Skip to contentSkip to conversation

preferred KERNEL_FEATURES format: netfilter or netfilter.scc?

DISCUSSION Robert P. J. Day 2 participants 6 messages Mar 3, 2015

lore
Share & tools
lore
#0 lore raw

kernel dev manual uses the KERNEL_FEATURES format of:

     KERNEL_FEATURES += "features/netfilter.scc"

while ref manual variable glossary uses

KERNEL_FEATURES="features/netfilter"

is there a preference for the usage of the .scc suffix? (i'm assuming they're both valid, i didn't actually check that. maybe i should check that ...)

oh, wait, here's something from kernel-yocto.bbclass that seems to
require the presence of the .scc suffix:

# returns all the elements from the src uri that are .scc files
def find_sccs(d):
    sources=src_patches(d, True)
    sources_list=[]
    for s in sources:
        base, ext = os.path.splitext(os.path.basename(s))
        if ext and ext in [".scc", ".cfg"]:
            sources_list.append(s)
        elif base and base in 'defconfig':
            sources_list.append(s)

return sources_list

i should note that the introductory comment is incorrect since that function clearly does not just find "sccs", it finds .cfg and defconfig files as well, no?

rday

Signature · 7 lines
========================================================================
Robert P. J. Day                                 Ottawa, Ontario, CANADA
                        http://crashcourse.ca

Twitter:                                       http://twitter.com/rpjday
LinkedIn:                               http://ca.linkedin.com/in/rpjday
========================================================================
#1 lore raw
Quoted 5 lines · Robert P. J. Day

kernel dev manual uses the KERNEL_FEATURES format of:

     KERNEL_FEATURES += "features/netfilter.scc"
Quoted 7 lines

while ref manual variable glossary uses

KERNEL_FEATURES="features/netfilter"

is there a preference for the usage of the .scc suffix? (i'm assuming they're both valid, i didn't actually check that. maybe i should check that ...)

hang on, i think there's an error in the docs ... if i check the
meta branch, i see that the actual structure for the netfilter feature is "features/netfilter/netfilter.scc", so "features/netfilter.scc" shouldn't even work, should it? historical remnant?

rday

Signature · 7 lines
========================================================================
Robert P. J. Day                                 Ottawa, Ontario, CANADA
                        http://crashcourse.ca

Twitter:                                       http://twitter.com/rpjday
LinkedIn:                               http://ca.linkedin.com/in/rpjday
========================================================================
#2 lore raw
Quoted 11 lines · Robert P. J. Day
Quoted 9 lines · Robert P. J. Day

kernel dev manual uses the KERNEL_FEATURES format of:

      KERNEL_FEATURES += "features/netfilter.scc"

while ref manual variable glossary uses

KERNEL_FEATURES="features/netfilter"

Quoted 8 lines
Quoted 3 lines

is there a preference for the usage of the .scc suffix? (i'm assuming they're both valid, i didn't actually check that. maybe i should check that ...)

hang on, i think there's an error in the docs ... if i check the
meta branch, i see that the actual structure for the netfilter feature is "features/netfilter/netfilter.scc", so "features/netfilter.scc" shouldn't even work, should it? historical remnant?

It works, but it is an old shortcut that isn't emphasized or used. So yes, its a historical remnant, and explicit .scc references are preferred.

Cheers,

Bruce

Quoted 3 lines

rday

#3 lore raw
Quoted 16 lines · Bruce Ashfield
Quoted 15 lines · Robert P. J. Day
Quoted 12 lines · Robert P. J. Day

kernel dev manual uses the KERNEL_FEATURES format of:

      KERNEL_FEATURES += "features/netfilter.scc"

while ref manual variable glossary uses

KERNEL_FEATURES="features/netfilter"

is there a preference for the usage of the .scc suffix? (i'm assuming they're both valid, i didn't actually check that. maybe i should check that ...)

