Bug#846366: ITP: bcc -- Command line tools for BPF Compiler Collection (BCC)

[Sébastien Delafond]

On Dec/29, Ritesh Raj Sarraf wrote:
> I've just pushed my changes to the git repo. Could you please review
> it once ?  I'd like you to have your comments/feedback before we
> decide on uploading it.
> 
> Apart from the main file name change, there are other minor changes.

It all looks good to me.

> PS: If you want to, please review and upload. Otherwise, if you want
> me to do it, please let me know.

I've pushed a change removing myself from the Uploaders field: it's
basically your package now :) Feel free to upload whenever you like !

Cheers,

--Seb

Bug#846366: ITP: bcc -- Command line tools for BPF Compiler Collection (BCC)

[Ritesh Raj Sarraf]

-BEGIN PGP SIGNED MESSAGE-
Hash: SHA512

Hello Sebastien,

I've just pushed my changes to the git repo. Could you please review it once ?
I'd like you to have your comments/feedback before we decide on uploading it.

Apart from the main file name change, there are other minor changes.


PS: If you want to, please review and upload. Otherwise, if you want me to do
it, please let me know.

Thanks,
Ritesh

On Thu, 2016-12-29 at 15:17 +0530, Ritesh Raj Sarraf wrote:
> On Thu, 2016-12-29 at 10:41 +0100, Sébastien Delafond wrote:
> > > So, unless there is a concern, I'd want to target this change for the
> > > next upload of both the tools. The bpfcc follow-up upload is pending
> > > because of a FTBFS bug. And on the perf-tools side, there haven't been
> > > any substantial changes lately, warranting an upload.
> > > 
> > > But whatever triggers the upload, we'll make this change included ?
> > 
> > Sounds good to me. Maybe worth documenting in README.Debian as well ?
> 
> Yes. Will do. Thanks for mentioning.
> 
- -- 
Ritesh Raj Sarraf | http://people.debian.org/~rrs
Debian - The Universal Operating System
-BEGIN PGP SIGNATURE-
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=ZgxB
-END PGP SIGNATURE-

Bug#846366: ITP: bcc -- Command line tools for BPF Compiler Collection (BCC)

[Ritesh Raj Sarraf]

-BEGIN PGP SIGNED MESSAGE-
Hash: SHA512

On Thu, 2016-12-29 at 10:41 +0100, Sébastien Delafond wrote:
> > So, unless there is a concern, I'd want to target this change for the
> > next upload of both the tools. The bpfcc follow-up upload is pending
> > because of a FTBFS bug. And on the perf-tools side, there haven't been
> > any substantial changes lately, warranting an upload.
> > 
> > But whatever triggers the upload, we'll make this change included ?
> 
> Sounds good to me. Maybe worth documenting in README.Debian as well ?

Yes. Will do. Thanks for mentioning.

- -- 
Ritesh Raj Sarraf | http://people.debian.org/~rrs
Debian - The Universal Operating System
-BEGIN PGP SIGNATURE-
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=5P2H
-END PGP SIGNATURE-

Bug#846366: ITP: bcc -- Command line tools for BPF Compiler Collection (BCC)

[Sébastien Delafond]

On Dec/29, Ritesh Raj Sarraf wrote:
> I think we should stick with this proposal of appending the type along
> with the name.
> 
> 1. On autocompletions, it'd autocomplete to "execsnoop-", which is an
> invalid name either way. This will expect the user to pay attention
> and fire the correct command.
> 
> 2. This approach is explicit, visible and allows for co-existence for
> both.
> 
> So, unless there is a concern, I'd want to target this change for the
> next upload of both the tools. The bpfcc follow-up upload is pending
> because of a FTBFS bug. And on the perf-tools side, there haven't been
> any substantial changes lately, warranting an upload.
> 
> But whatever triggers the upload, we'll make this change included ?

Sounds good to me. Maybe worth documenting in README.Debian as well ?

