Google Summer of Code

Google Summer of Code

[Joel Sherrill]

Hi

Is GDB planning on participating in GSoC this year at least under the
FSF umbrella?

From the RTEMS side, we have some simulator improvements we
would like to see that would be of general use. But ideally, there would
be a mentor from both our side and the gdb side.

-- 
Joel Sherrill, Ph.D.             Director of Research & Development
joel.sherrill@OARcorp.com        On-Line Applications Research
Ask me about RTEMS: a free RTOS  Huntsville AL 35805
Support Available                (256) 722-9985

Re: Google Summer of Code

[Jose E. Marchesi]

    
    Is GDB planning on participating in GSoC this year at least under the
    FSF umbrella?

Under the GNU umbrella in any case.  The Free Software Foundation does
not participate as a mentoring organization in GSOC.  The GNU project
does.

We will be sending our application in the following days.  GDB is of
course welcome to participate with us.

Re: Google Summer of Code

[Joel Sherrill]

On 2/16/2015 12:42 PM, Jose E. Marchesi wrote:
>     
>     Is GDB planning on participating in GSoC this year at least under the
>     FSF umbrella?
>
> Under the GNU umbrella in any case.  The Free Software Foundation does
> not participate as a mentoring organization in GSOC.  The GNU project
> does.
OK. 
> We will be sending our application in the following days.  GDB is of
> course welcome to participate with us.
>
And does GDB have any Open Projects? Mentors? I haven't seen any discussions
here.  I am the RTEMS organization administrator and have listed a GDB
enhancement
on our project ideas page. But that won't be feasible without a GDB
mentor. It is
also not RTEMS specific so may make sense as a purely GDB project. Either
way, it needs a mentor from GDB.

I can act as co-mentor given my RTEMS and GDB experience.

-- 
Joel Sherrill, Ph.D.             Director of Research & Development
joel.sherrill@OARcorp.com        On-Line Applications Research
Ask me about RTEMS: a free RTOS  Huntsville AL 35805
Support Available                (256) 722-9985

Re: google summer of code

[teawater]

Hi thiago,

On Sun, Mar 8, 2009 at 14:27, Thiago Jung Bauermann <bauerman@br.ibm.com> wrote:
> Hi teawater,
>
> El vie, 27-02-2009 a las 10:27 +0800, teawater escribió:
>> On Thu, Feb 26, 2009 at 22:25, Thiago Jung Bauermann
>> <bauerman@br.ibm.com> wrote:
>> > Some that may be worth considering for this:
>> >
>> > 1. Help with Python scripting support;
>>
>> It must can make a lot of people interested in.  When I told about gdb
>> in some place, must have a lot of people ask me if gdb support python
>> or not.  :)
>
> That's awesome!
>
>> > 2. Help with reversible debugging (not sure if teawater and Michael need
>> > a helping hand there);
>>
>> I am not very clear the GSOC want student get .
>> But maybe we can let these students do some test with reverse debug
>> and precord,  of course write some testsuite is better. :)
>> They can not only test reverse debug with precord, but also test
>> reverse debug with some remote-stub that support reverse-debug (For
>> example, vmware and Simics. Maybe we need help from Jakob and
>> Michael.)
>> I think it will help gdb reverse debug very much. :)
>
> So if I understood what you said, you think a GSoC student could write
> testcases for the reversible debugging feature? Would you (or Michael
> Snyder, perhaps) be willing to mentor someone working on this activity?

I am not sure the student can write it or not.  Actually, I don't have
experience with it. :(  My testsuite is written by Michael.
So maybe Michael can be the mentor.  But I didn't get his mail a long
time.  I am not sure he can answer us or not.

And maybe we can just let the student test reverse debug with hand.
Or let a people that have experience on write testsuite to be the
mentor.  If he doesn't have the experience on reverse debug, I can
help him.


Thanks,
Hui

Re: google summer of code

[Thiago Jung Bauermann]

Hi teawater,

El vie, 27-02-2009 a las 10:27 +0800, teawater escribió:
> On Thu, Feb 26, 2009 at 22:25, Thiago Jung Bauermann
> <bauerman@br.ibm.com> wrote:
> > Some that may be worth considering for this:
> >
> > 1. Help with Python scripting support;
> 
> It must can make a lot of people interested in.  When I told about gdb
> in some place, must have a lot of people ask me if gdb support python
> or not.  :)

That's awesome!

