97 is a different beast).  So what you're likely
> > seeing is UTF-8 characters on a terminal/program which is unequiped to
> > handle it.  Just set the target set like:
> >
> > export MDB_ICONV=3DISO-8859-1
> >
> > Which will work for most western european languages.  use "iconv
> > --list" to see the supported values.
>=20
> Phantastic - problem solved.
>=20
> Would you mind to add this to the docs?
>=20
> eden:doc[359.103] pwd
> /hardmnt/eden0/ssi/software/mdbtools/doc
> eden:doc[360.104] grep MDB_ICONV *
> eden:doc[361.105]
>=20
>=20
> > Eventually it would be nice to grab this info from LC_ALL or LANG, but
> > the string returned by those doesn't always map exactly to that
> > required by iconv.
>=20
> Ok.
>=20
> Thanks, Brian.
>=20
> Markus
>=20
> > Brian
> >
> > 2005/8/5, Markus Neteler <ne...@it...>:
> > > Hi again,
> > >
> > > after the fix for the 64bit platform I can now export the data.
> > > But I have the same problem as reported in
> > >  http://sourceforge.net/mailarchive/forum.php?thread_id=3D7699603&for=
um_id=3D5183
> > >
> > > mdb-ver StazioniAutomatiche.mdb
> > > JET4
> > >
> > > mdb-export -I StazioniAutomatiche.mdb Misure
> > > ...
> > > INSERT INTO Misure (Id Misura, Descrizione Misura, UM) VALUES (174,"U=
midit=E0 Massima 1440 min","%")
> > > ...
> > > INSERT INTO Misure (Id Misura, Descrizione Misura, UM) VALUES (6464,"=
Velocit=E0 vento Massima 15 min","m/s")
> > > ...
> > >
> > > Do you see a chance to maintain the character encoding somehow or
> > > map to ASCII? I am not sure if SQL is limited or permits accented
> > > characters.
> > >
> > > My intention is to convert the MDB into a PostgreSQL database.
> > >
> > > Thanks
> > >
> > >  Markus
> > >
> > >
> > > -------------------------------------------------------
> > > SF.Net email is Sponsored by the Better Software Conference & EXPO
> > > September 19-22, 2005 * San Francisco, CA * Development Lifecycle Pra=
ctices
> > > Agile & Plan-Driven Development * Managing Projects & Teams * Testing=
 & QA
> > > Security * Process Improvement & Measurement * http://www.sqe.com/bsc=
e5sf
> > > _______________________________________________
> > > mdbtools-dev mailing list
> > > mdb...@li...
> > > https://lists.sourceforge.net/lists/listinfo/mdbtools-dev
> > >
> >
> >
> > -------------------------------------------------------
> > SF.Net email is Sponsored by the Better Software Conference & EXPO
> > September 19-22, 2005 * San Francisco, CA * Development Lifecycle Pract=
ices
> > Agile & Plan-Driven Development * Managing Projects & Teams * Testing &=
 QA
> > Security * Process Improvement & Measurement * http://www.sqe.com/bsce5=
sf
> > _______________________________________________
> > mdbtools-dev mailing list
> > mdb...@li...
> > https://lists.sourceforge.net/lists/listinfo/mdbtools-dev
>=20
> --
> Markus Neteler     <neteler itc it>       http://mpa.itc.it
> ITC-irst -  Centro per la Ricerca Scientifica e Tecnologica
> MPBA - Predictive Models for Biol. & Environ. Data Analysis
> Via Sommarive, 18        -       38050 Povo (Trento), Italy
>

Re: [mdb-dev] Re: problems with JET4 and nonASCI characters

[Markus N. <ne...@it...>]

On Fri, Aug 05, 2005 at 12:29:56PM -0400, Brian Bruns wrote:
> The source/target character sets can be selected using the MDB_ICONV
> environment variable, otherwise the default is UTF-8 (this all applies
> to jet4, access 97 is a different beast).  So what you're likely
> seeing is UTF-8 characters on a terminal/program which is unequiped to
> handle it.  Just set the target set like:
>=20
> export MDB_ICONV=3DISO-8859-1
>=20
> Which will work for most western european languages.  use "iconv
> --list" to see the supported values.

Phantastic - problem solved.

Would you mind to add this to the docs?

eden:doc[359.103] pwd
/hardmnt/eden0/ssi/software/mdbtools/doc
eden:doc[360.104] grep MDB_ICONV *
eden:doc[361.105]=20

=20
> Eventually it would be nice to grab this info from LC_ALL or LANG, but
> the string returned by those doesn't always map exactly to that
> required by iconv.

Ok.

Thanks, Brian.

Markus
=20
> Brian
>=20
> 2005/8/5, Markus Neteler <ne...@it...>:
> > Hi again,
> >=20
> > after the fix for the 64bit platform I can now export the data.
> > But I have the same problem as reported in
> >  http://sourceforge.net/mailarchive/forum.php?thread_id=3D7699603&for=
um_id=3D5183
> >=20
> > mdb-ver StazioniAutomatiche.mdb
> > JET4
> >=20
> > mdb-export -I StazioniAutomatiche.mdb Misure
> > ...
> > INSERT INTO Misure (Id Misura, Descrizione Misura, UM) VALUES (174,"U=
midit=E0 Massima 1440 min","%")
> > ...
> > INSERT INTO Misure (Id Misura, Descrizione Misura, UM) VALUES (6464,"=
Velocit=E0 vento Massima 15 min","m/s")
> > ...
> >=20
> > Do you see a chance to maintain the character encoding somehow or
> > map to ASCII? I am not sure if SQL is limited or permits accented
> > characters.
> >=20
> > My intention is to convert the MDB into a PostgreSQL database.
> >=20
> > Thanks
> >=20
> >  Markus
> >=20
> >=20
> > -------------------------------------------------------
> > SF.Net email is Sponsored by the Better Software Conference & EXPO
> > September 19-22, 2005 * San Francisco, CA * Development Lifecycle Pra=
ctices
> > Agile & Plan-Driven Development * Managing Projects & Teams * Testing=
 & QA
> > Security * Process Improvement & Measurement * http://www.sqe.com/bsc=
e5sf
> > _______________________________________________
> > mdbtools-dev mailing list
> > mdb...@li...
> > https://lists.sourceforge.net/lists/listinfo/mdbtools-dev
> >
>=20
>=20
> -------------------------------------------------------
> SF.Net email is Sponsored by the Better Software Conference & EXPO
> September 19-22, 2005 * San Francisco, CA * Development Lifecycle Pract=
ices
> Agile & Plan-Driven Development * Managing Projects & Teams * Testing &=
 QA
> Security * Process Improvement & Measurement * http://www.sqe.com/bsce5=
sf
> _______________________________________________
> mdbtools-dev mailing list
> mdb...@li...
> https://lists.sourceforge.net/lists/listinfo/mdbtools-dev

--=20
Markus Neteler     <neteler itc it>       http://mpa.itc.it
ITC-irst -  Centro per la Ricerca Scientifica e Tecnologica
MPBA - Predictive Models for Biol. & Environ. Data Analysis
Via Sommarive, 18        -       38050 Povo (Trento), Italy

[mdb-dev] Jet 4 segmentation fault with mdb-sql

[Ron P. <ro...@sa...>]

Hi,
I have a Jet 4 data file that causes the tool mdb-sql to segmentation =
fault depending on which tables are accessed. The crash is repeatable =
and only with the longer tables. Is this something that someone would be =
interested in checking? If so, I could send the file.
Thanks,
Ron Platt

Re: [mdb-dev] Re: problems with JET4 and nonASCI characters

[Brian B. <bri...@gm...>]

The source/target character sets can be selected using the MDB_ICONV
environment variable, otherwise the default is UTF-8 (this all applies
to jet4, access 97 is a different beast).  So what you're likely
seeing is UTF-8 characters on a terminal/program which is unequiped to
handle it.  Just set the target set like:

export MDB_ICONV=3DISO-8859-1

Which will work for most western european languages.  use "iconv
--list" to see the supported values.