Cheers,

--Seb

Bug#846366: ITP: bcc -- Command line tools for BPF Compiler Collection (BCC)

[Ritesh Raj Sarraf]

-BEGIN PGP SIGNED MESSAGE-
Hash: SHA512

On Wed, 2016-11-30 at 23:18 +0530, Ritesh Raj Sarraf wrote:
> The binary file names are conflicting for bcc and perf-tools-unstable. I don't
> see a reason why one cannot co-install both the packages and use. But in its
> current form, it'll fail complaining file overwrites.
> 
> Personally, I don't think the alternatives route may be of much use because
> apart from this 2 packages, I don't see any other package using these names.
> Also, IMO, the alternatives feature is usually useful for common tools like
> editor, pager etc.
> 
> One option could be that we append the names of the binaries explicitly.
> Example, for bcc => execsnoop-bcc, perf-tools => execsnoop-perf
> 
> What do you say ? Or if you have any suggestions, please do mention.

I think we should stick with this proposal of appending the type along with the
name.

1. On autocompletions, it'd autocomplete to "execsnoop-", which is an invalid
name either way. This will expect the user to pay attention and fire the correct
command.

2. This approach is explicit, visible and allows for co-existence for both.


So, unless there is a concern, I'd want to target this change for the next
upload of both the tools. The bpfcc follow-up upload is pending because of a
FTBFS bug. And on the perf-tools side, there haven't been any substantial
changes lately, warranting an upload.

But whatever triggers the upload, we'll make this change included ?

- -- 
Ritesh Raj Sarraf | http://people.debian.org/~rrs
Debian - The Universal Operating System
-BEGIN PGP SIGNATURE-
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=zgA+
-END PGP SIGNATURE-

Bug#846366: ITP: bcc -- Command line tools for BPF Compiler Collection (BCC)

[Ritesh Raj Sarraf]

-BEGIN PGP SIGNED MESSAGE-
Hash: SHA512

Hi Sebastien,

On Thu, 2016-12-01 at 08:25 +0100, Sébastien Delafond wrote:
> Hi Ritesh,
> 
> I agree with you, there is no reason we can't coexist :)
> 
> However, perf-tools-unstable doesn't seem to be much more updated these
> days, and it sorta worries me, especially since Brendan Gregg mentions
> on his blog that bcc seems to be the future: in that light, do you still
> see a need for perf-tools-unstable at all ?
> 

Yes. I saw that too. Maybe you'd want to check with him first. If there are no
plans on maintaining it, then perhaps we should discontinue it.
I'd be surprised though, given his tools rely on ftrace, that he'd want to
discontinue them.

> If you do, would you be willing to take over its maintainership ? I'm
> trying to save more time to contribute to security work in Debian, so
> you'd be welcome to do that :)
> 

Yes. I can do that. But please, first lets start an email with Brendan (please
CC me too) and clarify on perf-tools-unstable's roadmap.

> Anyway, let me know what you think and we'll take it from there.

For now, I'll just mark perf-tools-unstable as a conflict against this package.
Later, once we've concluded with upstream, we can decide on renaming, or maybe
even merging both sources together, into a single package name.


- -- 
Ritesh Raj Sarraf | http://people.debian.org/~rrs
Debian - The Universal Operating System
-BEGIN PGP SIGNATURE-
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=FMuo
-END PGP SIGNATURE-

Bug#846366: ITP: bcc -- Command line tools for BPF Compiler Collection (BCC)

[Ritesh Raj Sarraf]

-BEGIN PGP SIGNED MESSAGE-
Hash: SHA512

On Wed, 2016-11-30 at 22:40 +0100, Christian Seiler wrote:
> > many thanks for the explanation, so from a technical point of
> > view there is no package naming conflict, although it is somewhat
> > counter-intuitive to end up with a source-package "bcc" and a
> > binary-package "bcc" where the latter isn't built from the former
> > but instead contains a completely different application.
> 
> Maybe the new source package could be named bpf-bcc? That way there
> would be no confusion with respect to bin:bcc vs. src:bcc, and the
> source package name is still quite short, yet descriptive. Just a
> suggestion.

How about ?
src: bpfcc
bin: bpfcc-tools, libbpfcc, libbpfcc-dev, bpfcc-lua, python-bpfcc

This relates well to what the tool is: BPF Compiler Collection, both in src and
binary names.

I think I'll stick with this name unless there are concerns.


Thanks.

