Android tracing and profiling

by Horia » Wed, 18 Jun 2008 16:56:47 GMT


Sponsored Links
 Hello,

The emulator and API provides a couple of ways of tracing and
profiling your applications:
1. the android.os.Debug provides some methods like:
startMethodTracing(), startNativeTracing()
2. the emulator could be run with "-trace" option in order to later
accept the start tracing command (F9)

The problem is nativeTracing (started by F9 or startNativeTracing()
method) generates some files I don't know how to use:
qtrace.bb
qtrace.exc
qtrace.pid
qtrace.method
qtrace.inst
qtrace.static

Does anyone know how can I view the contents of these files ?(they are
binaries, not ascii)

Thanks!

--~--~---------~--~----~------------~-------~--~----~



Android tracing and profiling

by Megha Joshi » Thu, 19 Jun 2008 03:14:50 GMT


 The tools to process and view these files are not shipped with the current
SDK. So , the native tracing isn't supported in the SDK yet.  You can only
use the method tracing for now.





--~--~---------~--~----~------------~-------~--~----~


Sponsored Links


Android tracing and profiling

by Horia » Fri, 20 Jun 2008 02:40:50 GMT


 Ok Joshi,
Thanks




--~--~---------~--~----~------------~-------~--~----~



Other Threads

1. Android and Funambol

Hi Stef,

I have spent a month off - skiing.

Nice to hear from Funambol that all companies below the scale of
mobile network carrier companies are
SMB businesses.

But that it is not really surprising me, I don't know if Funambol is
not willing or able to offer commercial licences to other businesses
than OEM and mobile carrier companies.
You are calling your one and only commercial license Funambol Carrier
license, this makes your product less attractive e.g. to (not so
small) multi-national insurance companies who on the one hand have
comparably low user numbers but on the other hand some of their
employees have a much higher demand for advanced data synchronization
applications.

In this field applications built around Funambol DS have to compete
with SAP MI (Mobile Infrastructure) and Oracle Database Lite solutions
with both offer advanced data synchronization  functionality out of
the box.

If for these "smaller" enterprise scale deployments you have to
license a "Carrier Edition", customers may feel greatly misunderstood
because their business needs are a little bit different from what
carriers offer to consumers and rather small shops. You simply cannot
compare carrier scale PIM synchronization with e.g. a management
information and support applications, that should also work on a long
business flight, where mobile web applications would fail totally.

Funambol Carrier Edition is definitely no special offer in these
scenarios, especially because we could assume that many not so small
enterprises are already Oracle or SAP costumers, if they aren't they
have often licensed at least Microsoft Exchange Server, which also
offers upto some degree mobile business data synchronization (Exchange
Stores).

Perhaps you may now see that neither a "Carrier Edition" (, which
costs US$ 120,000,) nor the fee-less "Funambol Community Edition"
copyleft license are hot candidates, if you want to tell a customer
the F/OSS business case, that he should invest into an open-source
mobile business infrastructure instead of taking established big-name
products.

Telling the enterprise clients the F/OSS business case, is in most
cases more complicated than telling them that they could save license
fees - that's often a minor aspect. They are used to get what they pay
for and they compare what they get from different vendors for their
money. These customers often use a software solution much longer than
a consumer his favourite electronic toy, they often expect that they
use it as long as 15 years and even longer. And that's the point where
SyncML becomes interesting, it's simply XML over any network protocol
you like. You are not trapped into difficult to understand product
specific binary protocols, which might produce a lot of headache if
you have either to reverse-engineer them or to migrate to something
totally new.

You are right it is too early to talk about what impact the AGPLv3
will have on products that are licensed under this license.

I have discussed this with some colleagues, we think that this will
have on the one hand a positive effect on the revenue of some open-
source companies, because some customers will feel the need to either
change to a commercial license or to less restrictive licensed
products, at the end of the day the only interesting numbers in
business life are your expenses and what you get in return, so
counting paying customers makes really sense.