Eventually it would be nice to grab this info from LC_ALL or LANG, but
the string returned by those doesn't always map exactly to that
required by iconv.

Brian

2005/8/5, Markus Neteler <ne...@it...>:
> Hi again,
>=20
> after the fix for the 64bit platform I can now export the data.
> But I have the same problem as reported in
>  http://sourceforge.net/mailarchive/forum.php?thread_id=3D7699603&forum_i=
d=3D5183
>=20
> mdb-ver StazioniAutomatiche.mdb
> JET4
>=20
> mdb-export -I StazioniAutomatiche.mdb Misure
> ...
> INSERT INTO Misure (Id Misura, Descrizione Misura, UM) VALUES (174,"Umidi=
t=E0 Massima 1440 min","%")
> ...
> INSERT INTO Misure (Id Misura, Descrizione Misura, UM) VALUES (6464,"Velo=
cit=E0 vento Massima 15 min","m/s")
> ...
>=20
> Do you see a chance to maintain the character encoding somehow or
> map to ASCII? I am not sure if SQL is limited or permits accented
> characters.
>=20
> My intention is to convert the MDB into a PostgreSQL database.
>=20
> Thanks
>=20
>  Markus
>=20
>=20
> -------------------------------------------------------
> SF.Net email is Sponsored by the Better Software Conference & EXPO
> September 19-22, 2005 * San Francisco, CA * Development Lifecycle Practic=
es
> Agile & Plan-Driven Development * Managing Projects & Teams * Testing & Q=
A
> Security * Process Improvement & Measurement * http://www.sqe.com/bsce5sf
> _______________________________________________
> mdbtools-dev mailing list
> mdb...@li...
> https://lists.sourceforge.net/lists/listinfo/mdbtools-dev
>

[mdb-dev] Re: problems with JET4 and nonASCI characters

[Markus N. <ne...@it...>]

Hi again,

after the fix for the 64bit platform I can now export the data.
But I have the same problem as reported in
 http://sourceforge.net/mailarchive/forum.php?thread_id=3D7699603&forum_i=
d=3D5183

mdb-ver StazioniAutomatiche.mdb
JET4

mdb-export -I StazioniAutomatiche.mdb Misure
...
INSERT INTO Misure (Id Misura, Descrizione Misura, UM) VALUES (174,"Umidi=
t=C3=A0 Massima 1440 min","%")
...
INSERT INTO Misure (Id Misura, Descrizione Misura, UM) VALUES (6464,"Velo=
cit=C3=A0 vento Massima 15 min","m/s")
...

Do you see a chance to maintain the character encoding somehow or
map to ASCII? I am not sure if SQL is limited or permits accented
characters.

My intention is to convert the MDB into a PostgreSQL database.

Thanks

 Markus

Re: [mdb-dev] Re: crash in mdb-export (64bit Linux)

[Markus N. <ne...@it...>]

Brian,

cool - works. Ready for CVS :-)

BTW: I am working in the GRASS GIS project where we('ll) have
the same problem. Now I know what to look for. Also if
another mdb tool crashes.

