barry-devel — Development discussion mailing list

Re: [Barry-devel] New tool: btarcmp (and other news)

[Chris F. <cd...@fo...>]

On Tue, Jan 24, 2012 at 01:23:47PM +0000, Toby Gray wrote:
> QNX gave an error with some of the overloads in FieldValueHandlerBase. 
> The problem appears to be that std::time_t on QNX is just a typedef for 
> std::uint32_t, where as on my linux system time_t is "long int". I can't 
> think of a solution which doesn't:
> a) require templates (which would defeat the point of FieldValueHandlerBase)
> b) require the explicit mention of every possible integer type (which is 
> not nice as it would require fiddling with where time_t is used in 
> places like r_calendar to avoid ambiguous overload calls)
> c) switch to using something like boost::variant, requiring boost to 
> compile Barry (which would mean finding ports of Boost for Android and QNX)
> 
> Do you have any thoughts on any other alternatives?

Good catch!  Thanks.

There's the d) option of creating a struct BarryTimeT which just wraps
time_t.

This might allow for useful functions to be attached to our time type
as well.

What do you think?

- Chris

Re: [Barry-devel] New tool: btarcmp (and other news)

[Nicolas <pr...@fr...>]

Le mardi 24 janvier 2012 à 03:42 -0500, Chris Frey a écrit :
> Hi list,
> 
> I've recently coded and added a new tool, which I needed in other Barry
> development in the Desktop GUI side of things.
> 
> The new tool is called btarcmp, which is intented to be a diff for
> Barry tar backup files.
> 
> Let me know if you find bugs or have suggestions for this tool.
> 
> Also, in writing this tool, I had a need to access all record fields
> in a programmatic way.  For btarcmp, I wanted to compare them, but I could
> see times when just being able to say "do X to all fields in Contact"
> would be very useful.
> 
> And so the FieldHandle<> template system was born.  If you call a record
> class's GetFieldHandles() member (and only if you call it) it will create
> a vector of field handles that describe all the fields.  You can then
> create a functor class which can be called for each field in the list.
> 
> It is documented in the doxygen docs, so I won't go into detail here.
> 
> There is an english description for each of the fields as well.  This
> brings with it the opportunity to have localized language support
> for field descriptions, directly in the library, in the field handles.
> 
> Would this be useful?  Nicolas?  Could you take a look at the docs and
> code and let me know if this is a good spot for translations?
> 
> Also, I'd appreciate if people would grab the latest git tree and compile
> it for their systems, and for their locales, before I release version
> 0.18 (planned for next month).  Nicolas, there's a new string in the
> backup GUI, which could use a French translation.
> 
> Thanks!
> - Chris

I'll try to find time to complete the french translation and do some
tests.

And I hope to find time to fix the last devel about evolution3 opensync
plugin.

Maybe in 15 days...

Regards,

Nicolas

Re: [Barry-devel] New tool: btarcmp (and other news)

[Nicolas <pr...@fr...>]

Le mardi 24 janvier 2012 à 13:23 +0000, Toby Gray a écrit :
> On 24/01/12 08:42, Chris Frey wrote:
> > Also, I'd appreciate if people would grab the latest git tree and compile
> > it for their systems, and for their locales, before I release version
> > 0.18 (planned for next month).  Nicolas, there's a new string in the
> > backup GUI, which could use a French translation.
> 
> I checked out the latest git tree (0bab6882) and it builds successfully 
> for Android. It builds fine on my desktop Linux system (Ubuntu 10.04 
> 64-bit).
> 
> QNX gave an error with some of the overloads in FieldValueHandlerBase. 
> The problem appears to be that std::time_t on QNX is just a typedef for 
> std::uint32_t, where as on my linux system time_t is "long int". I can't 
> think of a solution which doesn't:
> a) require templates (which would defeat the point of FieldValueHandlerBase)
> b) require the explicit mention of every possible integer type (which is 
> not nice as it would require fiddling with where time_t is used in 
> places like r_calendar to avoid ambiguous overload calls)
> c) switch to using something like boost::variant, requiring boost to 
> compile Barry (which would mean finding ports of Boost for Android and QNX)
> 
> Do you have any thoughts on any other alternatives?

No link with Barry.

I'm interested in QNX port. 

Which QNX framework do you use ? pkgsrc ? 

I think to start a new project for my PlayBook device : a DLNA player.
And I study how to port glib2, gupnp-* into the QNX system.