But the down-side of changing to a more restrictive license is that
the product has to compete with the commercial products of the big-
names.
 And even more worse it may reduce the number of open-source
professionals among the community members. Of course you may try to
compensate this with making little money gifts to others but that's
often seen as a big taboo, because then you make people to become
second-class workers of your enterprise, which work for less than your
own regular employees.

In any case where an in-house solution is a competitive
differentiator, there is a strong need to protect its internals from
disclosure.
If your enterprise is split into legally independent companies, which
is quite often the case in today's larger multi-national enterprises,
AGPLv3  is definitely a show-stopper, because ...
-) you cannot withdraw this license from a group company that leaves
the group but has used the AGPLv3 poisoned product.
-) you cannot outsource parts of the operational aspects of using the
AGPLv3 poisoned product.
-) you cannot practice the "no need to know" principle between legally
independent business entities of your enterprise group, if you use the
AGPLv3 poisoned product, even if it has been developed in-house.

Funambol is a product AFAIK that has a quite good reputation.

I do not see a reason why you should not also target "smaller" multi-
national enterprises beside network carriers and OEM , of course they
do not need the same feature set as Carrier Edition customers, their
needs are little bit different. ;-)


Your practice to force community members to make a copyright
assignment to a commercial company instead of assigning their rights
to an ethical higher integrity entity as the Free Software Foundation
(FSF), is definitely not the best choice because it interferes with
the original philosophy behind free software.

People who believe that it does not matter to whom you assign all your
natural rights as a open-source contributor, should better think twice
if that's really true, companies who want to make them believe that,
are in my eyes not really honest to the open-source community, because
they will be never doing the same, they will never assign their rights
to the FSF, they are only using FSF licenses for their own commercial
benefit, without contributing them self a single line of code to the
great foundation of GNU software. They do not share honestly at all
with the community. They reserve themselves rights that allow them to
fully control their open-source contributors work, a quite popular
trick is to use a protected trademark as a package name, which makes
it legally impossible to create a fully compatible vendor independent
fork, but patents work also really well.

The ethical rights of the open-source community would be better
protected if you would allow community members to assign their
copyrights directly to the Free Software Foundation.

No one knows who will own Funambol in a few years, it is not a
privately hold business, it's financed by venture capital, so no one
at Funambol can make promises, that Funambol will always offer and
maintain an open-source edition.

It would not be the first time where people become quite unhappy when
they have to learn that open-source companies and some of their
products may disappear if they get acquired by one of big name
companies.

Open-source governance matters - not only for the community but also
for commercial licensees.
It is beneficial to all users of an open-source product if the project
is ruled by an open community instead of a single company behind.

That's a strong advantage of Android it's backed up by a large GROUP
of industrial leading companies and Google is also building actively a
free open-source community around it.

It would be quite interesting to see if this would have been also
possible using a strong copyleft license.

I am in doubt, because strong copyleft licenses can have unpredictable
economic side-effects, in cases where a single commercial entity has
the right to determine the future of an open-source product, they can
simply stop development and maintenance of the open-source version,
and worse they may publish lists of security issues which they have
patched for their commercial users and recommend users who are using
the discontinued open-source version to become commercial licensees.

Many managers share my concerns about using strong copyleft licensed
software to implement critical business processes, you have to examine
really carefully  not only the impact of the license itself but also
try to estimate the projects individual risks, that can be a tough
task because you have to gather a lot of informations about the
projects copyright owners, if it is not a community ruled project.
Sometimes even a single big name is no guarantee for the projects open
future.
I am just thinking of a very large German software vendor who has
decided to discontinue his open-source SQL data base.
If you have trusted their promises bad luck for you and your business.


Cheers
 George







--~--~---------~--~----~------------~-------~--~----~

2. chage in home screen

I want to change android default menu, anyone know how to do that ?
--~--~---------~--~----~------------~-------~--~----~

3. ListView in Activity

4. Media album_art

5. Spotlight Series on Developers for Android

6. Looking for Developers to Share Experiences

7. File Transfer