non-versioned packages

RFC New Alex Suykov 5 participants 11 messages Sep 30, 2015 – Jan 25, 2016

Share & tools

Alex Suykov

author

2015-09-30 21:12 UTC #0 lore raw

RFC 1/2 non-versioned packages

In case (PKG)_VERSION is empty, do not attempt to download sources for the package. Attempts to download (package)-undefined.tar.gz are pointless, the file is not likely to exist.

Also, output/build/(package) looks better and makes more sense than output/build/(package)-undefined for such packages.

Non-versioned packages within buildroot: skeleton and mke2img.

Signed-off-by Alex Suykov <[email protected]>

---
 package/pkg-generic.mk | 6 ++----
 package/pkg-utils.mk   | 2 +-
 2 files changed, 3 insertions(+), 5 deletions(-)

The actual case the prompted me to do the change: I'm using buildroot as a test rig for building and running HEAD shapshots of a local git repo. "HEAD" is not a proper version, but without these changes I could only get "package-unknown" or "package-HEAD", and it took some effort to prevent download attempts.

I'm not sure how common this is, but it's possible. And since no proper downloadable package should ever come without a version, changing the way empty VERSION is handled should affect them.

package/pkg-generic.mk +2 -4

index 5201fca..cb5545b 100644
@@ -332,8 +332,6 @@ ifndef $(2)_VERSION
   $(2)_DL_VERSION := $$($(3)_DL_VERSION)
  else ifdef $(3)_VERSION
   $(2)_DL_VERSION := $$($(3)_VERSION)
- else
-  $(2)_DL_VERSION = undefined
  endif
 else
  $(2)_DL_VERSION := $$(strip $$($(2)_VERSION))

@@ -344,7 +342,7 @@ ifdef $(3)_OVERRIDE_SRCDIR
   $(2)_OVERRIDE_SRCDIR ?= $$($(3)_OVERRIDE_SRCDIR)
 endif
 
-$(2)_BASE_NAME	=  $(1)-$$($(2)_VERSION)
+$(2)_BASE_NAME	=  $$(if $$($(2)_VERSION),$(1)-$$($(2)_VERSION),$(1))
 $(2)_DL_DIR	=  $$(DL_DIR)/$$($(2)_BASE_NAME)
 $(2)_DIR	=  $$(BUILD_DIR)/$$($(2)_BASE_NAME)

@@ -374,7 +372,7 @@ endif
 ifndef $(2)_SOURCE
  ifdef $(3)_SOURCE
   $(2)_SOURCE = $$($(3)_SOURCE)
- else
+ else ifdef $(2)_VERSION
   $(2)_SOURCE			?= $$($(2)_RAWNAME)-$$($(2)_VERSION).tar.gz
  endif
 endif

package/pkg-utils.mk +1 -1

index 44bd2c9..c098b55 100644
@@ -83,7 +83,7 @@ INFLATE.tar  = cat
 suitable-extractor = $(INFLATE$(suffix $(1)))
 
 # MESSAGE Macro -- display a message in bold type
-MESSAGE = echo "$(TERM_BOLD)>>> $($(PKG)_NAME) $($(PKG)_VERSION) $(call qstrip,$(1))$(TERM_RESET)"
+MESSAGE = echo "$(TERM_BOLD)>>> $($(PKG)_BASE_NAME) $(call qstrip,$(1))$(TERM_RESET)"
 TERM_BOLD := $(shell tput smso)
 TERM_RESET := $(shell tput rmso)

Signature · 1 line

2.0.3

Alex Suykov

author

2015-09-30 21:13 UTC #1 lore raw

RFC 2/2 remove download suppression from mke2img and skeleton

With non-versioned packages patch applied, there is no need to set empty _SOURCE.

Signed-off-by Alex Suykov <[email protected]>

---
 package/mke2img/mke2img.mk   | 1 -
 package/skeleton/skeleton.mk | 3 ---
 2 files changed, 4 deletions(-)

package/mke2img/mke2img.mk +0 -1

index 04aaa8f..9de387a 100644
@@ -4,7 +4,6 @@
 #
 ################################################################################
 
-HOST_MKE2IMG_SOURCE =
 HOST_MKE2IMG_DEPENDENCIES = host-genext2fs host-e2fsprogs
 
 define HOST_MKE2IMG_INSTALL_CMDS

package/skeleton/skeleton.mk +0 -3

index 48e7085..c182b54 100644
@@ -4,9 +4,6 @@
 #
 ################################################################################
 
-# source included in buildroot
-SKELETON_SOURCE =
-
 # The skeleton can't depend on the toolchain, since all packages depends on the
 # skeleton and the toolchain is a target package, as is skeleton.
 # Hence, skeleton would depends on the toolchain and the toolchain would depend

Signature · 1 line

2.0.3

Yann E. MORIN

2016-01-19 22:33 UTC #2 lore raw

RFC 2/2 remove download suppression from mke2img and skeleton

Alex, All,

On 2015-10-01 00:13 +0300, Alex Suykov spake thusly:

Quoted 4 lines

With non-versioned packages patch applied, there is no need to set empty _SOURCE.
Signed-off-by Alex Suykov <[email protected]>

After the previous patch has been applied:

Acked-by "Yann E. MORIN" <[email protected]>

Regards,
Yann E. MORIN.