Thanks

 Markus

On Fri, Aug 05, 2005 at 10:54:19AM -0400, Brian Bruns wrote:
> Ok, think i may see the problem.  in src/libmdb/iconv.c in the
> function mdb_unicode2ascii(), iconv, according to the man page, is
> supposed to be passed pointers to size_t but this function is using
> "unsigned int" which obviously changes from 32 -> 64 bit.  Try
> changing:
>=20
>         char *tmp =3D NULL;
>         unsigned int tlen =3D 0;
>         unsigned int len_in, len_out;
>         char *in_ptr, *out_ptr;
>=20
> to=20
>         char *tmp =3D NULL;
>         size_t tlen =3D 0;
>         size_t len_in, len_out;
>         char *in_ptr, *out_ptr;
>=20
> at the top of the function and see how we do.
>=20
> Brian
>=20
> 2005/8/5, Markus Neteler <ne...@it...>:
> > On Thu, Aug 04, 2005 at 03:51:45PM +0200, Carsten Eggers wrote:
> > > Hi Markus,
> > >
> > > Somedays ago Brian Bruns wrote:
> > > >I'd be interested in fixing the CVS version.  I don't have access =
to a
> > > >solaris box though.  I could try it on AIX and perhaps find
> > > >something...do you have a sample database which demonstrates the
> > > >problem?  The other possibility is that it is a 32bit/64bit issue.
> > > >I'm not sure it is 64bit clean, can you set the compiler flags to
> > >      ^^^            ^^^^^^^^^^^
> > > >force 32bit mode? (not sure how to do it on sparc, sorry).
> > >
> > > This maybe also releated to you problem.
> >=20
> >=20
> > It looks like that. So I would appreciate to either know
> > how to compile on a 64bit platform in 32bit mode (maybe its
> > trivial, but I just don't know yet). Or I can offer some
> > time to debug the problem which depends on some indications
> > where to look for the problem (as I don't know the internals
> > of the mdb-tools).
> >=20
> > Thanks
> >=20
> >  Markus
> >=20
> > >
> > >  -- Carsten
> > >
> > >
> > > ----------
> > > Dear list,
> > >
> > > sorry to ask again... do you have any suggestion to
> > > get mdb-export running on a 64bit box?
> > >
> > > Thanks
> > >
> > >  Markus
> > >
> > >
> > > On Fri, Jul 15, 2005 at 05:04:23PM +0200, Markus Neteler wrote:
> > > > Hi,
> > > >
> > > > I have compiled the today's CVS version of the mdbtools
> > > > on a 64bit Linux machine. Unfortunately the following
> > > > command crashes:
> > > >
> > > > mdb-export -I StazioniAutomatiche.mdb Misure
> > > > Segmentation fault
> > > >
> > > > while it works well on a 32bit Linux machine. Both distros
> > > > are Redhat Enterprise 4 (RHEL4).
> > > >
> > > > A bit of debugging (I cannot sent the file since it is 1.4GB):
> > > >
> > > > strace mdb-export -I StazioniAutomatiche.mdb Misure
> > > > ...
> > > > read(3, "\1\1\326\16\0\0\0\0\0\0\0\0\4\0\273\17v\0171\17\354\16".=
.., 4096)
> > > =3D 4096
> > > > mmap(NULL, 200704, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOU=
S, -1, 0)
> > > =3D 0x2a95699000
> > > > mmap(NULL, 200704, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOU=
S, -1, 0)
> > > =3D 0x2a956ca000
> > > > mmap(NULL, 200704, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOU=
S, -1, 0)
> > > =3D 0x2a956fb000
> > > > fstat(3, {st_mode=3DS_IFREG|0644, st_size=3D1479475200, ...}) =3D=
 0
> > > > lseek(3, 73728, SEEK_SET)               =3D 73728
> > > > read(3, "\1\1V\6\20\0\0\0\0\0\0\0,\0\313\17\241\17e\0175\17\376".=
.., 4096)
> > > =3D 4096
> > > > --- SIGSEGV (Segmentation fault) @ 0 (0) ---
> > > > +++ killed by SIGSEGV +++
> > > >
> > > >
> > > > gdb `which mdb-export`
> > > > r -I StazioniAutomatiche.mdb Misure
> > > > Program received signal SIGSEGV, Segmentation fault.
> > > > 0x000000345dc227cc in __gconv_transform_ucs2_internal () from
> > > /lib64/tls/libc.so.6
> > > > bt
> > > > #0  0x000000345dc227cc in __gconv_transform_ucs2_internal () from
> > > /lib64/tls/libc.so.6
> > > > #1  0x000000345dc1d538 in __gconv () from /lib64/tls/libc.so.6
> > > > #2  0x000000345dc1cb9b in iconv () from /lib64/tls/libc.so.6
> > > > #3  0x0000002a955663dc in mdb_unicode2ascii (mdb=3D0x502600, src=3D=
0x503607
> > > "=B0", slen=3D0,
> > > >     dest=3D0x5199d0 "Temperatura aria Istantanea 60 min", dlen=3D=
16384) at
> > > iconv.c:73
> > > > #4  0x0000002a9555ff47 in mdb_col_to_string (mdb=3D0x502600, buf=3D=
0x502614,
> > > start=3D4047, datatype=3D10,
> > > >     size=3D36) at data.c:848
> > > > #5  0x0000002a9555e83b in mdb_xfer_bound_data (mdb=3D0x502600, st=
art=3D4047,
> > > col=3D0x515690, len=3D36)
> > > >     at data.c:222
> > > > #6  0x0000002a9555eadd in _mdb_attempt_bind (mdb=3D0x502600, col=3D=
0x515690,
> > > isnull=3D0 '\0',
> > > >     offset=3D4047, len=3D36) at data.c:306
> > > > #7  0x0000002a9555ea2f in mdb_read_row (table=3D0x515210, row=3D0=
) at
> > > data.c:284
> > > > #8  0x0000002a9555ee6e in mdb_fetch_row (table=3D0x515210) at dat=
a.c:402
> > > > #9  0x000000000040121f in main (argc=3D4, argv=3D0x7fbffff808) at
> > > mdb-export.c:157
> > > >
> > > > (gdb) bt full
> > > > ...
> > > >     is_fixed =3D 53 '5', colnum =3D 0, offset =3D 4196064}, {valu=
e =3D 0x515610,
> > > siz =3D 1569807600,
> > > >     start =3D 52, is_null =3D 48 '0', is_fixed =3D 246 '=F6', col=
num =3D 127, offset
> > > =3D -1073744400}, {
> > > >     value =3D 0x93635d7, siz =3D 4196064, start =3D 0, is_null =3D=
 0 '\0',
> > > is_fixed =3D 0 '\0', colnum =3D 0,
> > > >     offset =3D 1568700462}, {value =3D 0x0, siz =3D 1, start =3D =
0, is_null =3D 0