Regards,

Nicolas

Re: [Barry-devel] New tool: btarcmp (and other news)

[Toby G. <tob...@re...>]

On 24/01/12 08:42, Chris Frey wrote:
> Also, I'd appreciate if people would grab the latest git tree and compile
> it for their systems, and for their locales, before I release version
> 0.18 (planned for next month).  Nicolas, there's a new string in the
> backup GUI, which could use a French translation.

I checked out the latest git tree (0bab6882) and it builds successfully 
for Android. It builds fine on my desktop Linux system (Ubuntu 10.04 
64-bit).

QNX gave an error with some of the overloads in FieldValueHandlerBase. 
The problem appears to be that std::time_t on QNX is just a typedef for 
std::uint32_t, where as on my linux system time_t is "long int". I can't 
think of a solution which doesn't:
a) require templates (which would defeat the point of FieldValueHandlerBase)
b) require the explicit mention of every possible integer type (which is 
not nice as it would require fiddling with where time_t is used in 
places like r_calendar to avoid ambiguous overload calls)
c) switch to using something like boost::variant, requiring boost to 
compile Barry (which would mean finding ports of Boost for Android and QNX)

Do you have any thoughts on any other alternatives?

Regards,

Toby

[Barry-devel] New tool: btarcmp (and other news)

[Chris F. <cd...@fo...>]

Hi list,

I've recently coded and added a new tool, which I needed in other Barry
development in the Desktop GUI side of things.

The new tool is called btarcmp, which is intented to be a diff for
Barry tar backup files.

Let me know if you find bugs or have suggestions for this tool.

Also, in writing this tool, I had a need to access all record fields
in a programmatic way.  For btarcmp, I wanted to compare them, but I could
see times when just being able to say "do X to all fields in Contact"
would be very useful.

And so the FieldHandle<> template system was born.  If you call a record
class's GetFieldHandles() member (and only if you call it) it will create
a vector of field handles that describe all the fields.  You can then
create a functor class which can be called for each field in the list.

It is documented in the doxygen docs, so I won't go into detail here.

There is an english description for each of the fields as well.  This
brings with it the opportunity to have localized language support
for field descriptions, directly in the library, in the field handles.

Would this be useful?  Nicolas?  Could you take a look at the docs and
code and let me know if this is a good spot for translations?

Also, I'd appreciate if people would grab the latest git tree and compile
it for their systems, and for their locales, before I release version
0.18 (planned for next month).  Nicolas, there's a new string in the
backup GUI, which could use a French translation.

Thanks!
- Chris

Re: [Barry-devel] A couple of build system patches

[Chris F. <cd...@fo...>]

On Wed, Jan 11, 2012 at 04:19:25PM +0000, Toby Gray wrote:
> Hi,
> 
> I've got a couple of patches for the build system in my github branch at 
> git://github.com/tobygray/barry.git:
> 2c7f443777001cbef0db2f84f9f5e586746cb574 - correctly places the 
> networking library as a library rather than a linker flag, fixing a 
> build issue on QNX
> 00949a863e434a1d0f279bd7115b8e1a90022fb0 - Corrects the extension 
> expected on C++ source files, other the latest android build tools 
> reject it as not a valid extension.

Thanks!  Merged both.

- Chris

[Barry-devel] A couple of build system patches

[Toby G. <tob...@re...>]

Hi,

I've got a couple of patches for the build system in my github branch at 
git://github.com/tobygray/barry.git:
2c7f443777001cbef0db2f84f9f5e586746cb574 - correctly places the 
networking library as a library rather than a linker flag, fixing a 
build issue on QNX
00949a863e434a1d0f279bd7115b8e1a90022fb0 - Corrects the extension 
expected on C++ source files, other the latest android build tools 
reject it as not a valid extension.

Regards,

Toby

Re: [Barry-devel] Regarding Blackberry data recovery.

[Rakesh S. <rak...@gm...>]

Thanks for your help Chris. :-)

Regards
Rakesh S.

On Fri, Jan 6, 2012 at 5:00 PM, Chris Frey <cd...@fo...> wrote:

> On Fri, Jan 06, 2012 at 04:34:00PM +0530, Rakesh Sharnagate wrote:
> > Hi Chris,
> >
> > One last question I want to ask:
> > Which Initialization USB commands are required to extract records from
> any
> > database.
>
> They are pretty much all Bulk transfers, except for non-database things
> like
> setting the USB power levels.
>
> The Bulk transfers are arranged into packets, the format of which you can
> find in protostructs.h, including a socket number, size, command, etc.
> Some of these commands have sub-commands, and some command numbers
> are dynamic, such as tableCmd.
>
> For example, if you run 'btool -vd Memos' you'll see the USB traffic, for
> each stage: probe, mode selection (RIM Desktop), command code discovery
> (Database Access), DatabaseDatabase retrieval (a large block with a list
> of databases in the device... easy to find in the hex data), and then
> the dump of all records in the Memos database (using command 0x40 and
> the Database Access code number from the earlier command code discovery).
> You'll see sequence packets scattered throughout (command code 0x13).
>
> You can find the probe code in probe.cc.  Look at controller.cc and
> m_desktop.cc for the next stages.  But you'll have to match the packets
> against the right structs in protostructs.h to know what the bytes mean.
> Not all fields are understood yet, but enough to make it work.
>
> For database access, focus on the Packet struct in protostructs.h, and
> follow the other structs that are used in the unions.
>
> - Chris
>
>
>
> ------------------------------------------------------------------------------
> Ridiculously easy VDI. With Citrix VDI-in-a-Box, you don't need a complex
> infrastructure or vast IT resources to deliver seamless, secure access to
> virtual desktops. With this all-in-one solution, easily deploy virtual
> desktops for less than the cost of PCs and save 60% on VDI infrastructure
> costs. Try it free! http://p.sf.net/sfu/Citrix-VDIinabox
> _______________________________________________
> Barry-devel mailing list
> Bar...@li...
> https://lists.sourceforge.net/lists/listinfo/barry-devel
>

Re: [Barry-devel] Regarding Blackberry data recovery.

[Chris F. <cd...@fo...>]

On Fri, Jan 06, 2012 at 04:34:00PM +0530, Rakesh Sharnagate wrote:
> Hi Chris,
> 
> One last question I want to ask:
> Which Initialization USB commands are required to extract records from any
> database.

They are pretty much all Bulk transfers, except for non-database things like
setting the USB power levels.

The Bulk transfers are arranged into packets, the format of which you can
find in protostructs.h, including a socket number, size, command, etc.
Some of these commands have sub-commands, and some command numbers
are dynamic, such as tableCmd.

For example, if you run 'btool -vd Memos' you'll see the USB traffic, for
each stage: probe, mode selection (RIM Desktop), command code discovery
(Database Access), DatabaseDatabase retrieval (a large block with a list
of databases in the device... easy to find in the hex data), and then
the dump of all records in the Memos database (using command 0x40 and
the Database Access code number from the earlier command code discovery).
You'll see sequence packets scattered throughout (command code 0x13).

You can find the probe code in probe.cc.  Look at controller.cc and
m_desktop.cc for the next stages.  But you'll have to match the packets
against the right structs in protostructs.h to know what the bytes mean.
Not all fields are understood yet, but enough to make it work.

For database access, focus on the Packet struct in protostructs.h, and
follow the other structs that are used in the unions.

- Chris

Re: [Barry-devel] Regarding Blackberry data recovery.

[Rakesh S. <rak...@gm...>]

Hi Chris,

One last question I want to ask:
Which Initialization USB commands are required to extract records from any
database.

Thanks for your help

Regards
Rakesh S.

On Fri, Jan 6, 2012 at 3:20 PM, Chris Frey <cd...@fo...> wrote:

> On Fri, Jan 06, 2012 at 02:40:25PM +0530, Rakesh Sharnagate wrote:
> > Hi Chris,
> > Thanks for instant reply....
> >
> > I am trying to develop "Barry" like application for windows platform. Can
> > you please tell me from where or how do I get low level details.
>
> For low level details, it's the same answer I gave before. :-)
>
>        If you just want to learn the protocol, take a look at the
>        source code under src/.  Specifically, look at protocol.h,
>        protostructs.h, packet.{h,cc}, and socket.{h,cc}.  That should
>        get you started.
>
> Remember you can run the Barry command line tools, such as btool and bio,
> with the -v option, to see the low level USB traffic that Barry uses.
>
> But if you want to develop a Barry like application for Windows, you
> could just port Barry itself to Windows.  In theory, Barry should work just
> fine on Windows, but nobody has truly needed it yet.  You'll need GTK
> libraries, libusb for windows, and wxWidgets, and probably some POSIX
> wrapper...  maybe cygwin, or you can code wrappers for those yourself.
>
> The subject line said of this email thread says "data recovery",
> so if that's the ultimate goal, you might even want to switch to Linux
> instead of doing the Windows porting effort.
>
> Hope that helps,
> - Chris
>
>
>
> ------------------------------------------------------------------------------
> Ridiculously easy VDI. With Citrix VDI-in-a-Box, you don't need a complex
> infrastructure or vast IT resources to deliver seamless, secure access to
> virtual desktops. With this all-in-one solution, easily deploy virtual
> desktops for less than the cost of PCs and save 60% on VDI infrastructure
> costs. Try it free! http://p.sf.net/sfu/Citrix-VDIinabox
> _______________________________________________
> Barry-devel mailing list
> Bar...@li...
> https://lists.sourceforge.net/lists/listinfo/barry-devel
>

Re: [Barry-devel] Regarding Blackberry data recovery.

