Hi!
SBOM - Software Bill of Materials - often comes up in discussions in my projects. There’s a new working group in the IETF working on it and several other standardization bodies.
A starting point is identification of the license in each source code file with a parseable SPDX identifier.
- Is anyone against adding that to our source code? - Would it be beneficial for packaging in any way?
I think at some point in the future, a SBOM list in <pick format> will be included in packages, in order to be able to produce a SBOM for the container or the machine.
As we have multiple licenses in the source code it’s important to mark every file correctly.
I can start experimenting with http_client, then work myself around, if the dev community doesn’t scream and argue that it’s a bad thing (TM).
Read more here - SPDX - a linux foundation project ans ISO standard - https://spdx.dev - Tags in source code - https://spdx.dev/ids/
Cheers, /O
Hello,
I have nothing against it, just it should be done for the whole project (i.e., all files) in the repository if somebody decides to do it. Otherwise, we will end up with partial information, which might be misleading to some people rely on the identifier.
I know a bit about the SPDX standard, it sounds reasonable for me and its only one line added per file, so not much overhead.
Cheers,
Henning
-----Original Message----- From: sr-dev sr-dev-bounces@lists.kamailio.org On Behalf Of Olle E. Johansson Sent: Tuesday, August 16, 2022 10:43 AM To: Kamailio (SER) - Development Mailing List sr-dev@lists.kamailio.org Subject: [sr-dev] SPDX identifiers in source code
Hi!
SBOM - Software Bill of Materials - often comes up in discussions in my projects. There’s a new working group in the IETF working on it and several other standardization bodies.
A starting point is identification of the license in each source code file with a parseable SPDX identifier.
- Is anyone against adding that to our source code? - Would it be beneficial for packaging in any way?
I think at some point in the future, a SBOM list in <pick format> will be included in packages, in order to be able to produce a SBOM for the container or the machine.
As we have multiple licenses in the source code it’s important to mark every file correctly.
I can start experimenting with http_client, then work myself around, if the dev community doesn’t scream and argue that it’s a bad thing (TM).
Read more here - SPDX - a linux foundation project ans ISO standard - https://spdx.dev - Tags in source code - https://spdx.dev/ids/
Cheers, /O _______________________________________________ Kamailio (SER) - Development Mailing List sr-dev@lists.kamailio.org https://lists.kamailio.org/cgi-bin/mailman/listinfo/sr-dev
On 16 Aug 2022, at 14:53, Henning Westerholt hw@gilawa.com wrote:
Hello,
I have nothing against it, just it should be done for the whole project (i.e., all files) in the repository if somebody decides to do it. Otherwise, we will end up with partial information, which might be misleading to some people rely on the identifier.
Absolutely. It will affect all files but we don’t have to mark them all overnight, but can do it in a period between releases.
I know a bit about the SPDX standard, it sounds reasonable for me and its only one line added per file, so not much overhead.
Great. Are you aware of any good tools that parse and produce some interesting output?
Thank you for the feedback. /O
Cheers,
Henning
-----Original Message----- From: sr-dev sr-dev-bounces@lists.kamailio.org On Behalf Of Olle E. Johansson Sent: Tuesday, August 16, 2022 10:43 AM To: Kamailio (SER) - Development Mailing List sr-dev@lists.kamailio.org Subject: [sr-dev] SPDX identifiers in source code
Hi!
SBOM - Software Bill of Materials - often comes up in discussions in my projects. There’s a new working group in the IETF working on it and several other standardization bodies.
A starting point is identification of the license in each source code file with a parseable SPDX identifier.
- Is anyone against adding that to our source code?
- Would it be beneficial for packaging in any way?
I think at some point in the future, a SBOM list in <pick format> will be included in packages, in order to be able to produce a SBOM for the container or the machine.
As we have multiple licenses in the source code it’s important to mark every file correctly.
I can start experimenting with http_client, then work myself around, if the dev community doesn’t scream and argue that it’s a bad thing (TM).
Read more here
- SPDX - a linux foundation project ans ISO standard - https://spdx.dev
- Tags in source code - https://spdx.dev/ids/
Cheers, /O _______________________________________________ Kamailio (SER) - Development Mailing List sr-dev@lists.kamailio.org https://lists.kamailio.org/cgi-bin/mailman/listinfo/sr-dev
Hello Olle,
yes, it should be finalized in a shorter period and not going into multiple release cycles. If we have done it, it should be also probably added to the contribution guidelines e.g., for new modules.
Not sure what kind of tools you are referring to. For the addition to the source code, this could be probably done with some scripting and sed or similar tools.