> > 2. Help with reversible debugging (not sure if teawater and Michael need
> > a helping hand there);
> 
> I am not very clear the GSOC want student get .
> But maybe we can let these students do some test with reverse debug
> and precord,  of course write some testsuite is better. :)
> They can not only test reverse debug with precord, but also test
> reverse debug with some remote-stub that support reverse-debug (For
> example, vmware and Simics. Maybe we need help from Jakob and
> Michael.)
> I think it will help gdb reverse debug very much. :)

So if I understood what you said, you think a GSoC student could write
testcases for the reversible debugging feature? Would you (or Michael
Snyder, perhaps) be willing to mentor someone working on this activity?
-- 
[]'s
Thiago Jung Bauermann
IBM Linux Technology Center

Re: google summer of code

[Thiago Jung Bauermann]

Hi OÄŸuz,

El vie, 27-02-2009 a las 10:41 +0200, Oğuz Kayral escribió:
> Hi,
> The guys at Nmap preferred to apply for "Feature Creepers and Bug
> Wranglers" positions in order to have some students to work on lots of
> small bugs instead of one big project. I also read that, this method
> was pretty successful closing about 100 bugs in one summer. A similar
> approach can be taken in order to reduce the bug count in
> bugzilla(current 1365!, 52 of them for 6.8)

Agreed, That would be a very interesting project for GDB.

> Another thing is, the coding begins on May 23 and according to the
> release schedule gdb is planning its next release 15 days after that.
> Those days might be rough for students assuming that the mentors will
> be in a bug fixing/testing/documentation hurry and won't spend enough
> time with their students. On the other hand(according to the release
> schedule again), the students will have about 1 year after the summer
> perfecting their implementations before the next-next(7.x?)
> release(even though doing so is not obligatory).

I think a worse period will be right before GDB branches on May 08, when
we'll want to get the last patches in. According to the GSoC timeline,
that's more or less when the "Comunity Bonding Period" starts. So
mentors could be busy during the beginning of that period (not all of
it, though).

From what I saw up to now, GDB releases are not a big burden (except for
Joel :-) ). But perhaps this time we'll put in more effort since there
are many features going in (big breakage potential), and now we have a
bugzilla to track regressions from the previous release and try to get
them fixed. Maybe we'll have a bigger focus on stabilization in the
branch before the release...

> Thiago > 3. Support pipes in the run command (this may be too small for a GSoC);
> 
> I'm planning to spend some nights on this before GSoC starts.

Nice. Let me know if you need help. I think you won't have to mess with
ugly parts of GDB internals, so it's a good feature to start.

Have you seen the DeveloperTips page on the GDB wiki? It links to a
document from Jeremy Bennett which has a good explanation of GDB.
There's also the GDB Internals document. It's very incomplete, but I
think it's still useful to read it.
-- 
[]'s
Thiago Jung Bauermann
IBM Linux Technology Center

Re: google summer of code

[Oğuz Kayral]

Hi,
The guys at Nmap preferred to apply for "Feature Creepers and Bug
Wranglers" positions in order to have some students to work on lots of
small bugs instead of one big project. I also read that, this method
was pretty successful closing about 100 bugs in one summer. A similar
approach can be taken in order to reduce the bug count in
bugzilla(current 1365!, 52 of them for 6.8)

Another thing is, the coding begins on May 23 and according to the
release schedule gdb is planning its next release 15 days after that.
Those days might be rough for students assuming that the mentors will
be in a bug fixing/testing/documentation hurry and won't spend enough
time with their students. On the other hand(according to the release
schedule again), the students will have about 1 year after the summer
perfecting their implementations before the next-next(7.x?)
release(even though doing so is not obligatory).

Thiago > 3. Support pipes in the run command (this may be too small for a GSoC);

I'm planning to spend some nights on this before GSoC starts.