[Chris F. <cd...@fo...>]

On Fri, Jan 06, 2012 at 02:40:25PM +0530, Rakesh Sharnagate wrote:
> Hi Chris,
> Thanks for instant reply....
> 
> I am trying to develop "Barry" like application for windows platform. Can
> you please tell me from where or how do I get low level details.

For low level details, it's the same answer I gave before. :-)

	If you just want to learn the protocol, take a look at the
	source code under src/.  Specifically, look at protocol.h,
	protostructs.h, packet.{h,cc}, and socket.{h,cc}.  That should
	get you started.

Remember you can run the Barry command line tools, such as btool and bio,
with the -v option, to see the low level USB traffic that Barry uses.

But if you want to develop a Barry like application for Windows, you
could just port Barry itself to Windows.  In theory, Barry should work just
fine on Windows, but nobody has truly needed it yet.  You'll need GTK
libraries, libusb for windows, and wxWidgets, and probably some POSIX
wrapper...  maybe cygwin, or you can code wrappers for those yourself.

The subject line said of this email thread says "data recovery",
so if that's the ultimate goal, you might even want to switch to Linux
instead of doing the Windows porting effort.

Hope that helps,
- Chris

Re: [Barry-devel] Regarding Blackberry data recovery.

[Rakesh S. <rak...@gm...>]

Hi Chris,
Thanks for instant reply....

I am trying to develop "Barry" like application for windows platform. Can
you please tell me from where or how do I get low level details.

Regards
Rakesh S.

On Fri, Jan 6, 2012 at 2:20 PM, Chris Frey <cd...@fo...> wrote:

> On Fri, Jan 06, 2012 at 02:13:38PM +0530, Rakesh Sharnagate wrote:
> > Hi Chris,
> > I am snooping USB traffic when Blackberry's Desktop Manager is taking
> > backup of my Bold 9700 under Windows XP.
> >
> > Frankly speaking I am not understanding how to use USB log to start
> > interacting with blackberry device and desktop application.
> > How do I initialize BB for data transaction ?
>
> Hi Rakesh,
>
> You initial question seemed to be about how to reverse engineer the
> BlackBerry USB protocol.  I was assuming you wanted lots of low level
> detail.
>
> If you just want to use Barry to access your devices, you can do that
> via C++ using the Barry library, or even the command line tools like
> btool or bio, or the GUI which does backups and restores.
>
> Please let me know which direction you're heading.
>
> Thanks,
> - Chris
>
>
>
> ------------------------------------------------------------------------------
> Ridiculously easy VDI. With Citrix VDI-in-a-Box, you don't need a complex
> infrastructure or vast IT resources to deliver seamless, secure access to
> virtual desktops. With this all-in-one solution, easily deploy virtual
> desktops for less than the cost of PCs and save 60% on VDI infrastructure
> costs. Try it free! http://p.sf.net/sfu/Citrix-VDIinabox
> _______________________________________________
> Barry-devel mailing list
> Bar...@li...
> https://lists.sourceforge.net/lists/listinfo/barry-devel
>

Re: [Barry-devel] Regarding Blackberry data recovery.

[Chris F. <cd...@fo...>]

On Fri, Jan 06, 2012 at 02:13:38PM +0530, Rakesh Sharnagate wrote:
> Hi Chris,
> I am snooping USB traffic when Blackberry's Desktop Manager is taking
> backup of my Bold 9700 under Windows XP.
> 
> Frankly speaking I am not understanding how to use USB log to start
> interacting with blackberry device and desktop application.
> How do I initialize BB for data transaction ?

Hi Rakesh,

You initial question seemed to be about how to reverse engineer the
BlackBerry USB protocol.  I was assuming you wanted lots of low level
detail.

If you just want to use Barry to access your devices, you can do that
via C++ using the Barry library, or even the command line tools like
btool or bio, or the GUI which does backups and restores.

Please let me know which direction you're heading.

Thanks,
- Chris

Re: [Barry-devel] Regarding Blackberry data recovery.

[Rakesh S. <rak...@gm...>]

Hi Chris,
I am snooping USB traffic when Blackberry's Desktop Manager is taking
backup of my Bold 9700 under Windows XP.