Quoted 38 lines

---
 package/mke2img/mke2img.mk   | 1 -
 package/skeleton/skeleton.mk | 3 ---
 2 files changed, 4 deletions(-)

package/mke2img/mke2img.mk +0 -1

index 04aaa8f..9de387a 100644
@@ -4,7 +4,6 @@
 #
 ################################################################################
 
-HOST_MKE2IMG_SOURCE =
 HOST_MKE2IMG_DEPENDENCIES = host-genext2fs host-e2fsprogs
 
 define HOST_MKE2IMG_INSTALL_CMDS

package/skeleton/skeleton.mk +0 -3

index 48e7085..c182b54 100644
@@ -4,9 +4,6 @@
 #
 ################################################################################
 
-# source included in buildroot
-SKELETON_SOURCE =
-
 # The skeleton can't depend on the toolchain, since all packages depends on the
 # skeleton and the toolchain is a target package, as is skeleton.
 # Hence, skeleton would depends on the toolchain and the toolchain would depend

Signature · 6 lines

2.0.3

_______________________________________________
buildroot mailing list
buildroot at busybox.net
http://lists.busybox.net/mailman/listinfo/buildroot

Signature · 6 lines

.-----------------.--------------------.------------------.--------------------.
|  Yann E. MORIN  | Real-Time Embedded | /"\ ASCII RIBBON | Erics' conspiracy: |
| +33 662 376 056 | Software  Designer | \ / CAMPAIGN     |  ___               |
| +33 223 225 172 `------------.-------:  X  AGAINST      |  \e/  There is no  |
| http://ymorin.is-a-geek.org/ | _/*\_ | / \ HTML MAIL    |   v   conspiracy.  |
'------------------------------^-------^------------------^--------------------'

Thomas De Schampheleire

2016-01-20 14:14 UTC #3 lore raw

RFC 2/2 remove download suppression from mke2img and skeleton

Quoted 40 lines · Yann E. MORIN

Alex, All,
On 2015-10-01 00:13 +0300, Alex Suykov spake thusly:
Quoted 4 lines
With non-versioned packages patch applied, there is no need to set empty _SOURCE.
Signed-off-by Alex Suykov <[email protected]>
After the previous patch has been applied:
Acked-by "Yann E. MORIN" <[email protected]>
Regards,
Yann E. MORIN.
Quoted 25 lines
---
 package/mke2img/mke2img.mk   | 1 -
 package/skeleton/skeleton.mk | 3 ---
 2 files changed, 4 deletions(-)
package/mke2img/mke2img.mk +0 -1
index 04aaa8f..9de387a 100644
@@ -4,7 +4,6 @@
 #
 ################################################################################
 
-HOST_MKE2IMG_SOURCE =
 HOST_MKE2IMG_DEPENDENCIES = host-genext2fs host-e2fsprogs
 
 define HOST_MKE2IMG_INSTALL_CMDS
package/skeleton/skeleton.mk +0 -0
index 48e7085..c182b54 100644
@@ -4,9 +4,6 @@
 #
 ################################################################################
 

Quoted 7 lines

package/skeleton/skeleton.mk line 9

Quoted 6 lines

package/skeleton/skeleton.mk +0 -3

-# source included in buildroot
-SKELETON_SOURCE =
-
 # The skeleton can't depend on the toolchain, since all packages depends on the
 # skeleton and the toolchain is a target package, as is skeleton.
 # Hence, skeleton would depends on the toolchain and the toolchain would depend

Question is whether it is correct that mke2img and skeleton are unversioned. Packages that are similar are makedevs, mkpasswd and getent, where the _VERSION is explicitly set to match the Buildroot version:
MAKEDEVS_VERSION = buildroot-$(BR2_VERSION)
I think that mke2img is in the same boat as these packages, and so should follow the same strategy.