> > > '\0', is_fixed =3D 0 '\0',
> > > >     colnum =3D 0, offset =3D -1789540033}, {value =3D 0x1000, siz=
 =3D 73728, start
> > > =3D 0,
> > > >     is_null =3D 2 '\002', is_fixed =3D 9 '\t', colnum =3D 0, offs=
et =3D 13795360},
> > > {value =3D 0x1,
> > > >     siz =3D 33188, start =3D 1338, is_null =3D 204 '=CC', is_fixe=
d =3D 0 '\0',
> > > colnum =3D 0, offset =3D 0}, {
> > > >     value =3D 0x582f0000, siz =3D 4096, start =3D 0, is_null =3D =
144 '\220',
> > > is_fixed =3D 34 '"',
> > > >     colnum =3D 0, offset =3D 1121438862}, {value =3D 0x325bbbe9, =
siz =3D
> > > 1111569942, start =3D 0,
> > > >     is_null =3D 0 '\0', is_fixed =3D 0 '\0', colnum =3D 0, offset=
 =3D 1121152771},
> > > {value =3D 0x0, siz =3D 0,
> > > >     start =3D 0, is_null =3D 0 '\0', is_fixed =3D 0 '\0', colnum =
=3D 0, offset =3D
> > > 0}, {value =3D 0x1000,
> > > >     siz =3D 18, start =3D 0, is_null =3D 20 '\024', is_fixed =3D =
38 '&', colnum =3D
> > > 0, offset =3D 5252608}, {
> > > >     value =3D 0x7fbffff630, siz =3D -1789540359, start =3D 42, is=
_null =3D 0 '\0',
> > > is_fixed =3D 16 '\020',
> > > >     colnum =3D 0, offset =3D 4096}}
> > > >         num_fields =3D 3
> > > > #8  0x0000002a9555ee6e in mdb_fetch_row (table=3D0x515210) at dat=
a.c:402
> > > >         mdb =3D (MdbHandle *) 0x502600
> > > >         fmt =3D (MdbFormatConstants *) 0x2a9566a780
> > > >         rows =3D 44
> > > >         rc =3D 5329424
> > > >         pg =3D 0
> > > > #9  0x000000000040121f in main (argc=3D4, argv=3D0x7fbffff808) at
> > > mdb-export.c:157
> > > >         j =3D 3
> > > >         mdb =3D (MdbHandle *) 0x502600
> > > >         table =3D (MdbTableDef *) 0x515210
> > > >         col =3D (MdbColumn *) 0x0
> > > >         bound_values =3D (char **) 0x517290
> > > >         bound_lens =3D (int *) 0x5159b0
> > > >         delimiter =3D 0x502010 ","
> > > >         row_delimiter =3D 0x502030 "\n"
> > > >         header_row =3D 0 '\0'
> > > >         quote_text =3D 1 '\001'
> > > >         insert_statements =3D 1 '\001'
> > > >         sanitize =3D 0 '\0'
> > > >         opt =3D -1
> > > >
> > > >
> > > > Using 'ddd' it dies in
> > > > mdb-export.c:157
> > > >  while(mdb_fetch_row(table)) {
> > > >
> > > > Do you have any ideas? I'm not sure how to continue
> > > > debugging... The bug seems to be in
> > > > mdbtools/src/libmdb/data.c
> > > > Maybe related to guint32? (sorry, just guessing).
> > > >
> > > > Thanks anyway for maintaining the mdbtools!
> > > >
> > > > Markus
> > > >
> > > > --
> > > > Markus Neteler     <neteler itc it>       http://mpa.itc.it
> > > > ITC-irst -  Centro per la Ricerca Scientifica e Tecnologica
> > > > MPBA - Predictive Models for Biol. & Environ. Data Analysis
> > > > Via Sommarive, 18        -       38050 Povo (Trento), Italy
> > >
> > >
> > > -------------------------------------------------------
> > > SF.Net email is Sponsored by the Better Software Conference & EXPO
> > > September 19-22, 2005 * San Francisco, CA * Development Lifecycle P=
ractices
> > > Agile & Plan-Driven Development * Managing Projects & Teams * Testi=
ng & QA
> > > Security * Process Improvement & Measurement * http://www.sqe.com/b=
sce5sf
> > > _______________________________________________
> > > mdbtools-dev mailing list
> > > mdb...@li...
> > > https://lists.sourceforge.net/lists/listinfo/mdbtools-dev
> > >
> > >
> > >
> > >
> > >
> > > -------------------------------------------------------
> > > SF.Net email is Sponsored by the Better Software Conference & EXPO
> > > September 19-22, 2005 * San Francisco, CA * Development Lifecycle P=
ractices
> > > Agile & Plan-Driven Development * Managing Projects & Teams * Testi=
ng & QA
> > > Security * Process Improvement & Measurement * http://www.sqe.com/b=
sce5sf
> > > _______________________________________________
> > > mdbtools-dev mailing list
> > > mdb...@li...
> > > https://lists.sourceforge.net/lists/listinfo/mdbtools-dev
> >=20
> >=20
> > -------------------------------------------------------
> > SF.Net email is Sponsored by the Better Software Conference & EXPO
> > September 19-22, 2005 * San Francisco, CA * Development Lifecycle Pra=
ctices
> > Agile & Plan-Driven Development * Managing Projects & Teams * Testing=
 & QA
> > Security * Process Improvement & Measurement * http://www.sqe.com/bsc=
e5sf
> > _______________________________________________
> > mdbtools-dev mailing list
> > mdb...@li...
> > https://lists.sourceforge.net/lists/listinfo/mdbtools-dev
> >

--=20
Markus Neteler     <neteler itc it>       http://mpa.itc.it
ITC-irst -  Centro per la Ricerca Scientifica e Tecnologica
MPBA - Predictive Models for Biol. & Environ. Data Analysis
Via Sommarive, 18        -       38050 Povo (Trento), Italy

Re: [mdb-dev] Re: crash in mdb-export (64bit Linux)

[Brian B. <bri...@gm...>]

Ok, think i may see the problem.  in src/libmdb/iconv.c in the
function mdb_unicode2ascii(), iconv, according to the man page, is
supposed to be passed pointers to size_t but this function is using
"unsigned int" which obviously changes from 32 -> 64 bit.  Try
changing:

       char *tmp =3D NULL;
       unsigned int tlen =3D 0;
       unsigned int len_in, len_out;
       char *in_ptr, *out_ptr;

to
       char *tmp =3D NULL;
       size_t tlen =3D 0;
       size_t len_in, len_out;
       char *in_ptr, *out_ptr;

at the top of the function and see how we do.