Frankly speaking I am not understanding how to use USB log to start
interacting with blackberry device and desktop application.
How do I initialize BB for data transaction ?

Regards
Rakesh S.


On Thu, Jan 5, 2012 at 3:14 AM, Chris Frey <cd...@fo...> wrote:

> On Wed, Jan 04, 2012 at 01:48:28PM +0530, Rakesh Sharnagate wrote:
> > I am hobbiest and freelance software programmer. I am working on
> Blackberry
> > data recovery using USB interface from last 3 months. I read USB protocol
> > from USB.org and beyondlogic.org and very well aware of USB 2.0
> protocol.
> > I am trying to reverse engineer Blackberry USB protocol using USBlyzer,
> but
> > didn't achieve success.
> >
> > Expecting help !
>
> Hi Rakesh,
>
> When you're snooping the USB traffic, what software is running?  Are you
> doing a backup using the Windows desktop software?
>
> Barry already handles the protocol work for backups, restores, javaloader,
> modem, screenshots, date/time, etc.  There are a few missing features,
> but overall, it is very useable.
>
> If you just want to learn the protocol, take a look at the source code
> under src/.  Specifically, look at protocol.h, protostructs.h,
> packet.{h,cc},
> and socket.{h,cc}.  That should get you started.
>
> Hope that helps,
> - Chris
>
>
>
> ------------------------------------------------------------------------------
> Ridiculously easy VDI. With Citrix VDI-in-a-Box, you don't need a complex
> infrastructure or vast IT resources to deliver seamless, secure access to
> virtual desktops. With this all-in-one solution, easily deploy virtual
> desktops for less than the cost of PCs and save 60% on VDI infrastructure
> costs. Try it free! http://p.sf.net/sfu/Citrix-VDIinabox
> _______________________________________________
> Barry-devel mailing list
> Bar...@li...
> https://lists.sourceforge.net/lists/listinfo/barry-devel
>

Re: [Barry-devel] btool save database to device help

[Chris F. <cd...@fo...>]

On Thu, Jan 05, 2012 at 07:20:19PM -0600, Bruce Guenter wrote:
> > Are you using 0.2x opensync, or the devel tree (upcoming 0.4x)?
> 
> I've been using 0.39 on Gentoo.

If you're using the devel tree of opensync (all 0.3x versions are development
versions), you might consider switching to the latest git version.

I have a git repo called binary-meta, which builds opensync 0.2x and 0.3x/0.4x
as well as some of the plugins.  It does source builds as well as Debian
and RPM package builds.  Gentoo is one that I haven't tackled.

In any case, my opensync development is happening on git branches
that you can see here:

	http://repo.or.cz/w/opensync.git (scroll to the bottom to see the forks)

The latest opensync tree is not the above repo, but this fork:

	http://repo.or.cz/w/opensync/opensync-cdf.git

If you want to build using binary-meta, grab the following tree first, which
will grab the others it needs, including Barry:

	http://repo.or.cz/w/opensync/binary-meta.git

If repo.or.cz gives you disconnect problems, just keep trying.  Sometimes
it has load trouble.

Once you have binary-meta, see the README files for more information, and
the help text when you type 'make'.

Binary-meta is the source of the next Barry release, when it happens.
The next Barry release is planned to include opensync dependencies and
some plugins as well.  Ideally, I want it easy for people to add Barry
to their apt-get sources or yum repos and keep up to date that way with
binary packages.  If you know someone on the Gentoo side who would like
to help make an ebuild for binary-meta, that would be great, since tracking
that tree is where the latest changes appear, not SVN anymore.


Ok, onward to your actual problem :-)

> Here's what I did:
> 
> osynctool --addgroup g
> osynctool --addmember g barry-sync
> osynctool --addmember g file-sync
> osynctool --configure g 1
> # stuck in the PIN and turned off everything but memos
> # the default config includes debug info turned on
> osynctool --configure g 2
> # put in a real path
> osynctool --discover g
> osynctool --sync g
> 
> This, of course, prints out a lot of info, and ends with:
> 
> Received an entry 1571527172 (vjournal) from member 1 (barry-sync).  Changetype ADDED
> note sink of member 1 of type barry-sync just sent all changes
> Main sink of member 1 of type barry-sync just sent all changes
> All clients sent changes or error
> All changes got mapped
> All conflicts have been reported
> The sync failed: Couldn't find Object Type Engine for Object Type "note" while preparing for write phase.
> Main sink of member 2 of type file-sync just disconnected
> note sink of member 1 of type barry-sync had an error: Broken Pipe
> Main sink of member 1 of type barry-sync had an error: Broken Pipe
> All clients have disconnected

I did a quick test here with the latest code, and it worked for me, but
I had to change the file-sync <ObjType> block to <ObjType>note</ObjType>.