Tom > I already got email off-list about this one :)

Now you have an on-list one :)

Regards,
Oguz

2009/2/26 Thiago Jung Bauermann <bauerman@br.ibm.com>
>
> Hi,
>
> I was wondering whether we want to participate in the Google Summer of
> Code. The application period for mentoring organizations is from March 9
> until March 13th, so now is a good time to decide if it will be
> beneficial for us.
>
> I think two benefits that we get from the program are: 1. possibility of
> attracting new gdb developers and 2. a helping hand for features we want
> to develop.
>
> There are some disadvantages too. From what I read, the main problem is
> that mentoring a student takes some time and effort. From the GSoC
> Mentors Wiki:
>
> "However talented the students are who you may select, they will require
> significant attention from your developers, so release schedules and the
> desire of your community to be supportive are important considerations.
>
> Many organizations have found great rewards in mentoring, but most agree
> that they underestimated the amount of time that mentoring requires.
>
> As a potential administrator you should also consider that you will be
> responsible for getting reports and forms in to the GSoC website on
> time. You should make sure that your potential mentors realize that they
> will need to take time to evaluate students at the mid and final points
> of the 12 week period.
>
> Failure to meet the reporting requirements or mentoring requirements of
> GSoC is a serious infraction, and reflects poorly on your organization.
> If you have a big release coming out during the summer that will require
> all hands, you may wish to reconsider applying."
>
> We have some ideas for projects that a student could implement in:
>
> http://sourceware.org/gdb/wiki/ProjectIdeas
>
> Some that may be worth considering for this:
>
> 1. Help with Python scripting support;
> 2. Help with reversible debugging (not sure if teawater and Michael need
> a helping hand there);
> 3. Support pipes in the run command (this may be too small for a GSoC);
> 4. Support tracepoints in gdbserver.
>
> There may be some wishlist bugs in the bugzilla that we could lump
> together with 3. to keep someone busy for the summer...
>
> Some pointers about the GSoC:
>
> Main page: http://code.google.com/soc/
> FAQ: http://code.google.com/opensource/gsoc/2009/faqs.html
> Mentors Wiki: http://gsoc-wiki.osuosl.org/index.php/Main_Page
>
> --
> []'s
> Thiago Jung Bauermann
> IBM Linux Technology Center
>

Re: google summer of code

[teawater]

Thanks Thiago, I think it's a very cool idea.

On Thu, Feb 26, 2009 at 22:25, Thiago Jung Bauermann
<bauerman@br.ibm.com> wrote:
> Hi,
>
> I was wondering whether we want to participate in the Google Summer of
> Code. The application period for mentoring organizations is from March 9
> until March 13th, so now is a good time to decide if it will be
> beneficial for us.
>
> I think two benefits that we get from the program are: 1. possibility of
> attracting new gdb developers and 2. a helping hand for features we want
> to develop.
>
> There are some disadvantages too. From what I read, the main problem is
> that mentoring a student takes some time and effort. From the GSoC
> Mentors Wiki:
>
> "However talented the students are who you may select, they will require
> significant attention from your developers, so release schedules and the
> desire of your community to be supportive are important considerations.
>
> Many organizations have found great rewards in mentoring, but most agree
> that they underestimated the amount of time that mentoring requires.
>
> As a potential administrator you should also consider that you will be
> responsible for getting reports and forms in to the GSoC website on
> time. You should make sure that your potential mentors realize that they
> will need to take time to evaluate students at the mid and final points
> of the 12 week period.
>
> Failure to meet the reporting requirements or mentoring requirements of
> GSoC is a serious infraction, and reflects poorly on your organization.
> If you have a big release coming out during the summer that will require
> all hands, you may wish to reconsider applying."
>
> We have some ideas for projects that a student could implement in:
>
> http://sourceware.org/gdb/wiki/ProjectIdeas
>
> Some that may be worth considering for this:
>
> 1. Help with Python scripting support;

It must can make a lot of people interested in.  When I told about gdb
in some place, must have a lot of people ask me if gdb support python
or not.  :)

> 2. Help with reversible debugging (not sure if teawater and Michael need
> a helping hand there);