Brian

2005/8/5, Markus Neteler <ne...@it...>:
> On Thu, Aug 04, 2005 at 03:51:45PM +0200, Carsten Eggers wrote:
> > Hi Markus,
> >
> > Somedays ago Brian Bruns wrote:
> > >I'd be interested in fixing the CVS version.  I don't have access to a
> > >solaris box though.  I could try it on AIX and perhaps find
> > >something...do you have a sample database which demonstrates the
> > >problem?  The other possibility is that it is a 32bit/64bit issue.
> > >I'm not sure it is 64bit clean, can you set the compiler flags to
> >      ^^^            ^^^^^^^^^^^
> > >force 32bit mode? (not sure how to do it on sparc, sorry).
> >
> > This maybe also releated to you problem.
>=20
>=20
> It looks like that. So I would appreciate to either know
> how to compile on a 64bit platform in 32bit mode (maybe its
> trivial, but I just don't know yet). Or I can offer some
> time to debug the problem which depends on some indications
> where to look for the problem (as I don't know the internals
> of the mdb-tools).
>=20
> Thanks
>=20
>  Markus
>=20
> >
> >  -- Carsten
> >
> >
> > ----------
> > Dear list,
> >
> > sorry to ask again... do you have any suggestion to
> > get mdb-export running on a 64bit box?
> >
> > Thanks
> >
> >  Markus
> >
> >
> > On Fri, Jul 15, 2005 at 05:04:23PM +0200, Markus Neteler wrote:
> > > Hi,
> > >
> > > I have compiled the today's CVS version of the mdbtools
> > > on a 64bit Linux machine. Unfortunately the following
> > > command crashes:
> > >
> > > mdb-export -I StazioniAutomatiche.mdb Misure
> > > Segmentation fault
> > >
> > > while it works well on a 32bit Linux machine. Both distros
> > > are Redhat Enterprise 4 (RHEL4).
> > >
> > > A bit of debugging (I cannot sent the file since it is 1.4GB):
> > >
> > > strace mdb-export -I StazioniAutomatiche.mdb Misure
> > > ...
> > > read(3, "\1\1\326\16\0\0\0\0\0\0\0\0\4\0\273\17v\0171\17\354\16"..., =
4096)
> > =3D 4096
> > > mmap(NULL, 200704, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -=
1, 0)
> > =3D 0x2a95699000
> > > mmap(NULL, 200704, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -=
1, 0)
> > =3D 0x2a956ca000
> > > mmap(NULL, 200704, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -=
1, 0)
> > =3D 0x2a956fb000
> > > fstat(3, {st_mode=3DS_IFREG|0644, st_size=3D1479475200, ...}) =3D 0
> > > lseek(3, 73728, SEEK_SET)               =3D 73728
> > > read(3, "\1\1V\6\20\0\0\0\0\0\0\0,\0\313\17\241\17e\0175\17\376"..., =
4096)
> > =3D 4096
> > > --- SIGSEGV (Segmentation fault) @ 0 (0) ---
> > > +++ killed by SIGSEGV +++
> > >
> > >
> > > gdb `which mdb-export`
> > > r -I StazioniAutomatiche.mdb Misure
> > > Program received signal SIGSEGV, Segmentation fault.
> > > 0x000000345dc227cc in __gconv_transform_ucs2_internal () from
> > /lib64/tls/libc.so.6
> > > bt
> > > #0  0x000000345dc227cc in __gconv_transform_ucs2_internal () from
> > /lib64/tls/libc.so.6
> > > #1  0x000000345dc1d538 in __gconv () from /lib64/tls/libc.so.6
> > > #2  0x000000345dc1cb9b in iconv () from /lib64/tls/libc.so.6
> > > #3  0x0000002a955663dc in mdb_unicode2ascii (mdb=3D0x502600, src=3D0x=
503607
> > "=B0", slen=3D0,
> > >     dest=3D0x5199d0 "Temperatura aria Istantanea 60 min", dlen=3D1638=
4) at
> > iconv.c:73
> > > #4  0x0000002a9555ff47 in mdb_col_to_string (mdb=3D0x502600, buf=3D0x=
502614,
> > start=3D4047, datatype=3D10,
> > >     size=3D36) at data.c:848
> > > #5  0x0000002a9555e83b in mdb_xfer_bound_data (mdb=3D0x502600, start=
=3D4047,
> > col=3D0x515690, len=3D36)
> > >     at data.c:222
> > > #6  0x0000002a9555eadd in _mdb_attempt_bind (mdb=3D0x502600, col=3D0x=
515690,
> > isnull=3D0 '\0',
> > >     offset=3D4047, len=3D36) at data.c:306
> > > #7  0x0000002a9555ea2f in mdb_read_row (table=3D0x515210, row=3D0) at
> > data.c:284
> > > #8  0x0000002a9555ee6e in mdb_fetch_row (table=3D0x515210) at data.c:=
402
> > > #9  0x000000000040121f in main (argc=3D4, argv=3D0x7fbffff808) at
> > mdb-export.c:157
> > >
> > > (gdb) bt full
> > > ...
> > >     is_fixed =3D 53 '5', colnum =3D 0, offset =3D 4196064}, {value =
=3D 0x515610,
> > siz =3D 1569807600,
> > >     start =3D 52, is_null =3D 48 '0', is_fixed =3D 246 '=F6', colnum =
=3D 127, offset
> > =3D -1073744400}, {
> > >     value =3D 0x93635d7, siz =3D 4196064, start =3D 0, is_null =3D 0 =
'\0',
> > is_fixed =3D 0 '\0', colnum =3D 0,
> > >     offset =3D 1568700462}, {value =3D 0x0, siz =3D 1, start =3D 0, i=
s_null =3D 0
> > '\0', is_fixed =3D 0 '\0',
> > >     colnum =3D 0, offset =3D -1789540033}, {value =3D 0x1000, siz =3D=
 73728, start
> > =3D 0,
> > >     is_null =3D 2 '\002', is_fixed =3D 9 '\t', colnum =3D 0, offset =
=3D 13795360},
> > {value =3D 0x1,
> > >     siz =3D 33188, start =3D 1338, is_null =3D 204 '=CC', is_fixed =
=3D 0 '\0',
> > colnum =3D 0, offset =3D 0}, {
> > >     value =3D 0x582f0000, siz =3D 4096, start =3D 0, is_null =3D 144 =
'\220',
> > is_fixed =3D 34 '"',
> > >     colnum =3D 0, offset =3D 1121438862}, {value =3D 0x325bbbe9, siz =
=3D
> > 1111569942, start =3D 0,
> > >     is_null =3D 0 '\0', is_fixed =3D 0 '\0', colnum =3D 0, offset =3D=
 1121152771},
> > {value =3D 0x0, siz =3D 0,
> > >     start =3D 0, is_null =3D 0 '\0', is_fixed =3D 0 '\0', colnum =3D =
0, offset =3D
> > 0}, {value =3D 0x1000,
> > >     siz =3D 18, start =3D 0, is_null =3D 20 '\024', is_fixed =3D 38 '=
&', colnum =3D
> > 0, offset =3D 5252608}, {
> > >     value =3D 0x7fbffff630, siz =3D -1789540359, start =3D 42, is_nul=
l =3D 0 '\0',
> > is_fixed =3D 16 '\020',
> > >     colnum =3D 0, offset =3D 4096}}
> > >         num_fields =3D 3
> > > #8  0x0000002a9555ee6e in mdb_fetch_row (table=3D0x515210) at data.c:=
402
> > >         mdb =3D (MdbHandle *) 0x502600
> > >         fmt =3D (MdbFormatConstants *) 0x2a9566a780
> > >         rows =3D 44
> > >         rc =3D 5329424
> > >         pg =3D 0
> > > #9  0x000000000040121f in main (argc=3D4, argv=3D0x7fbffff808) at
> > mdb-export.c:157
> > >         j =3D 3
> > >         mdb =3D (MdbHandle *) 0x502600
> > >         table =3D (MdbTableDef *) 0x515210
> > >         col =3D (MdbColumn *) 0x0
> > >         bound_values =3D (char **) 0x517290
> > >         bound_lens =3D (int *) 0x5159b0
> > >         delimiter =3D 0x502010 ","
> > >         row_delimiter =3D 0x502030 "\n"
> > >         header_row =3D 0 '\0'
> > >         quote_text =3D 1 '\001'
> > >         insert_statements =3D 1 '\001'
> > >         sanitize =3D 0 '\0'
> > >         opt =3D -1
> > >
> > >
> > > Using 'ddd' it dies in
> > > mdb-export.c:157
> > >  while(mdb_fetch_row(table)) {
> > >
> > > Do you have any ideas? I'm not sure how to continue
> > > debugging... The bug seems to be in
> > > mdbtools/src/libmdb/data.c
> > > Maybe related to guint32? (sorry, just guessing).
> > >
> > > Thanks anyway for maintaining the mdbtools!
> > >
> > > Markus
> > >
> > > --
> > > Markus Neteler     <neteler itc it>       http://mpa.itc.it
> > > ITC-irst -  Centro per la Ricerca Scientifica e Tecnologica
> > > MPBA - Predictive Models for Biol. & Environ. Data Analysis
> > > Via Sommarive, 18        -       38050 Povo (Trento), Italy
> >
> >
> > -------------------------------------------------------
> > SF.Net email is Sponsored by the Better Software Conference & EXPO
> > September 19-22, 2005 * San Francisco, CA * Development Lifecycle Pract=
ices
> > Agile & Plan-Driven Development * Managing Projects & Teams * Testing &=
 QA
> > Security * Process Improvement & Measurement * http://www.sqe.com/bsce5=
sf
> > _______________________________________________
> > mdbtools-dev mailing list
> > mdb...@li...
> > https://lists.sourceforge.net/lists/listinfo/mdbtools-dev
> >
> >
> >
> >
> >
> > -------------------------------------------------------
> > SF.Net email is Sponsored by the Better Software Conference & EXPO
> > September 19-22, 2005 * San Francisco, CA * Development Lifecycle Pract=
ices
> > Agile & Plan-Driven Development * Managing Projects & Teams * Testing &=
 QA
> > Security * Process Improvement & Measurement * http://www.sqe.com/bsce5=
sf
> > _______________________________________________
> > mdbtools-dev mailing list
> > mdb...@li...
> > https://lists.sourceforge.net/lists/listinfo/mdbtools-dev
>=20
>=20
> -------------------------------------------------------
> SF.Net email is Sponsored by the Better Software Conference & EXPO
> September 19-22, 2005 * San Francisco, CA * Development Lifecycle Practic=
es
> Agile & Plan-Driven Development * Managing Projects & Teams * Testing & Q=
A
> Security * Process Improvement & Measurement * http://www.sqe.com/bsce5sf
> _______________________________________________
> mdbtools-dev mailing list
> mdb...@li...
> https://lists.sourceforge.net/lists/listinfo/mdbtools-dev
>