When I did the discover, it showed a mismatch (file-sync defaults to contact)
until I changed it.  Even when syncing, though, it didn't give me an error,
it just didn't sync two different types.

I suspect that either you're missing a plugin, or there's a bug in your
older version of opensync.

Try osynctool --listformats and see if a 'memo' format with 'note' object type
is available.

- Chris

Re: [Barry-devel] btool save database to device help

[Bruce G. <br...@un...>]

On Thu, Jan 05, 2012 at 06:53:08PM -0500, Chris Frey wrote:
> You're welcome to ask here, or if you prefer on opensync-devel.
> I kinda ended up being the opensync maintainer too.  So you've
> reached the right guy. :-)

Cool LOL

> Are you using 0.2x opensync, or the devel tree (upcoming 0.4x)?

I've been using 0.39 on Gentoo.

Here's what I did:

osynctool --addgroup g
osynctool --addmember g barry-sync
osynctool --addmember g file-sync
osynctool --configure g 1
# stuck in the PIN and turned off everything but memos
# the default config includes debug info turned on
osynctool --configure g 2
# put in a real path
osynctool --discover g
osynctool --sync g

This, of course, prints out a lot of info, and ends with:

Received an entry 1571527172 (vjournal) from member 1 (barry-sync).  Changetype ADDED
note sink of member 1 of type barry-sync just sent all changes
Main sink of member 1 of type barry-sync just sent all changes
All clients sent changes or error
All changes got mapped
All conflicts have been reported
The sync failed: Couldn't find Object Type Engine for Object Type "note" while preparing for write phase.
Main sink of member 2 of type file-sync just disconnected
note sink of member 1 of type barry-sync had an error: Broken Pipe
Main sink of member 1 of type barry-sync had an error: Broken Pipe
All clients have disconnected

-- 
Bruce Guenter <br...@un...>                http://untroubled.org/

Re: [Barry-devel] btool save database to device help

[Chris F. <cd...@fo...>]

On Thu, Jan 05, 2012 at 08:38:28AM -0600, Bruce Guenter wrote:
> I retried, and can confirm it did actually work.  I hadn't actually
> exited the memos application on the BB to check if anything was stored,
> and it showed nothing.  Exiting and restarting MemoPad worked.  Between
> that and the confusing output from btool, I was convinced I had wiped
> all the memos and they weren't uploading again.

Glad it worked for you!  I've changed the status message so it's not
so confusing anymore.


> Actually, what I really want is to sync my memos to/from files, but was
> having problems with osynctool (between barry-sync and file-sync).
> Using btool (or bio) directly seemed to be a useful stop-gap.  Should I
> ask about that here, or is there a better forum?

You're welcome to ask here, or if you prefer on opensync-devel.
I kinda ended up being the opensync maintainer too.  So you've
reached the right guy. :-)

Are you using 0.2x opensync, or the devel tree (upcoming 0.4x)?

- Chris

Re: [Barry-devel] Gcc 4.7

[Chris F. <cd...@fo...>]

On Thu, Jan 05, 2012 at 01:47:56PM -0700, Nathanael Noblet wrote:
> Hello,
> 
>   Fedora 17 will include gcc-4.7 and as such the packages have all been
> rebuilt. Barry failed because of a change in gcc. The following patch fixes
> the issue.

Thanks!  Applied.

- Chris

Re: [Barry-devel] btool save database to device help

[Bruce G. <br...@un...>]

On Thu, Jan 05, 2012 at 05:20:53AM -0500, Chris Frey wrote:
> I've just tried this very test with git master here, and it works.
> It does report 0 records, but the records are restored.

I retried, and can confirm it did actually work.  I hadn't actually
exited the memos application on the BB to check if anything was stored,
and it showed nothing.  Exiting and restarting MemoPad worked.  Between
that and the confusing output from btool, I was convinced I had wiped
all the memos and they weren't uploading again.

> If you're running git master, there is also the bio tool, which is
> more flexible if you want to work with data in a stream-like style.
> 
> The nice thing about bio is that it supports the tar backup format,
> which gets away from the boost file format, which can be brittle.

Apparently this all works with 0.17 too -- I played around with it for a
while and it does much more of what I want than btool can.

