[patch] Number matching support

Joachim Breitner mail at joachim-breitner.de
Mon Jul 21 23:02:14 CEST 2008


Am Montag, den 21.07.2008, 12:22 +0200 schrieb Holger Freyther:
> On Saturday 19 July 2008 20:34:11 Joachim Breitner wrote:
> > I have added (seemingly) complete support for matching human-written
> > phone number (07032/123 456) in the phone book or when dialing, instead
> > of just dumb string comparison, in both openmoko-dialer2 and
> > openmoko-messages2. The latter required a small, non-obstrusive change
> > to eds and libjana. The related bug report is
> > https://docs.openmoko.org/trac/ticket/1606
> >
> > Please consider applying these git patches, which are against the
> > org.openmoko.dev branch, or otherwise tell my what needs to be changed
> > that they can be applied.
> Hey,
> I had hoped someone else, more qualified would step up. I think the issues you 
> bring up are.
> 1. Who is reading the bug report and acking it?
> 2. Who is reviewing patches and landing them?
> 3. Who makes sure people get updated packages?
> 1. I don't think we have anyone in Taipei responsible for going through 
> incoming OM2007.2 bugs yet. I know that Tony was asked to start organizing 
> support though. Yeah, I'm well aware how this sounds. :)
> Sometimes I go through the bugs....but I'm more than busy with ASU.
> 2. Most of OM2007.2 was done by o-hand.com and the contract with them on that 
> ended (to the best of my knowledge). So we can not ask them to fix/handle 
> these kind of bugs, we can do this on a private developer developer base 
> though. We can ask erin but these patches are tricky.
> 3. I don't know if XorA feels responsible for providing upgraded packages? If 
> opkg upgrade from factory to image to the latest .dev build is working at 
> all. I know that for ASU that this will be a task for the distro team 
> (Julian, Graeme, me).

Have you considered asking the community if someone wants to take part
of the work? I.e., reviewing patches might be something that a
contributor might do for you? And if none can be found to maintain the
2007.2 branch (or at least help so far as required), then that branch
can probably declared dead (but that needs to be stated explicitly and

> 1.)
> So I still hope for someone else to play bugmaster and manage, triage the 
> incoming bug reports as my workload is already incredible high. Let's see if 
> I can talk brenda into playing our bugmaster as well... 
> 2.)
> Patch review. Certainly all the three/four touched components are not in my 
> area of expertise. I'm a trained monkey/reviewer at webkit.org and we have 
> quite strict coding style guidelines[1]. But I try to ignore these for now.
> Technical:
> 	Comparing numbers will only work if they are exact matches or have at least 
> seven numbers? Where is that seven coming from?

We had a discussion on this list, and I brought up the seven as a
ballpark figure that is “probably good enough”, at least until people
complain. A very rough guess said that even a very international user of
the Freerunner has only a 2% chance of a collision, and this type of
person will probably use international prefixes every anyways.

> Approach:
> 	Can we somehow avoid the duplication of this code? Can we move the matching 
> somewhere else? Can we move the matching code into an evolution library (if 
> that exists)?

Someone on the list suggested a static library for it. Putting it in
some evolution library might work, but then it’s that messages and
dialer don’t link directly to any evolution library...

> 	You added patches for OpenEmbedded. Personally I would prefer if these things 
> could end up upstream. And at this state it might makes sense to talk to Ross 
> and Chris from o-hand.com as they work on evolution and control libjana....

Ok, I’ll talk to them. Probably the eds upstream does not want to link a
new library with that code and probably want it “inlined”, so this seems
a bit contradictional to the wish of avoiding duplicate code.

> So we can put the patches to Openmoko apps into the Openmoko repository (after 
> cleaning the code a bit). Then we still would have libjana and eds to go, 
> which we could patch in the metadata. But I would really feel a lot better if 
> we would add the numbering comparing only once... any ideas where such a 
> place could be?

No, not really. Not even -messages and -dialer link against some common
library that seems suitable. Maybe  libmokogsmd2. So, for this
relatively small and self-contained code, copy’n’paste might be
approriate – or I’m not creative enough :-).

Maybe some larger refactoring or pushing upstream can be done once the
code has been used a week or two in the 2007.2 archive.

>	Some more comments would be cool.

Point taken, when I have to submit more changes, I’ll add some comments.

Thanks for taking the time to review it!


Joachim "nomeata" Breitner
  mail: mail at joachim-breitner.de | ICQ# 74513189 | GPG-Key: 4743206C
  JID: nomeata at joachim-breitner.de | http://www.joachim-breitner.de/
  Debian Developer: nomeata at debian.org
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 197 bytes
Desc: Dies ist ein digital signierter Nachrichtenteil
Url : http://lists.openmoko.org/pipermail/devel/attachments/20080721/67243235/attachment.pgp 

More information about the devel mailing list