Re: [mdb-dev] Re: crash in mdb-export (64bit Linux)

[Markus N. <ne...@it...>]

On Thu, Aug 04, 2005 at 03:51:45PM +0200, Carsten Eggers wrote:
> Hi Markus,
>=20
> Somedays ago Brian Bruns wrote:
> >I'd be interested in fixing the CVS version.  I don't have access to a
> >solaris box though.  I could try it on AIX and perhaps find
> >something...do you have a sample database which demonstrates the
> >problem?  The other possibility is that it is a 32bit/64bit issue.
> >I'm not sure it is 64bit clean, can you set the compiler flags to
>      ^^^            ^^^^^^^^^^^
> >force 32bit mode? (not sure how to do it on sparc, sorry).
>=20
> This maybe also releated to you problem.


It looks like that. So I would appreciate to either know
how to compile on a 64bit platform in 32bit mode (maybe its
trivial, but I just don't know yet). Or I can offer some
time to debug the problem which depends on some indications
where to look for the problem (as I don't know the internals
of the mdb-tools).

Thanks=20

 Markus
=20
>=20
>  -- Carsten
>=20
>=20
> ----------
> Dear list,
>=20
> sorry to ask again... do you have any suggestion to
> get mdb-export running on a 64bit box?
>=20
> Thanks
>=20
>  Markus
>=20
>=20
> On Fri, Jul 15, 2005 at 05:04:23PM +0200, Markus Neteler wrote:
> > Hi,
> >
> > I have compiled the today's CVS version of the mdbtools
> > on a 64bit Linux machine. Unfortunately the following
> > command crashes:
> >
> > mdb-export -I StazioniAutomatiche.mdb Misure
> > Segmentation fault
> >
> > while it works well on a 32bit Linux machine. Both distros
> > are Redhat Enterprise 4 (RHEL4).
> >
> > A bit of debugging (I cannot sent the file since it is 1.4GB):
> >
> > strace mdb-export -I StazioniAutomatiche.mdb Misure
> > ...
> > read(3, "\1\1\326\16\0\0\0\0\0\0\0\0\4\0\273\17v\0171\17\354\16"..., =
4096)
> =3D 4096
> > mmap(NULL, 200704, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -=
1, 0)
> =3D 0x2a95699000
> > mmap(NULL, 200704, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -=
1, 0)
> =3D 0x2a956ca000
> > mmap(NULL, 200704, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -=
1, 0)
> =3D 0x2a956fb000
> > fstat(3, {st_mode=3DS_IFREG|0644, st_size=3D1479475200, ...}) =3D 0
> > lseek(3, 73728, SEEK_SET)               =3D 73728
> > read(3, "\1\1V\6\20\0\0\0\0\0\0\0,\0\313\17\241\17e\0175\17\376"..., =
4096)
> =3D 4096
> > --- SIGSEGV (Segmentation fault) @ 0 (0) ---
> > +++ killed by SIGSEGV +++
> >
> >
> > gdb `which mdb-export`
> > r -I StazioniAutomatiche.mdb Misure
> > Program received signal SIGSEGV, Segmentation fault.
> > 0x000000345dc227cc in __gconv_transform_ucs2_internal () from
> /lib64/tls/libc.so.6
> > bt
> > #0  0x000000345dc227cc in __gconv_transform_ucs2_internal () from
> /lib64/tls/libc.so.6
> > #1  0x000000345dc1d538 in __gconv () from /lib64/tls/libc.so.6
> > #2  0x000000345dc1cb9b in iconv () from /lib64/tls/libc.so.6
> > #3  0x0000002a955663dc in mdb_unicode2ascii (mdb=3D0x502600, src=3D0x=
503607
> "=B0", slen=3D0,
> >     dest=3D0x5199d0 "Temperatura aria Istantanea 60 min", dlen=3D1638=
4) at
> iconv.c:73
> > #4  0x0000002a9555ff47 in mdb_col_to_string (mdb=3D0x502600, buf=3D0x=
502614,
> start=3D4047, datatype=3D10,
> >     size=3D36) at data.c:848
> > #5  0x0000002a9555e83b in mdb_xfer_bound_data (mdb=3D0x502600, start=3D=
4047,
> col=3D0x515690, len=3D36)
> >     at data.c:222
> > #6  0x0000002a9555eadd in _mdb_attempt_bind (mdb=3D0x502600, col=3D0x=
515690,
> isnull=3D0 '\0',
> >     offset=3D4047, len=3D36) at data.c:306
> > #7  0x0000002a9555ea2f in mdb_read_row (table=3D0x515210, row=3D0) at
> data.c:284
> > #8  0x0000002a9555ee6e in mdb_fetch_row (table=3D0x515210) at data.c:=
402
> > #9  0x000000000040121f in main (argc=3D4, argv=3D0x7fbffff808) at
> mdb-export.c:157
> >
> > (gdb) bt full
> > ...
> >     is_fixed =3D 53 '5', colnum =3D 0, offset =3D 4196064}, {value =3D=
 0x515610,
> siz =3D 1569807600,
> >     start =3D 52, is_null =3D 48 '0', is_fixed =3D 246 '=F6', colnum =
=3D 127, offset
> =3D -1073744400}, {
> >     value =3D 0x93635d7, siz =3D 4196064, start =3D 0, is_null =3D 0 =
'\0',
> is_fixed =3D 0 '\0', colnum =3D 0,
> >     offset =3D 1568700462}, {value =3D 0x0, siz =3D 1, start =3D 0, i=
s_null =3D 0
> '\0', is_fixed =3D 0 '\0',
> >     colnum =3D 0, offset =3D -1789540033}, {value =3D 0x1000, siz =3D=
 73728, start
> =3D 0,
> >     is_null =3D 2 '\002', is_fixed =3D 9 '\t', colnum =3D 0, offset =3D=
 13795360},
> {value =3D 0x1,
> >     siz =3D 33188, start =3D 1338, is_null =3D 204 '=CC', is_fixed =3D=
 0 '\0',
> colnum =3D 0, offset =3D 0}, {
> >     value =3D 0x582f0000, siz =3D 4096, start =3D 0, is_null =3D 144 =
'\220',
> is_fixed =3D 34 '"',
> >     colnum =3D 0, offset =3D 1121438862}, {value =3D 0x325bbbe9, siz =
=3D
> 1111569942, start =3D 0,
> >     is_null =3D 0 '\0', is_fixed =3D 0 '\0', colnum =3D 0, offset =3D=
 1121152771},
> {value =3D 0x0, siz =3D 0,
> >     start =3D 0, is_null =3D 0 '\0', is_fixed =3D 0 '\0', colnum =3D =
0, offset =3D
> 0}, {value =3D 0x1000,
> >     siz =3D 18, start =3D 0, is_null =3D 20 '\024', is_fixed =3D 38 '=
&', colnum =3D
> 0, offset =3D 5252608}, {
> >     value =3D 0x7fbffff630, siz =3D -1789540359, start =3D 42, is_nul=
l =3D 0 '\0',
> is_fixed =3D 16 '\020',
> >     colnum =3D 0, offset =3D 4096}}
> >         num_fields =3D 3
> > #8  0x0000002a9555ee6e in mdb_fetch_row (table=3D0x515210) at data.c:=
402
> >         mdb =3D (MdbHandle *) 0x502600
> >         fmt =3D (MdbFormatConstants *) 0x2a9566a780
> >         rows =3D 44
> >         rc =3D 5329424
> >         pg =3D 0
> > #9  0x000000000040121f in main (argc=3D4, argv=3D0x7fbffff808) at
> mdb-export.c:157
> >         j =3D 3
> >         mdb =3D (MdbHandle *) 0x502600
> >         table =3D (MdbTableDef *) 0x515210
> >         col =3D (MdbColumn *) 0x0
> >         bound_values =3D (char **) 0x517290
> >         bound_lens =3D (int *) 0x5159b0
> >         delimiter =3D 0x502010 ","
> >         row_delimiter =3D 0x502030 "\n"
> >         header_row =3D 0 '\0'
> >         quote_text =3D 1 '\001'
> >         insert_statements =3D 1 '\001'
> >         sanitize =3D 0 '\0'
> >         opt =3D -1
> >
> >
> > Using 'ddd' it dies in
> > mdb-export.c:157
> >  while(mdb_fetch_row(table)) {
> >
> > Do you have any ideas? I'm not sure how to continue
> > debugging... The bug seems to be in
> > mdbtools/src/libmdb/data.c
> > Maybe related to guint32? (sorry, just guessing).
> >
> > Thanks anyway for maintaining the mdbtools!
> >
> > Markus
> >
> > --
> > Markus Neteler     <neteler itc it>       http://mpa.itc.it
> > ITC-irst -  Centro per la Ricerca Scientifica e Tecnologica
> > MPBA - Predictive Models for Biol. & Environ. Data Analysis
> > Via Sommarive, 18        -       38050 Povo (Trento), Italy
>=20
>=20
> -------------------------------------------------------
> SF.Net email is Sponsored by the Better Software Conference & EXPO
> September 19-22, 2005 * San Francisco, CA * Development Lifecycle Pract=
ices
> Agile & Plan-Driven Development * Managing Projects & Teams * Testing &=
 QA
> Security * Process Improvement & Measurement * http://www.sqe.com/bsce5=
sf
> _______________________________________________
> mdbtools-dev mailing list
> mdb...@li...
> https://lists.sourceforge.net/lists/listinfo/mdbtools-dev
>=20
>=20
>=20
>=20
>=20
> -------------------------------------------------------
> SF.Net email is Sponsored by the Better Software Conference & EXPO
> September 19-22, 2005 * San Francisco, CA * Development Lifecycle Pract=
ices
> Agile & Plan-Driven Development * Managing Projects & Teams * Testing &=
 QA
> Security * Process Improvement & Measurement * http://www.sqe.com/bsce5=
sf
> _______________________________________________
> mdbtools-dev mailing list
> mdb...@li...
> https://lists.sourceforge.net/lists/listinfo/mdbtools-dev

AW: [mdb-dev] Re: crash in mdb-export (64bit Linux)

[Carsten E. <ce...@if...>]

Hi Markus,

Somedays ago Brian Bruns wrote:
>I'd be interested in fixing the CVS version.  I don't have access to a
>solaris box though.  I could try it on AIX and perhaps find
>something...do you have a sample database which demonstrates the
>problem?  The other possibility is that it is a 32bit/64bit issue.
>I'm not sure it is 64bit clean, can you set the compiler flags to
     ^^^            ^^^^^^^^^^^
>force 32bit mode? (not sure how to do it on sparc, sorry).

This maybe also releated to you problem.


 -- Carsten


----------
Dear list,

sorry to ask again... do you have any suggestion to
get mdb-export running on a 64bit box?