Actually, what I really want is to sync my memos to/from files, but was
having problems with osynctool (between barry-sync and file-sync).
Using btool (or bio) directly seemed to be a useful stop-gap.  Should I
ask about that here, or is there a better forum?

Thanks.

-- 
Bruce Guenter <br...@un...>                http://untroubled.org/

Re: [Barry-devel] btool save database to device help

[Chris F. <cd...@fo...>]

On Wed, Jan 04, 2012 at 08:27:22PM -0600, Bruce Guenter wrote:
> Hi.
> 
> How do I make btool store a database onto my BlackBerry?  This seems
> like it should be simple but I have been unable to get it to work.  No
> matter what I have tried with btool -s, all I get is "Store counted 0
> records." and an empty database on the BB.

Hi Bruce,

I should fix that message.  Store only counts the records that are
taken from the device, to memory (and eventually to the boost file).
It will report 0 when restoring from file to device.


> Here's what I've done.  I am trying to transfer some memos from my
> computer to my BB.  As a test, I dumped out the current "Memos" database
> to a file with btool -d Memos.  For good measure, I stored it in the
> default format, vcard format, with null conversion, and as a boost dump
> with -f.
> 
> Now I am trying to reload that data back onto the BB.  When I do:
> 
> 	btool -s Memos -f Memos.boost
> 
> it does load and parse the memos in that file, reports "# records
> loaded" but still stores nothing.

I've just tried this very test with git master here, and it works.
It does report 0 records, but the records are restored.

If you're running git master, there is also the bio tool, which is
more flexible if you want to work with data in a stream-like style.
You can take record data from one of the supported inputs, and write
it to multiple outputs.

For example:

This simulates btools's -d output:

	bio -i device -d Memos -o dump

This reads a boost file and dumps to stdout:

	bio -i boost -f Memos.boost -o dump

This reads a boost file and writes to the device:

	bio -i boost -f Memos.boost -o device -w erase


The nice thing about bio is that it supports the tar backup format,
which gets away from the boost file format, which can be brittle.


This is just additional information... btool should work just fine
for you.  I would recommend trying this as a confirmation:

	btool -d Memos -f Memos.boost
	bio -i boost -f Memos.boost -o dump

If you see data, then the boost file is valid, and the problem is in
the btool -s stage.  If this is the case, try adding -v to the btool
command line, and posting the results to see where it's failing.

- Chris

[Barry-devel] btool save database to device help

[Bruce G. <br...@un...>]

Hi.

How do I make btool store a database onto my BlackBerry?  This seems
like it should be simple but I have been unable to get it to work.  No
matter what I have tried with btool -s, all I get is "Store counted 0
records." and an empty database on the BB.

Here's what I've done.  I am trying to transfer some memos from my
computer to my BB.  As a test, I dumped out the current "Memos" database
to a file with btool -d Memos.  For good measure, I stored it in the
default format, vcard format, with null conversion, and as a boost dump
with -f.

Now I am trying to reload that data back onto the BB.  When I do:

	btool -s Memos -f Memos.boost

it does load and parse the memos in that file, reports "# records
loaded" but still stores nothing.

I have tried this with both version 0.17.1 and git master.

What am I doing wrong?

-- 
Bruce Guenter <br...@un...>                http://untroubled.org/

Re: [Barry-devel] Regarding Blackberry data recovery.

[Chris F. <cd...@fo...>]

On Wed, Jan 04, 2012 at 01:48:28PM +0530, Rakesh Sharnagate wrote:
> I am hobbiest and freelance software programmer. I am working on Blackberry
> data recovery using USB interface from last 3 months. I read USB protocol
> from USB.org and beyondlogic.org and very well aware of USB 2.0 protocol.
> I am trying to reverse engineer Blackberry USB protocol using USBlyzer, but
> didn't achieve success.
> 
> Expecting help !

Hi Rakesh,

When you're snooping the USB traffic, what software is running?  Are you
doing a backup using the Windows desktop software?

Barry already handles the protocol work for backups, restores, javaloader,
modem, screenshots, date/time, etc.  There are a few missing features,
but overall, it is very useable.

If you just want to learn the protocol, take a look at the source code
under src/.  Specifically, look at protocol.h, protostructs.h, packet.{h,cc},
and socket.{h,cc}.  That should get you started.

Hope that helps,
- Chris

[Barry-devel] Regarding Blackberry data recovery.

[Rakesh S. <rak...@gm...>]

I am hobbiest and freelance software programmer. I am working on Blackberry
data recovery using USB interface from last 3 months. I read USB protocol
from USB.org and beyondlogic.org and very well aware of USB 2.0 protocol.
I am trying to reverse engineer Blackberry USB protocol using USBlyzer, but
didn't achieve success.