- -- 
Ritesh Raj Sarraf | http://people.debian.org/~rrs
Debian - The Universal Operating System
-BEGIN PGP SIGNATURE-
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=9yGo
-END PGP SIGNATURE-

Bug#846366: ITP: bcc -- Command line tools for BPF Compiler Collection (BCC)

[Sébastien Delafond]

Hi Ritesh,

I agree with you, there is no reason we can't coexist :)

However, perf-tools-unstable doesn't seem to be much more updated these
days, and it sorta worries me, especially since Brendan Gregg mentions
on his blog that bcc seems to be the future: in that light, do you still
see a need for perf-tools-unstable at all ?

If you do, would you be willing to take over its maintainership ? I'm
trying to save more time to contribute to security work in Debian, so
you'd be welcome to do that :)

Anyway, let me know what you think and we'll take it from there.

Cheers,

--Seb

On Nov/30, Ritesh Raj Sarraf wrote:
> -BEGIN PGP SIGNED MESSAGE-
> Hash: SHA512
> 
> Hello Sebastien,
> 
> First of all, thanks for maintaining perf-tools-unstable. I use it many a 
> times.
> 
> I just completed the major chunk of bcc packaging. I intend to maintain it 
> under
> collab-maint/ too.
> 
> 
> The binary file names are conflicting for bcc and perf-tools-unstable. I don't
> see a reason why one cannot co-install both the packages and use. But in its
> current form, it'll fail complaining file overwrites.
> 
> Personally, I don't think the alternatives route may be of much use because
> apart from this 2 packages, I don't see any other package using these names.
> Also, IMO, the alternatives feature is usually useful for common tools like
> editor, pager etc.
> 
> One option could be that we append the names of the binaries explicitly.
> Example, for bcc => execsnoop-bcc, perf-tools => execsnoop-perf
> 
> What do you say ? Or if you have any suggestions, please do mention.
> 
> Thanks,
> Ritesh
> 
> On Wed, 2016-11-30 at 22:50 +0530, Ritesh Raj Sarraf wrote:
> > Package: wnpp
> > Severity: wishlist
> > Owner: Ritesh Raj Sarraf 
> > 
> > * Package name: bcc
> >   Version : 0.2.0
> >   Upstream Author : IO Visor Project (https://github.com/iovisor)
> > * URL : https://github.com/iovisor/bcc
> > * License : Apache 2.0
> >   Programming Lang: C, Python
> >   Description : Tools for BPF Compiler Collection (BCC)
> > 
> >  BPF Compiler Collection (BCC) is a toolkit for creating efficient
> >  kernel tracing and manipulation programs
> >  .
> >  It makes use of extended BPF (Berkeley Package Filter) and provides tools
> >  for BPF based Linux IO analysis, networking, monitoring and more
> > 
> > This is a great tool to debug and instrument your kernel and
> > applications. It also is a a performance characterization and analysis
> > tool
> > 
> > I have an interest in Debugging and Analysis and I intend to use and
> > maintain it under Debian.
> > 
> > As with most of my packages, I intend to maintain it under collab-maint
> - -- 
> Ritesh Raj Sarraf | http://people.debian.org/~rrs
> Debian - The Universal Operating System
> -BEGIN PGP SIGNATURE-
> 
> iQIzBAEBCgAdFiEEQCVDstmIVAB/Yn02pjpYo/LhdWkFAlg/EPQACgkQpjpYo/Lh
> dWle6xAAsRRfIO3oUvk6wN4NXanqQark//nvE5ezt2RPlx4SUYWA3WWJas1bbGA8
> AdnJMKvqYD1wkBmuSRU5S+Ne6kXjat8GQCPiYBi1IS1mcUQXRT/MoKce3YmxJGvt
> p6jxGBP8d9CeUZe7eOHujKmsBj4OYghX1e6VmYtwdGeaCEw4x5vnxO7GuX8ZkgxZ
> XE/zSyuKB7GLAJUbg11VGUKoLVEP0kOprflj17DsNofXFNDrETL2OykDTNNIsJTP
> c+RuTQE2xlD4EGCbnrYQ5A4tVl6ZtFao8LZXzcOmmAu9yR+6aYVMdEJnFe6iXTSY
> M74pIBLZs3gq5gkhfm0x+sVadkXx/xuG3fslFxHYJIUWXS2aaB1pryrvDKdP1SF7
> F1RzJe6JtflzkkRu0DNBvZdkOztoX+jhldeoWXjZ7qcwKNFC7COvT4piEi+ivIXb
> q3gvuadoGN4SL8M4sJJD2nbmKiDo74UuQ+HisY1xrAM6+ksf+/DxwVHOOX1Akpii
> b77uguwPf6kxmScfI0VrL7LJr9y3vOI4GeQsAqq1ttUghi59u/dZ6qGjbA3xEnTz
> a2w9kHGtwVUTIZaMT3v2nXveI87P+vHNhVDA9GyTmDvdqdVpTQp3nfCMzeCj1/L6
> 5fK78HJnQkdhMcITGVYmgcUAuegws2X1f0pw2/9jzpGWE39Sjjc=
> =YYEe
> -END PGP SIGNATURE-
> 
> 