Thanks

 Markus


On Fri, Jul 15, 2005 at 05:04:23PM +0200, Markus Neteler wrote:
> Hi,
>
> I have compiled the today's CVS version of the mdbtools
> on a 64bit Linux machine. Unfortunately the following
> command crashes:
>
> mdb-export -I StazioniAutomatiche.mdb Misure
> Segmentation fault
>
> while it works well on a 32bit Linux machine. Both distros
> are Redhat Enterprise 4 (RHEL4).
>
> A bit of debugging (I cannot sent the file since it is 1.4GB):
>
> strace mdb-export -I StazioniAutomatiche.mdb Misure
> ...
> read(3, "\1\1\326\16\0\0\0\0\0\0\0\0\4\0\273\17v\0171\17\354\16"..., 4096)
= 4096
> mmap(NULL, 200704, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0)
= 0x2a95699000
> mmap(NULL, 200704, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0)
= 0x2a956ca000
> mmap(NULL, 200704, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0)
= 0x2a956fb000
> fstat(3, {st_mode=S_IFREG|0644, st_size=1479475200, ...}) = 0
> lseek(3, 73728, SEEK_SET)               = 73728
> read(3, "\1\1V\6\20\0\0\0\0\0\0\0,\0\313\17\241\17e\0175\17\376"..., 4096)
= 4096
> --- SIGSEGV (Segmentation fault) @ 0 (0) ---
> +++ killed by SIGSEGV +++
>
>
> gdb `which mdb-export`
> r -I StazioniAutomatiche.mdb Misure
> Program received signal SIGSEGV, Segmentation fault.
> 0x000000345dc227cc in __gconv_transform_ucs2_internal () from
/lib64/tls/libc.so.6
> bt
> #0  0x000000345dc227cc in __gconv_transform_ucs2_internal () from
/lib64/tls/libc.so.6
> #1  0x000000345dc1d538 in __gconv () from /lib64/tls/libc.so.6
> #2  0x000000345dc1cb9b in iconv () from /lib64/tls/libc.so.6
> #3  0x0000002a955663dc in mdb_unicode2ascii (mdb=0x502600, src=0x503607
"°", slen=0,
>     dest=0x5199d0 "Temperatura aria Istantanea 60 min", dlen=16384) at
iconv.c:73
> #4  0x0000002a9555ff47 in mdb_col_to_string (mdb=0x502600, buf=0x502614,
start=4047, datatype=10,
>     size=36) at data.c:848
> #5  0x0000002a9555e83b in mdb_xfer_bound_data (mdb=0x502600, start=4047,
col=0x515690, len=36)
>     at data.c:222
> #6  0x0000002a9555eadd in _mdb_attempt_bind (mdb=0x502600, col=0x515690,
isnull=0 '\0',
>     offset=4047, len=36) at data.c:306
> #7  0x0000002a9555ea2f in mdb_read_row (table=0x515210, row=0) at
data.c:284
> #8  0x0000002a9555ee6e in mdb_fetch_row (table=0x515210) at data.c:402
> #9  0x000000000040121f in main (argc=4, argv=0x7fbffff808) at
mdb-export.c:157
>
> (gdb) bt full
> ...
>     is_fixed = 53 '5', colnum = 0, offset = 4196064}, {value = 0x515610,
siz = 1569807600,
>     start = 52, is_null = 48 '0', is_fixed = 246 'ö', colnum = 127, offset
= -1073744400}, {
>     value = 0x93635d7, siz = 4196064, start = 0, is_null = 0 '\0',
is_fixed = 0 '\0', colnum = 0,
>     offset = 1568700462}, {value = 0x0, siz = 1, start = 0, is_null = 0
'\0', is_fixed = 0 '\0',
>     colnum = 0, offset = -1789540033}, {value = 0x1000, siz = 73728, start
= 0,
>     is_null = 2 '\002', is_fixed = 9 '\t', colnum = 0, offset = 13795360},
{value = 0x1,
>     siz = 33188, start = 1338, is_null = 204 'Ì', is_fixed = 0 '\0',
colnum = 0, offset = 0}, {
>     value = 0x582f0000, siz = 4096, start = 0, is_null = 144 '\220',
is_fixed = 34 '"',
>     colnum = 0, offset = 1121438862}, {value = 0x325bbbe9, siz =
1111569942, start = 0,
>     is_null = 0 '\0', is_fixed = 0 '\0', colnum = 0, offset = 1121152771},
{value = 0x0, siz = 0,
>     start = 0, is_null = 0 '\0', is_fixed = 0 '\0', colnum = 0, offset =
0}, {value = 0x1000,
>     siz = 18, start = 0, is_null = 20 '\024', is_fixed = 38 '&', colnum =
0, offset = 5252608}, {
>     value = 0x7fbffff630, siz = -1789540359, start = 42, is_null = 0 '\0',
is_fixed = 16 '\020',
>     colnum = 0, offset = 4096}}
>         num_fields = 3
> #8  0x0000002a9555ee6e in mdb_fetch_row (table=0x515210) at data.c:402
>         mdb = (MdbHandle *) 0x502600
>         fmt = (MdbFormatConstants *) 0x2a9566a780
>         rows = 44
>         rc = 5329424
>         pg = 0
> #9  0x000000000040121f in main (argc=4, argv=0x7fbffff808) at
mdb-export.c:157
>         j = 3
>         mdb = (MdbHandle *) 0x502600
>         table = (MdbTableDef *) 0x515210
>         col = (MdbColumn *) 0x0
>         bound_values = (char **) 0x517290
>         bound_lens = (int *) 0x5159b0
>         delimiter = 0x502010 ","
>         row_delimiter = 0x502030 "\n"
>         header_row = 0 '\0'
>         quote_text = 1 '\001'
>         insert_statements = 1 '\001'
>         sanitize = 0 '\0'
>         opt = -1
>
>
> Using 'ddd' it dies in
> mdb-export.c:157
>  while(mdb_fetch_row(table)) {
>
> Do you have any ideas? I'm not sure how to continue
> debugging... The bug seems to be in
> mdbtools/src/libmdb/data.c
> Maybe related to guint32? (sorry, just guessing).
>
> Thanks anyway for maintaining the mdbtools!
>
> Markus
>
> --
> Markus Neteler     <neteler itc it>       http://mpa.itc.it
> ITC-irst -  Centro per la Ricerca Scientifica e Tecnologica
> MPBA - Predictive Models for Biol. & Environ. Data Analysis
> Via Sommarive, 18        -       38050 Povo (Trento), Italy


-------------------------------------------------------
SF.Net email is Sponsored by the Better Software Conference & EXPO
September 19-22, 2005 * San Francisco, CA * Development Lifecycle Practices
Agile & Plan-Driven Development * Managing Projects & Teams * Testing & QA
Security * Process Improvement & Measurement * http://www.sqe.com/bsce5sf
_______________________________________________
mdbtools-dev mailing list
mdb...@li...
https://lists.sourceforge.net/lists/listinfo/mdbtools-dev

[mdb-dev] Re: crash in mdb-export (64bit Linux)

[Markus N. <ne...@it...>]

Dear list,

sorry to ask again... do you have any suggestion to
get mdb-export running on a 64bit box?