I am not very clear the GSOC want student get .
But maybe we can let these students do some test with reverse debug
and precord,  of course write some testsuite is better. :)
They can not only test reverse debug with precord, but also test
reverse debug with some remote-stub that support reverse-debug (For
example, vmware and Simics. Maybe we need help from Jakob and
Michael.)
I think it will help gdb reverse debug very much. :)


Thanks,
Hui

> 3. Support pipes in the run command (this may be too small for a GSoC);
> 4. Support tracepoints in gdbserver.
>
> There may be some wishlist bugs in the bugzilla that we could lump
> together with 3. to keep someone busy for the summer...
>
> Some pointers about the GSoC:
>
> Main page: http://code.google.com/soc/
> FAQ: http://code.google.com/opensource/gsoc/2009/faqs.html
> Mentors Wiki: http://gsoc-wiki.osuosl.org/index.php/Main_Page
>
> --
> []'s
> Thiago Jung Bauermann
> IBM Linux Technology Center
>
>

Re: google summer of code

[Joel Brobecker]

> > Note that it would be nice if the project was focusing on the FSF GDB,
> > but if it makes things simpler for whoever submits an entry and/or
> > mentors the project, it's fine with me if the project gets attached
> > to Archer, for instance. This decision can only be at the discretion
> > of the people who spend the effort on this project, after all...
> 
> At least for the Python support, IMHO we could use the Archer repo while
> the student is working on it, and then have a requirement of having the
> feature submitted and accepted upstream for the project to be considered
> finished.

Seems very reasonable to me. I would go even farther and say that
control over what goes in the FSF tree is held by a group of persons.
So it might be risky to have inclusion of the patches in the FSF tree
as a "requirement" in the strict sense. From our limited experience
with Archer, it looks like Archer is a nice greenhouse for maturing
some new features before pushing them onto the FSF gdb. As long as
the feature is properly implemented there, I think we'll eventually
manage to incorporate it in the FSF gdb, one way or the other.

(I can be obsessed with flowers sometimes :-)

-- 
Joel

Re: google summer of code

[Thiago Jung Bauermann]

El jue, 26-02-2009 a las 09:34 -0800, Joel Brobecker escribió:
> Hey Thiago,
> 
> Nice initiative :)

Thanks!

> > I was wondering whether we want to participate in the Google Summer of
> > Code. The application period for mentoring organizations is from March 9
> > until March 13th, so now is a good time to decide if it will be
> > beneficial for us.
> 
> I think help is beneficial to the project, but only if we know we have
> GMs available for review, or at least Area Maintainers for the projects
> that we are considering entering...

Indeed. So we can discuss project ideas, and then evaluate this point.

> > There are some disadvantages too. From what I read, the main problem is
> > that mentoring a student takes some time and effort. From the GSoC
> > Mentors Wiki:
> 
> I agree. I just tutored someone whom we contracted out to give us
> a hand finishing one internal project (which was not GDB, but it was
> something more approachable in a short amount of time), and I sometimes
> felt sucked into the project. This should definitely be weighed in
> anyone's decision to mentor someone.

I'll probably volunteer to mentor someone. I just want to learn a bit
more about what's involved, and writeups from people who did it in the
past.

> > 1. Help with Python scripting support;
> 
> To me, this would be a very promising project because:
>   . The project is very exciting (I think)
>   . We have active developers
>   . Reviews seem to be done very quickly.

Agreed!

> Note that it would be nice if the project was focusing on the FSF GDB,
> but if it makes things simpler for whoever submits an entry and/or
> mentors the project, it's fine with me if the project gets attached
> to Archer, for instance. This decision can only be at the discretion
> of the people who spend the effort on this project, after all...

At least for the Python support, IMHO we could use the Archer repo while
the student is working on it, and then have a requirement of having the
feature submitted and accepted upstream for the project to be considered
finished.
-- 
[]'s
Thiago Jung Bauermann
IBM Linux Technology Center

Re: google summer of code

[Tom Tromey]

>>>>> "Thiago" == Thiago Jung Bauermann <bauerman@br.ibm.com> writes:

Thiago> I was wondering whether we want to participate in the Google
Thiago> Summer of Code.

Yes!