Quoted 8 lines
Quoted 4 lines

hang on, i think there's an error in the docs ... if i check the
meta branch, i see that the actual structure for the netfilter feature is "features/netfilter/netfilter.scc", so "features/netfilter.scc" shouldn't even work, should it? historical remnant?

It works, but it is an old shortcut that isn't emphasized or used. So yes, its a historical remnant, and explicit .scc references are preferred.

wait ... just to be clear, you're saying that referring to a
features *directory* is a historical remnant? note above that i'm asking about what appears to be a historical remnant where it looks like netfilter used to be a .scc file directly under features/, now it's in a subdirectory.

as it stands now, it would *appear* that the statement:

KERNEL_FEATURES += "features/netfilter.scc"

is simply wrong and can't possibly work, correct?

rday

Signature · 7 lines
========================================================================
Robert P. J. Day                                 Ottawa, Ontario, CANADA
                        http://crashcourse.ca

Twitter:                                       http://twitter.com/rpjday
LinkedIn:                               http://ca.linkedin.com/in/rpjday
========================================================================
#4 lore raw
Quoted 27 lines · Robert P. J. Day
Quoted 16 lines · Bruce Ashfield
Quoted 15 lines · Robert P. J. Day
Quoted 12 lines · Robert P. J. Day

kernel dev manual uses the KERNEL_FEATURES format of:

KERNEL_FEATURES += "features/netfilter.scc"

while ref manual variable glossary uses

KERNEL_FEATURES="features/netfilter"

is there a preference for the usage of the .scc suffix? (i'm assuming they're both valid, i didn't actually check that. maybe i should check that ...)

Quoted 8 lines
Quoted 4 lines

hang on, i think there's an error in the docs ... if i check the
meta branch, i see that the actual structure for the netfilter feature is "features/netfilter/netfilter.scc", so "features/netfilter.scc" shouldn't even work, should it? historical remnant?

It works, but it is an old shortcut that isn't emphasized or used. So yes, its a historical remnant, and explicit .scc references are preferred.

Quoted 5 lines

wait ... just to be clear, you're saying that referring to a
features *directory* is a historical remnant? note above that i'm asking about what appears to be a historical remnant where it looks like netfilter used to be a .scc file directly under features/, now it's in a subdirectory.

This format: KERNEL_FEATURES="features/netfilter"

Is the historical remnant. The tools can convert that to features/netfilter/netfiler.scc internally. It was a shortcut to make the features more human readable when they followed that defined form "<dirname>/<dirname>.scc"

Quoted 6 lines

as it stands now, it would *appear* that the statement:

   KERNEL_FEATURES += "features/netfilter.scc"

is simply wrong and can't possibly work, correct?

That one is wrong, but the tools can also do a basename before running the test, which converts it to that other format, and then it can be expanded to find the feature description.

Both are non-obvious, and error prone .. hence why they are not used anymore.

So the answer is that the docs should be updated to purge the old references.

Bruce

Quoted 3 lines

rday

#5 lore raw

On Tue, 3 Mar 2015, Bruce Ashfield wrote:

... snip ...

Quoted 18 lines

This format: KERNEL_FEATURES="features/netfilter"

Is the historical remnant. The tools can convert that to features/netfilter/netfiler.scc internally. It was a shortcut to make the features more human readable when they followed that defined form "<dirname>/<dirname>.scc"

Quoted 6 lines

as it stands now, it would *appear* that the statement:

   KERNEL_FEATURES += "features/netfilter.scc"

is simply wrong and can't possibly work, correct?

That one is wrong, but the tools can also do a basename before running the test, which converts it to that other format, and then it can be expanded to find the feature description.

Quoted 5 lines

Both are non-obvious, and error prone .. hence why they are not used anymore.

So the answer is that the docs should be updated to purge the old references.

thank you, that was the "A" answer.

rday

Signature · 7 lines
========================================================================
Robert P. J. Day                                 Ottawa, Ontario, CANADA
                        http://crashcourse.ca

Twitter:                                       http://twitter.com/rpjday
LinkedIn:                               http://ca.linkedin.com/in/rpjday
========================================================================