Expecting help !


Regards
Rakesh S.

Re: [Barry-devel] USB Logs

[Chris F. <cd...@fo...>]

Thanks to you both!  I've updated USB-capture.txt to reflect this info.

- Chris


On Fri, Dec 23, 2011 at 11:34:10AM +0100, Nicolas CARRIER wrote:
> Hello,
> Personally I run a windows XP on a virtualbox, hosted on a linux desktop.
> Wireshark+usbmon can then sniff all the USB packets even when the device is attached to the virtual machine.
> It works very well.
> On 12/23/11 11:24, Toby Gray wrote:
> > On 23/12/11 02:01, Chris Frey wrote:
> >>> Are there any newer ways to get low-level logging information from running
> >>> loader that anyone is aware of?
> >> I'm afraid that USB-capture.txt is as complete as I know of. :-)
> >
> > There are a couple of other windows tools (all XP only I think) listed
> > at http://wiki.wireshark.org/Tools#USB_capture
> >
> > I think I used SnoopyPro in the end.
> >
> > If you've got (quite a lot of) money to spend then you could also go for
> > a hardware USB analyzer, such as the LeCroy Advisor T3
> > (http://www.lecroy.com/protocolanalyzer/protocoloverview.aspx?seriesid=280).
> >
> >
> > If you've got a BeagleBoard (http://beagleboard.org/), or possibly other
> > Linux hardware with a host and OTG USB port, then you might want to have
> > a look at: http://elinux.org/BeagleBoard/GSoC/2010_Projects/USBSniffer
> >
> > Regards,
> >
> > Toby
> >
> > ------------------------------------------------------------------------------
> > Write once. Port to many.
> > Get the SDK and tools to simplify cross-platform app development. Create
> > new or port existing apps to sell to consumers worldwide. Explore the
> > Intel AppUpSM program developer opportunity. appdeveloper.intel.com/join
> > http://p.sf.net/sfu/intel-appdev
> > _______________________________________________
> > Barry-devel mailing list
> > Bar...@li...
> > https://lists.sourceforge.net/lists/listinfo/barry-devel
> >
> 
> 
> -- 
> Nicolas CARRIER - 6081 - 2?me centre
> 
> ------------------------------------------------------------------------------
> Write once. Port to many.
> Get the SDK and tools to simplify cross-platform app development. Create 
> new or port existing apps to sell to consumers worldwide. Explore the 
> Intel AppUpSM program developer opportunity. appdeveloper.intel.com/join
> http://p.sf.net/sfu/intel-appdev
> _______________________________________________
> Barry-devel mailing list
> Bar...@li...
> https://lists.sourceforge.net/lists/listinfo/barry-devel

Re: [Barry-devel] USB Logs

[Nicolas C. <nic...@pa...>]

Hello,
Personally I run a windows XP on a virtualbox, hosted on a linux desktop.
Wireshark+usbmon can then sniff all the USB packets even when the device is attached to the virtual machine.
It works very well.
On 12/23/11 11:24, Toby Gray wrote:
> On 23/12/11 02:01, Chris Frey wrote:
>>> Are there any newer ways to get low-level logging information from running
>>> loader that anyone is aware of?
>> I'm afraid that USB-capture.txt is as complete as I know of. :-)
>
> There are a couple of other windows tools (all XP only I think) listed
> at http://wiki.wireshark.org/Tools#USB_capture
>
> I think I used SnoopyPro in the end.
>
> If you've got (quite a lot of) money to spend then you could also go for
> a hardware USB analyzer, such as the LeCroy Advisor T3
> (http://www.lecroy.com/protocolanalyzer/protocoloverview.aspx?seriesid=280).
>
>
> If you've got a BeagleBoard (http://beagleboard.org/), or possibly other
> Linux hardware with a host and OTG USB port, then you might want to have
> a look at: http://elinux.org/BeagleBoard/GSoC/2010_Projects/USBSniffer
>
> Regards,
>
> Toby
>
> ------------------------------------------------------------------------------
> Write once. Port to many.
> Get the SDK and tools to simplify cross-platform app development. Create
> new or port existing apps to sell to consumers worldwide. Explore the
> Intel AppUpSM program developer opportunity. appdeveloper.intel.com/join
> http://p.sf.net/sfu/intel-appdev
> _______________________________________________
> Barry-devel mailing list
> Bar...@li...
> https://lists.sourceforge.net/lists/listinfo/barry-devel
>


-- 
Nicolas CARRIER - 6081 - 2ème centre