Blackberry -> WM sync fails

[Adam Williamson]

Hi, all.

I just had cause to try something slightly unusual. My partner got an
HTC Touch Diamond from his work as a demo model for a couple of months,
so he wanted to transfer all his contacts from his Blackberry Curve to
the HTC.

This should work great with opensync, so I tried it. I set up a sync
group with the Barry and synce plugins and synchronized. It went through
great. There's only one problem - the contacts arrive at the WM phone
without phone numbers :) they have email addresses, but no phone
numbers.

I suspect this is a problem somehow in the synce plugin's handling of
the data received from the Blackberry? To this end, I set up another
group which just syncs the Blackberry to the file sync plugin, and I'm
attaching an example contact file from this sync. You can see from this
the format the Blackberry spits out, I guess. Can someone then see why
the phone number info would not sync to the WM device? I changed the
name and number, BTW. Thanks!

[Adam Williamson]

Post by Adam Williamson
I suspect this is a problem somehow in the synce plugin's handling of
the data received from the Blackberry? To this end, I set up another
group which just syncs the Blackberry to the file sync plugin, and I'm
attaching an example contact file from this sync. You can see from this
the format the Blackberry spits out, I guess. Can someone then see why
the phone number info would not sync to the WM device? I changed the
name and number, BTW. Thanks!

I just did the same thing with a WM device, and I notice the following
difference. The BlackBerry uses this format for the numbers:

TEL;work:6048359373

WM device uses:

TEL;WORK;VOICE:(425) 1657643

I suspect the missing 'VOICE' is the key?

[Adam Williamson]

Post by Adam Williamson
I just did the same thing with a WM device, and I notice the following
TEL;work:6048359373
TEL;WORK;VOICE:(425) 1657643
I suspect the missing 'VOICE' is the key?

OK, solved it. It's not *just* the VOICE bit. It's either that bit
*plus* the case, or just the case.

I tried changing TEL;work;number to TEL;work;voice;number (via sed) in
the Blackberry file sync and then syncing those files back to the WM
device. No good. But then I changed to TEL;WORK;VOICE;number and did it
again - and boom, works fine, the numbers show up.

Hopefully someone can now fix this so you don't have to sync to file,
run a sed voodoo command on the files, and then sync them back to the WM
device. =) Thanks all.

[Mark Ellis]

Post by Adam Williamson

Post by Adam Williamson
I just did the same thing with a WM device, and I notice the following
TEL;work:6048359373
TEL;WORK;VOICE:(425) 1657643
I suspect the missing 'VOICE' is the key?

OK, solved it. It's not *just* the VOICE bit. It's either that bit
*plus* the case, or just the case.
I tried changing TEL;work;number to TEL;work;voice;number (via sed) in
the Blackberry file sync and then syncing those files back to the WM
device. No good. But then I changed to TEL;WORK;VOICE;number and did it
again - and boom, works fine, the numbers show up.
Hopefully someone can now fix this so you don't have to sync to file,
run a sed voodoo command on the files, and then sync them back to the WM
device. =) Thanks all.
--

Adam, what opensync version are you on ? Apparently the phone number foo
in 0.3x is broken, and John decided there is no point fixing it until
they stabilise, just wondered if you're on 0.3x you might like to try
0.2x.

Mark

[John Carr]

Post by Mark Ellis

Post by Adam Williamson

Post by Adam Williamson
I just did the same thing with a WM device, and I notice the following
TEL;work:6048359373
TEL;WORK;VOICE:(425) 1657643
I suspect the missing 'VOICE' is the key?

OK, solved it. It's not *just* the VOICE bit. It's either that bit
*plus* the case, or just the case.
I tried changing TEL;work;number to TEL;work;voice;number (via sed) in
the Blackberry file sync and then syncing those files back to the WM
device. No good. But then I changed to TEL;WORK;VOICE;number and did it
again - and boom, works fine, the numbers show up.
Hopefully someone can now fix this so you don't have to sync to file,
run a sed voodoo command on the files, and then sync them back to the WM
device. =) Thanks all.
--

Adam, what opensync version are you on ? Apparently the phone number foo
in 0.3x is broken, and John decided there is no point fixing it until
they stabilise, just wondered if you're on 0.3x you might like to try
0.2x.
Mark

This is a known bug in 0.22 (and presumably 0.36, 0.37 doesnt work at
all). See the thread entitled "Certain phone types not syncing from
Nokia to WM6 device" from a few weeks ago. If someone else acks the
XSL attached to the last message in the thread, we should commit it.