Thiago> The application period for mentoring organizations is from
Thiago> March 9 until March 13th, so now is a good time to decide if
Thiago> it will be beneficial for us.

I think we should apply as an organization.  If we're rejected, we can
still have students submit their proposals under the GNU
project... but this should be the fallback plan, IMO, because it
lessens the chance of any gdb proposal being approved.

Thiago> We have some ideas for projects that a student could implement in:
Thiago> http://sourceware.org/gdb/wiki/ProjectIdeas
Thiago> Some that may be worth considering for this:
Thiago> 1. Help with Python scripting support;

I already got email off-list about this one :)
I suppose there are any number of possible projects in here.

Thiago> 2. Help with reversible debugging (not sure if teawater and Michael need
Thiago> a helping hand there);
Thiago> 3. Support pipes in the run command (this may be too small for a GSoC);
Thiago> 4. Support tracepoints in gdbserver.

I'm wondering if there is any summer-sized multi-inferior project.

FWIW, I'm willing to mentor a student.

Tom

Re: google summer of code

[Joel Brobecker]

Hey Thiago,

Nice initiative :)

> I was wondering whether we want to participate in the Google Summer of
> Code. The application period for mentoring organizations is from March 9
> until March 13th, so now is a good time to decide if it will be
> beneficial for us.

I think help is beneficial to the project, but only if we know we have
GMs available for review, or at least Area Maintainers for the projects
that we are considering entering...

> There are some disadvantages too. From what I read, the main problem is
> that mentoring a student takes some time and effort. From the GSoC
> Mentors Wiki:

I agree. I just tutored someone whom we contracted out to give us
a hand finishing one internal project (which was not GDB, but it was
something more approachable in a short amount of time), and I sometimes
felt sucked into the project. This should definitely be weighed in
anyone's decision to mentor someone.

> 1. Help with Python scripting support;

To me, this would be a very promising project because:
  . The project is very exciting (I think)
  . We have active developers
  . Reviews seem to be done very quickly.

Note that it would be nice if the project was focusing on the FSF GDB,
but if it makes things simpler for whoever submits an entry and/or
mentors the project, it's fine with me if the project gets attached
to Archer, for instance. This decision can only be at the discretion
of the people who spend the effort on this project, after all...

-- 
Joel

google summer of code

[Thiago Jung Bauermann]

Hi,

I was wondering whether we want to participate in the Google Summer of
Code. The application period for mentoring organizations is from March 9
until March 13th, so now is a good time to decide if it will be
beneficial for us.

I think two benefits that we get from the program are: 1. possibility of
attracting new gdb developers and 2. a helping hand for features we want
to develop.

There are some disadvantages too. From what I read, the main problem is
that mentoring a student takes some time and effort. From the GSoC
Mentors Wiki:

"However talented the students are who you may select, they will require
significant attention from your developers, so release schedules and the
desire of your community to be supportive are important considerations.

Many organizations have found great rewards in mentoring, but most agree
that they underestimated the amount of time that mentoring requires.

As a potential administrator you should also consider that you will be
responsible for getting reports and forms in to the GSoC website on
time. You should make sure that your potential mentors realize that they
will need to take time to evaluate students at the mid and final points
of the 12 week period.

Failure to meet the reporting requirements or mentoring requirements of
GSoC is a serious infraction, and reflects poorly on your organization.
If you have a big release coming out during the summer that will require
all hands, you may wish to reconsider applying."

We have some ideas for projects that a student could implement in:

http://sourceware.org/gdb/wiki/ProjectIdeas

Some that may be worth considering for this:

1. Help with Python scripting support;
2. Help with reversible debugging (not sure if teawater and Michael need
a helping hand there);
3. Support pipes in the run command (this may be too small for a GSoC);
4. Support tracepoints in gdbserver.

There may be some wishlist bugs in the bugzilla that we could lump
together with 3. to keep someone busy for the summer...

Some pointers about the GSoC:

Main page: http://code.google.com/soc/
FAQ: http://code.google.com/opensource/gsoc/2009/faqs.html
Mentors Wiki: http://gsoc-wiki.osuosl.org/index.php/Main_Page

-- 
[]'s
Thiago Jung Bauermann
IBM Linux Technology Center