I looked again to the SPDX standard. This is actually quite extensive and there are many fields that could be added. Example: https://github.com/spdx/spdx-spec/blob/development/v2.2.2/examples/SPDXTagEx...
Additionally, it can be added as tags (as above), XML, JSON etc.. This has grown quite a lot since I have last investigated it.
You were not proposing to add multiple tags to the source code files, right? If yes, this should be discussed in a larger round, maybe in some online developer meeting or similar. It would be difficult to maintain, if we compare e.g., to the Doxygen topic which was done for many modules, but not all parts of the code.
Cheers,
Henning
On 17 Aug 2022, at 11:15, Henning Westerholt hw@gilawa.com wrote:
Hello Olle,
yes, it should be finalized in a shorter period and not going into multiple release cycles. If we have done it, it should be also probably added to the contribution guidelines e.g., for new modules.
Right
Not sure what kind of tools you are referring to. For the addition to the source code, this could be probably done with some scripting and sed or similar tools.
Tools that generate an overview of the product. I’ll read on here. But tagging source is a starting point.
I looked again to the SPDX standard. This is actually quite extensive and there are many fields that could be added. Example: https://github.com/spdx/spdx-spec/blob/development/v2.2.2/examples/SPDXTagEx...
I think that’s the generated product, the SBOM, not what you add to the source code. The tag from the source is here: LicenseInfoInFile: GPL-2.0-only
(row 49 in that document)
Additionally, it can be added as tags (as above), XML, JSON etc.. This has grown quite a lot since I have last investigated it.
You were not proposing to add multiple tags to the source code files, right? If yes, this should be discussed in a larger round, maybe in some online developer meeting or similar. It would be difficult to maintain, if we compare e.g., to the Doxygen topic which was done for many modules, but not all parts of the code.
No, I was just proposing to add a single tag alongside the license text.
There are tools for scanning files and creating base info for debian packages, if I understand it right. At some point, I suspect it will happen.
We have an issue there with files under different licenses, but a combined license for the running software, but that is another issue to handle when that problem comes up.
/O
Cheers,
Henning
-- Henning Westerholt – https://skalatan.de/blog/ Kamailio services – https://gilawa.com
-----Original Message----- From: Olle E. Johansson oej@edvina.net Sent: Tuesday, August 16, 2022 3:44 PM To: Henning Westerholt hw@gilawa.com Cc: Kamailio (SER) - Development Mailing List sr-dev@lists.kamailio.org Subject: Re: [sr-dev] SPDX identifiers in source code
On 16 Aug 2022, at 14:53, Henning Westerholt hw@gilawa.com wrote:
Hello,
I have nothing against it, just it should be done for the whole project (i.e., all files) in the repository if somebody decides to do it. Otherwise, we will end up with partial information, which might be misleading to some people rely on the identifier.
Absolutely. It will affect all files but we don’t have to mark them all overnight, but can do it in a period between releases.
I know a bit about the SPDX standard, it sounds reasonable for me and its only one line added per file, so not much overhead.
Great. Are you aware of any good tools that parse and produce some interesting output?
Thank you for the feedback. /O
Cheers,
Henning
-----Original Message----- From: sr-dev sr-dev-bounces@lists.kamailio.org On Behalf Of Olle E. Johansson Sent: Tuesday, August 16, 2022 10:43 AM To: Kamailio (SER) - Development Mailing List sr-dev@lists.kamailio.org Subject: [sr-dev] SPDX identifiers in source code
Hi!
SBOM - Software Bill of Materials - often comes up in discussions in my projects. There’s a new working group in the IETF working on it and several other standardization bodies.
A starting point is identification of the license in each source code file with a parseable SPDX identifier.
- Is anyone against adding that to our source code?
- Would it be beneficial for packaging in any way?
I think at some point in the future, a SBOM list in <pick format> will be included in packages, in order to be able to produce a SBOM for the container or the machine.
As we have multiple licenses in the source code it’s important to mark every file correctly.
I can start experimenting with http_client, then work myself around, if the dev community doesn’t scream and argue that it’s a bad thing (TM).
Read more here
- SPDX - a linux foundation project ans ISO standard - https://spdx.dev
- Tags in source code - https://spdx.dev/ids/
Cheers, /O _______________________________________________ Kamailio (SER) - Development Mailing List sr-dev@lists.kamailio.org https://lists.kamailio.org/cgi-bin/mailman/listinfo/sr-dev
-
Henning Westerholt -
Olle E. Johansson