John

[Adam Williamson]

Post by John Carr

Post by Mark Ellis
Adam, what opensync version are you on ? Apparently the phone number foo
in 0.3x is broken, and John decided there is no point fixing it until
they stabilise, just wondered if you're on 0.3x you might like to try
0.2x.
Mark

This is a known bug in 0.22 (and presumably 0.36, 0.37 doesnt work at
all). See the thread entitled "Certain phone types not syncing from
Nokia to WM6 device" from a few weeks ago. If someone else acks the
XSL attached to the last message in the thread, we should commit it.

Right, I'm on 0.22 as I always am.

I forgot that thread. I'll test that XSL file.

[Adam Williamson]

Post by Adam Williamson

Post by John Carr

Post by Mark Ellis
Adam, what opensync version are you on ? Apparently the phone number foo
in 0.3x is broken, and John decided there is no point fixing it until
they stabilise, just wondered if you're on 0.3x you might like to try
0.2x.
Mark

This is a known bug in 0.22 (and presumably 0.36, 0.37 doesnt work at
all). See the thread entitled "Certain phone types not syncing from
Nokia to WM6 device" from a few weeks ago. If someone else acks the
XSL attached to the last message in the thread, we should commit it.

Right, I'm on 0.22 as I always am.
I forgot that thread. I'll test that XSL file.

Actually, eyeballing it, that XSL file fixes the VOICE keyword issue but
does not fix the case issue. How do I phrase "'HOME' or Type='home') and
Type!='FAX'" correctly in, er, .xsl-language? :)

[Adam Williamson]

Post by Adam Williamson

Post by Adam Williamson

Post by John Carr
This is a known bug in 0.22 (and presumably 0.36, 0.37 doesnt work at
all). See the thread entitled "Certain phone types not syncing from
Nokia to WM6 device" from a few weeks ago. If someone else acks the
XSL attached to the last message in the thread, we should commit it.

Right, I'm on 0.22 as I always am.
I forgot that thread. I'll test that XSL file.

Actually, eyeballing it, that XSL file fixes the VOICE keyword issue but
does not fix the case issue. How do I phrase "'HOME' or Type='home') and
Type!='FAX'" correctly in, er, .xsl-language? :)

If someone could fix this I'd be really grateful. I just don't know how
to gracefully make the match case-insensitive. The issue is exactly as I
described it: we need to accept the number even without the voice/VOICE
keyword, and we need to make the matching of all keywords (voice, home,
cell, work etc) case-insensitive. I would do this if I could, but I just
don't have the knowledge.

[Dr J A Gow]

Post by Adam Williamson

Post by Adam Williamson

Post by Adam Williamson

Post by John Carr
This is a known bug in 0.22 (and presumably 0.36, 0.37 doesnt work at
all). See the thread entitled "Certain phone types not syncing from
Nokia to WM6 device" from a few weeks ago. If someone else acks the
XSL attached to the last message in the thread, we should commit it.

Right, I'm on 0.22 as I always am.
I forgot that thread. I'll test that XSL file.

Actually, eyeballing it, that XSL file fixes the VOICE keyword issue but
does not fix the case issue. How do I phrase "'HOME' or Type='home') and
Type!='FAX'" correctly in, er, .xsl-language? :)

If someone could fix this I'd be really grateful. I just don't know how
to gracefully make the match case-insensitive. The issue is exactly as I
described it: we need to accept the number even without the voice/VOICE
keyword, and we need to make the matching of all keywords (voice, home,
cell, work etc) case-insensitive. I would do this if I could, but I just
don't have the knowledge.

I'll look at this in the next batch of fixes.

John.

[Adam Williamson]

Post by Dr J A Gow

Post by Adam Williamson
If someone could fix this I'd be really grateful. I just don't know how
to gracefully make the match case-insensitive. The issue is exactly as I
described it: we need to accept the number even without the voice/VOICE
keyword, and we need to make the matching of all keywords (voice, home,
cell, work etc) case-insensitive. I would do this if I could, but I just
don't have the knowledge.

I'll look at this in the next batch of fixes.

That's great, thanks so much John.