Thanks

 Markus


On Fri, Jul 15, 2005 at 05:04:23PM +0200, Markus Neteler wrote:
> Hi,
>=20
> I have compiled the today's CVS version of the mdbtools
> on a 64bit Linux machine. Unfortunately the following
> command crashes:
>=20
> mdb-export -I StazioniAutomatiche.mdb Misure =20
> Segmentation fault
>=20
> while it works well on a 32bit Linux machine. Both distros
> are Redhat Enterprise 4 (RHEL4).
>=20
> A bit of debugging (I cannot sent the file since it is 1.4GB):
>=20
> strace mdb-export -I StazioniAutomatiche.mdb Misure
> ...
> read(3, "\1\1\326\16\0\0\0\0\0\0\0\0\4\0\273\17v\0171\17\354\16"..., 40=
96) =3D 4096
> mmap(NULL, 200704, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1,=
 0) =3D 0x2a95699000
> mmap(NULL, 200704, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1,=
 0) =3D 0x2a956ca000
> mmap(NULL, 200704, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1,=
 0) =3D 0x2a956fb000
> fstat(3, {st_mode=3DS_IFREG|0644, st_size=3D1479475200, ...}) =3D 0
> lseek(3, 73728, SEEK_SET)               =3D 73728
> read(3, "\1\1V\6\20\0\0\0\0\0\0\0,\0\313\17\241\17e\0175\17\376"..., 40=
96) =3D 4096
> --- SIGSEGV (Segmentation fault) @ 0 (0) ---
> +++ killed by SIGSEGV +++
>=20
>=20
> gdb `which mdb-export`
> r -I StazioniAutomatiche.mdb Misure
> Program received signal SIGSEGV, Segmentation fault.
> 0x000000345dc227cc in __gconv_transform_ucs2_internal () from /lib64/tl=
s/libc.so.6
> bt
> #0  0x000000345dc227cc in __gconv_transform_ucs2_internal () from /lib6=
4/tls/libc.so.6
> #1  0x000000345dc1d538 in __gconv () from /lib64/tls/libc.so.6
> #2  0x000000345dc1cb9b in iconv () from /lib64/tls/libc.so.6
> #3  0x0000002a955663dc in mdb_unicode2ascii (mdb=3D0x502600, src=3D0x50=
3607 "=B0", slen=3D0,
>     dest=3D0x5199d0 "Temperatura aria Istantanea 60 min", dlen=3D16384)=
 at iconv.c:73
> #4  0x0000002a9555ff47 in mdb_col_to_string (mdb=3D0x502600, buf=3D0x50=
2614, start=3D4047, datatype=3D10,
>     size=3D36) at data.c:848
> #5  0x0000002a9555e83b in mdb_xfer_bound_data (mdb=3D0x502600, start=3D=
4047, col=3D0x515690, len=3D36)
>     at data.c:222
> #6  0x0000002a9555eadd in _mdb_attempt_bind (mdb=3D0x502600, col=3D0x51=
5690, isnull=3D0 '\0',
>     offset=3D4047, len=3D36) at data.c:306
> #7  0x0000002a9555ea2f in mdb_read_row (table=3D0x515210, row=3D0) at d=
ata.c:284
> #8  0x0000002a9555ee6e in mdb_fetch_row (table=3D0x515210) at data.c:40=
2
> #9  0x000000000040121f in main (argc=3D4, argv=3D0x7fbffff808) at mdb-e=
xport.c:157
>=20
> (gdb) bt full
> ...
>     is_fixed =3D 53 '5', colnum =3D 0, offset =3D 4196064}, {value =3D =
0x515610, siz =3D 1569807600,
>     start =3D 52, is_null =3D 48 '0', is_fixed =3D 246 '=F6', colnum =3D=
 127, offset =3D -1073744400}, {
>     value =3D 0x93635d7, siz =3D 4196064, start =3D 0, is_null =3D 0 '\=
0', is_fixed =3D 0 '\0', colnum =3D 0,
>     offset =3D 1568700462}, {value =3D 0x0, siz =3D 1, start =3D 0, is_=
null =3D 0 '\0', is_fixed =3D 0 '\0',
>     colnum =3D 0, offset =3D -1789540033}, {value =3D 0x1000, siz =3D 7=
3728, start =3D 0,
>     is_null =3D 2 '\002', is_fixed =3D 9 '\t', colnum =3D 0, offset =3D=
 13795360}, {value =3D 0x1,
>     siz =3D 33188, start =3D 1338, is_null =3D 204 '=CC', is_fixed =3D =
0 '\0', colnum =3D 0, offset =3D 0}, {
>     value =3D 0x582f0000, siz =3D 4096, start =3D 0, is_null =3D 144 '\=
220', is_fixed =3D 34 '"',
>     colnum =3D 0, offset =3D 1121438862}, {value =3D 0x325bbbe9, siz =3D=
 1111569942, start =3D 0,
>     is_null =3D 0 '\0', is_fixed =3D 0 '\0', colnum =3D 0, offset =3D 1=
121152771}, {value =3D 0x0, siz =3D 0,
>     start =3D 0, is_null =3D 0 '\0', is_fixed =3D 0 '\0', colnum =3D 0,=
 offset =3D 0}, {value =3D 0x1000,
>     siz =3D 18, start =3D 0, is_null =3D 20 '\024', is_fixed =3D 38 '&'=
, colnum =3D 0, offset =3D 5252608}, {
>     value =3D 0x7fbffff630, siz =3D -1789540359, start =3D 42, is_null =
=3D 0 '\0', is_fixed =3D 16 '\020',
>     colnum =3D 0, offset =3D 4096}}
>         num_fields =3D 3
> #8  0x0000002a9555ee6e in mdb_fetch_row (table=3D0x515210) at data.c:40=
2
>         mdb =3D (MdbHandle *) 0x502600
>         fmt =3D (MdbFormatConstants *) 0x2a9566a780
>         rows =3D 44
>         rc =3D 5329424
>         pg =3D 0
> #9  0x000000000040121f in main (argc=3D4, argv=3D0x7fbffff808) at mdb-e=
xport.c:157
>         j =3D 3
>         mdb =3D (MdbHandle *) 0x502600
>         table =3D (MdbTableDef *) 0x515210
>         col =3D (MdbColumn *) 0x0
>         bound_values =3D (char **) 0x517290
>         bound_lens =3D (int *) 0x5159b0
>         delimiter =3D 0x502010 ","
>         row_delimiter =3D 0x502030 "\n"
>         header_row =3D 0 '\0'
>         quote_text =3D 1 '\001'
>         insert_statements =3D 1 '\001'
>         sanitize =3D 0 '\0'
>         opt =3D -1
>=20
>=20
> Using 'ddd' it dies in
> mdb-export.c:157
>  while(mdb_fetch_row(table)) {
>=20
> Do you have any ideas? I'm not sure how to continue
> debugging... The bug seems to be in=20
> mdbtools/src/libmdb/data.c
> Maybe related to guint32? (sorry, just guessing).
>=20
> Thanks anyway for maintaining the mdbtools!
>=20
> Markus
>=20
> --=20
> Markus Neteler     <neteler itc it>       http://mpa.itc.it
> ITC-irst -  Centro per la Ricerca Scientifica e Tecnologica
> MPBA - Predictive Models for Biol. & Environ. Data Analysis
> Via Sommarive, 18        -       38050 Povo (Trento), Italy

Re: [mdb-dev] table problem , data wrong type

[Jeff S. <why...@ya...>]

--- "Barry A. Tigner" <ti...@ms...> wrote:
> I have an apparent problem with an access db.
> One of the tables appears to have incorrect
> data type detection.

> I am using the current version(0.5_6) of MDBTools running
> on FreeBSD 5.4 .

While this does appear to be the most recent port compiled
for FreeBSD, it is severely out-of-date (about 4 years from
what I can gather).  Several issues with column ordering
have been fixed in the mean time.

Please try compiling from CVS.  If you have problems with
it, I may be able to be of more help (although I have
no experience with *BSD).

 -- Jeff Smith


		
____________________________________________________
Start your day with Yahoo! - make it your home page 
http://www.yahoo.com/r/hs 
 

[mdb-dev] table problem , data wrong type

[Barry A. T. <ti...@ms...>]

I have an apparent problem with an access db.
One of the tables appears to have incorrect
data type detection.

mdb-sql reports the following description of a table
in my database. The EmployeeID is supposed to be
a 50 character string, but it shows up as a date.
The ExpirationDate is supposed to be a short date,
but shows up as a memo/hyperlink.

The CardNumber shows up as a 2 byte interger, but
the text lengths appear to be 2X the size reported
in MS OFfice 2003 Access. I understand this may be
due to Office using 2 byte character representations ?

I am using the current version(0.5_6) of MDBTools running
on FreeBSD 5.4 .

+------------------------------+--------------------+----------+
|Column Name                   |Type                |Size      |
+------------------------------+--------------------+----------+
|CardNumber                    |Integer             |2         |
|PIN                           |Text                |24        |
|EmployeeID                    |DateTime (Short)    |8         |
|DateAssigned                  |DateTime (Short)    |8         |
|DateReturned                  |DateTime (Short)    |8         |
|ExpirationDate                |Memo/Hyperlink      |0         |
|Notes                         |Text                |100       |
|Email                         |Text                |8         |
+------------------------------+--------------------+----------+

-- 
Barry A. Tigner
Electronics Shop Mgr.
Physics and Astronomy dept.
Michigan State University
ti...@ms...
517-355-9200 x 2538
room 1223 Biomedical Physical Sciences bldg.