Bug#846366: ITP: bcc -- Command line tools for BPF Compiler Collection (BCC)

[Ben Hutchings]

On Thu, 2016-12-01 at 00:56 +0530, Ritesh Raj Sarraf wrote:
> Hello Karsten,
> 
> On Wed, 2016-11-30 at 20:05 +0100, Karsten Merker wrote:
> > Hello,
> > 
> > bcc is a package (and executable) name that is already in use for
> > another program in Debian. From https://packages.debian.org/sid/bcc:
> 
> I'm aware of it. bcc is an already packaged binary package. It build from 
> source package: linux86
> 
> For this package, I've tried to be close to what upstream has already named.
> So, for Debian, only the source package name is: bcc.
[...]

Please don't do this.  When the same name is used for a binary package
and for a source package that doesn't build that binary, it tends to
result in mis-assigned bugs as BTS users don't consistently specify
which they mean.

Ben.

-- 
Ben Hutchings
A free society is one where it is safe to be unpopular. - Adlai
Stevenson



signature.asc
Description: This is a digitally signed message part

Bug#846366: ITP: bcc -- Command line tools for BPF Compiler Collection (BCC)

[Christian Seiler]

On 11/30/2016 10:32 PM, Karsten Merker wrote:
> On Thu, Dec 01, 2016 at 12:56:14AM +0530, Ritesh Raj Sarraf wrote:
>> On Wed, 2016-11-30 at 20:05 +0100, Karsten Merker wrote:
>>> bcc is a package (and executable) name that is already in use for
>>> another program in Debian. From https://packages.debian.org/sid/bcc:
>>
>> I'm aware of it. bcc is an already packaged binary package. It
>> build from source package: linux86
>>
>> For this package, I've tried to be close to what upstream has already named.
>> So, for Debian, only the source package name is: bcc.
>> The binary packages are:
>>
>> rrs@learner:~/rrs-home/Community/Packaging/bcc (master)$ grep Package: 
>> debian/control 
>> Package: libbcc
>> Package: libbcc-dev
>> Package: python-bcc
>> Package: bcc-tools
>> Package: bcc-lua
>> 2016-12-01 / 00:52:49 ♒♒♒  ☺  
>>
>> Does it make sense ?
>>
>> If you have suggestions, please mention them, because it'll be
>> easier to make the name changes now.
> 
> many thanks for the explanation, so from a technical point of
> view there is no package naming conflict, although it is somewhat
> counter-intuitive to end up with a source-package "bcc" and a
> binary-package "bcc" where the latter isn't built from the former
> but instead contains a completely different application.

Maybe the new source package could be named bpf-bcc? That way there
would be no confusion with respect to bin:bcc vs. src:bcc, and the
source package name is still quite short, yet descriptive. Just a
suggestion.