[Adam Williamson]

Post by Adam Williamson

Post by Dr J A Gow

Post by Adam Williamson
If someone could fix this I'd be really grateful. I just don't know how
to gracefully make the match case-insensitive. The issue is exactly as I
described it: we need to accept the number even without the voice/VOICE
keyword, and we need to make the matching of all keywords (voice, home,
cell, work etc) case-insensitive. I would do this if I could, but I just
don't have the knowledge.

I'll look at this in the next batch of fixes.

That's great, thanks so much John.

Just to note - as we're getting rather close to our release, I have
applied Pawel's .xsl fix (to handle phone numbers with no VOICE subtype)
and a hack to the barry opensync plugin (so it outputs its vcards with
the relevant types in UPPER CASE) to our packages. Together, this
*should* mean Blackberry -> WM sync (and probably Nokia -> WM sync...)
will work more or less right in MDV 2009, even if Dr. Gow doesn't get
around to fixing it properly in time for me to get it in.

If Dr. Gow's fix turns up in time, I'll revert the Barry hack and use
that instead.

Tomorrow I can get my hands on a Blackberry again (partner switched to
another WM phone, sigh, I have to harangue him to bring a spare
Blackberry home) and actually test it. Testing, that's a novel
concept! :D

[Adam Williamson]

Post by Adam Williamson
Tomorrow I can get my hands on a Blackberry again (partner switched to
another WM phone, sigh, I have to harangue him to bring a spare
Blackberry home) and actually test it. Testing, that's a novel
concept! :D

Just for interest - the FrankenFix works. :) I just sync'ed a couple
hundred contacts, complete with phone numbers, direct from a Blackberry
Curve to an HTC Apache.

So, er, yay. Take that, other operating systems! Muahaha.

[Vladimir Andreev]

I found dhclient on gentoo, it's called "net-misc/dhcp-3.1.1" and
synce-hal really works with it. What I find a bit strange, that it did
not if I removed the dynamic configuration from the hal addon script and
let it use the fixed addresses..
As a side note, dhcpcd-4* which is "out" by now, has evolved also into
support dhclient-alike syntax, and is much more lightweight (as the
author says in the History at http://roy.marples.name/dhcpcd so it might
be worth trying to make synce-hal work with both. Otherwise it would be
really enforcing a specific dependency... So if you don't mind the idea,
I'm ready to try and help in any way I can (also with my limited spare
time, as everyone) with trying and such, I'm also a c/c++ programmer,
only don't have python and dhcp knowledge...
Anyway, I'd also like to assist Iain with gentoo packaging, as he seems
to be very busy, but showed up on IRC yesterday...

There is a problem then i try use net-misc/dhcp-3.1.1 in Gentoo to obtain ip
address :
# dhcpcd eth1
err, eth1: cannot request a link local address
err, eth1: cannot request a link local address
err, eth1: cannot request a link local address
...

The 169.254.0.0/16 set as local sublet in /etc/networks:
# /etc/networks
#
# This file describes a number of netname-to-adress
# mappings for the TCP/IP subsytem. It is mostly
# used at boot time, when no name servers are running.
#

loopback 127.0.0.0
link-local 169.254.0

So i think, the best way to use dhcp in synce-hal is add
DEPEND="net-misc/dhcp" in ebuild.

[Dr J A Gow]

Post by Adam Williamson

Post by Adam Williamson

Post by Dr J A Gow

Post by Adam Williamson
If someone could fix this I'd be really grateful. I just don't know how
to gracefully make the match case-insensitive. The issue is exactly as I
described it: we need to accept the number even without the voice/VOICE
keyword, and we need to make the matching of all keywords (voice, home,
cell, work etc) case-insensitive. I would do this if I could, but I just
don't have the knowledge.

I'll look at this in the next batch of fixes.

That's great, thanks so much John.

Just to note - as we're getting rather close to our release, I have
applied Pawel's .xsl fix (to handle phone numbers with no VOICE subtype)
and a hack to the barry opensync plugin (so it outputs its vcards with
the relevant types in UPPER CASE) to our packages. Together, this
*should* mean Blackberry -> WM sync (and probably Nokia -> WM sync...)
will work more or less right in MDV 2009, even if Dr. Gow doesn't get
around to fixing it properly in time for me to get it in.
If Dr. Gow's fix turns up in time, I'll revert the Barry hack and use
that instead.

That's fine, I may not get the fix in in time for the release due to
other commitments at the moment. I will check it thoroughly after the
release (I am also waiting on Opensync to fix the known bug in the
OS0.3x schemas to avoid duplicating work. This has been scheduled for a
future point release of OS).

John.