Regards,
Christian

Bug#846366: ITP: bcc -- Command line tools for BPF Compiler Collection (BCC)

[Ritesh Raj Sarraf]

-BEGIN PGP SIGNED MESSAGE-
Hash: SHA512

Hello Karsten,

On Wed, 2016-11-30 at 20:05 +0100, Karsten Merker wrote:
> Hello,
> 
> bcc is a package (and executable) name that is already in use for
> another program in Debian. From https://packages.debian.org/sid/bcc:

I'm aware of it. bcc is an already packaged binary package. It build from 
source package: linux86

For this package, I've tried to be close to what upstream has already named.
So, for Debian, only the source package name is: bcc.
The binary packages are:

rrs@learner:~/rrs-home/Community/Packaging/bcc (master)$ grep Package: 
debian/control 
Package: libbcc
Package: libbcc-dev
Package: python-bcc
Package: bcc-tools
Package: bcc-lua
2016-12-01 / 00:52:49 ♒♒♒  ☺  



Does it make sense ?
If you have suggestions, please mention them, because it'll be easier to make 
the name changes now.


- -- 
Ritesh Raj Sarraf | http://people.debian.org/~rrs
Debian - The Universal Operating System
-BEGIN PGP SIGNATURE-
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=emD/
-END PGP SIGNATURE-

Bug#846366: ITP: bcc -- Command line tools for BPF Compiler Collection (BCC)

[Ritesh Raj Sarraf]

-BEGIN PGP SIGNED MESSAGE-
Hash: SHA512

Hello Sebastien,

First of all, thanks for maintaining perf-tools-unstable. I use it many a times.

I just completed the major chunk of bcc packaging. I intend to maintain it under
collab-maint/ too.


The binary file names are conflicting for bcc and perf-tools-unstable. I don't
see a reason why one cannot co-install both the packages and use. But in its
current form, it'll fail complaining file overwrites.

Personally, I don't think the alternatives route may be of much use because
apart from this 2 packages, I don't see any other package using these names.
Also, IMO, the alternatives feature is usually useful for common tools like
editor, pager etc.

One option could be that we append the names of the binaries explicitly.
Example, for bcc => execsnoop-bcc, perf-tools => execsnoop-perf

What do you say ? Or if you have any suggestions, please do mention.

Thanks,
Ritesh

On Wed, 2016-11-30 at 22:50 +0530, Ritesh Raj Sarraf wrote:
> Package: wnpp
> Severity: wishlist
> Owner: Ritesh Raj Sarraf 
> 
> * Package name: bcc
>   Version : 0.2.0
>   Upstream Author : IO Visor Project (https://github.com/iovisor)
> * URL : https://github.com/iovisor/bcc
> * License : Apache 2.0
>   Programming Lang: C, Python
>   Description : Tools for BPF Compiler Collection (BCC)
> 
>  BPF Compiler Collection (BCC) is a toolkit for creating efficient
>  kernel tracing and manipulation programs
>  .
>  It makes use of extended BPF (Berkeley Package Filter) and provides tools
>  for BPF based Linux IO analysis, networking, monitoring and more
> 
> This is a great tool to debug and instrument your kernel and
> applications. It also is a a performance characterization and analysis
> tool
> 
> I have an interest in Debugging and Analysis and I intend to use and
> maintain it under Debian.
> 
> As with most of my packages, I intend to maintain it under collab-maint
- -- 
Ritesh Raj Sarraf | http://people.debian.org/~rrs
Debian - The Universal Operating System
-BEGIN PGP SIGNATURE-
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=YYEe
-END PGP SIGNATURE-

Bug#846366: ITP: bcc -- Command line tools for BPF Compiler Collection (BCC)

[Ritesh Raj Sarraf]

Package: wnpp
Severity: wishlist
Owner: Ritesh Raj Sarraf 

* Package name: bcc
  Version : 0.2.0
  Upstream Author : IO Visor Project (https://github.com/iovisor)
* URL : https://github.com/iovisor/bcc
* License : Apache 2.0
  Programming Lang: C, Python
  Description : Tools for BPF Compiler Collection (BCC)

 BPF Compiler Collection (BCC) is a toolkit for creating efficient
 kernel tracing and manipulation programs
 .
 It makes use of extended BPF (Berkeley Package Filter) and provides tools
 for BPF based Linux IO analysis, networking, monitoring and more

This is a great tool to debug and instrument your kernel and
applications. It also is a a performance characterization and analysis
tool

I have an interest in Debugging and Analysis and I intend to use and
maintain it under Debian.

As with most of my packages, I intend to maintain it under collab-maint