/Thomas

Arnout Vandecappelle

2016-01-25 00:16 UTC #4 lore raw

RFC 2/2 remove download suppression from mke2img and skeleton

Quoted 12 lines · Thomas De Schampheleire

Quoted 11 lines · Yann E. MORIN
Alex, All,
On 2015-10-01 00:13 +0300, Alex Suykov spake thusly:
Quoted 4 lines
With non-versioned packages patch applied, there is no need to set empty _SOURCE.
Signed-off-by Alex Suykov <[email protected]>
After the previous patch has been applied:
Acked-by "Yann E. MORIN" <[email protected]>

[snip]

Quoted 8 lines

Quoted 7 lines

package/skeleton/skeleton.mk line 9

Quoted 6 lines

package/skeleton/skeleton.mk +0 -3

-# source included in buildroot
-SKELETON_SOURCE =
-
 # The skeleton can't depend on the toolchain, since all packages depends on the
 # skeleton and the toolchain is a target package, as is skeleton.
 # Hence, skeleton would depends on the toolchain and the toolchain would depend

Quoted 6 lines

Question is whether it is correct that mke2img and skeleton are unversioned. Packages that are similar are makedevs, mkpasswd and getent, where the _VERSION is explicitly set to match the Buildroot version:
MAKEDEVS_VERSION = buildroot-$(BR2_VERSION)
I think that mke2img is in the same boat as these packages, and so should follow the same strategy.

+1 to that.

Alex, will you send a v2 that sets the version of mke2img and skeleton to
buildroot-$(BR2_VERSION) instead?

Regards,
Arnout

Signature · 6 lines

Arnout Vandecappelle                          arnout at mind be
Senior Embedded Software Architect            +32-16-286500
Essensium/Mind                                http://www.mind.be
G.Geenslaan 9, 3001 Leuven, Belgium           BE 872 984 063 RPR Leuven
LinkedIn profile: http://www.linkedin.com/in/arnoutvandecappelle
GPG fingerprint:  7493 020B C7E3 8618 8DEC 222C 82EB F404 F9AC 0DDF

Yann E. MORIN

2015-09-30 22:15 UTC #5 lore raw

RFC 1/2 non-versioned packages

In reply to Alex Suykov

Alex, All,

On 2015-10-01 00:12 +0300, Alex Suykov spake thusly:

Quoted 8 lines

In case (PKG)_VERSION is empty, do not attempt to download sources for the package. Attempts to download (package)-undefined.tar.gz are pointless, the file is not likely to exist.
Also, output/build/(package) looks better and makes more sense than output/build/(package)-undefined for such packages.
Non-versioned packages within buildroot: skeleton and mke2img.

Wrong. We also have:
googlefontdirectory
gadgetfs-test
* initscripts
r+ * mcookie
cache-calibrator

[*] like skeleton and mke2img, those are internal packages. so it's normal they have no version.
[+] mcookie is now provided by util-linux, so we should probably get rid of our internal copy.

Also, there has been a new package discussed on the list today that had no version either (libsoil):

    http://lists.busybox.net/pipermail/buildroot/2015-September/140370.html
    http://lists.busybox.net/pipermail/buildroot/2015-September/140480.html
    http://lists.busybox.net/pipermail/buildroot/2015-September/140508.html

And yes, I would like to keep being able to have packages without a version.Even if upstream are not playing by the rules, it is perfectly legit that we have support for that.

Regards,
Yann E. MORIN.

Quoted 68 lines

Signed-off-by Alex Suykov <[email protected]>

---
 package/pkg-generic.mk | 6 ++----
 package/pkg-utils.mk   | 2 +-
 2 files changed, 3 insertions(+), 5 deletions(-)

I'm not sure how common this is, but it's possible. And since no proper downloadable package should ever come without a version, changing the way empty VERSION is handled should affect them.

package/pkg-generic.mk +2 -4

index 5201fca..cb5545b 100644
@@ -332,8 +332,6 @@ ifndef $(2)_VERSION
   $(2)_DL_VERSION := $$($(3)_DL_VERSION)
  else ifdef $(3)_VERSION
   $(2)_DL_VERSION := $$($(3)_VERSION)
- else
-  $(2)_DL_VERSION = undefined
  endif
 else
  $(2)_DL_VERSION := $$(strip $$($(2)_VERSION))

@@ -344,7 +342,7 @@ ifdef $(3)_OVERRIDE_SRCDIR
   $(2)_OVERRIDE_SRCDIR ?= $$($(3)_OVERRIDE_SRCDIR)
 endif
 
-$(2)_BASE_NAME	=  $(1)-$$($(2)_VERSION)
+$(2)_BASE_NAME	=  $$(if $$($(2)_VERSION),$(1)-$$($(2)_VERSION),$(1))
 $(2)_DL_DIR	=  $$(DL_DIR)/$$($(2)_BASE_NAME)
 $(2)_DIR	=  $$(BUILD_DIR)/$$($(2)_BASE_NAME)

@@ -374,7 +372,7 @@ endif
 ifndef $(2)_SOURCE
  ifdef $(3)_SOURCE
   $(2)_SOURCE = $$($(3)_SOURCE)
- else
+ else ifdef $(2)_VERSION
   $(2)_SOURCE			?= $$($(2)_RAWNAME)-$$($(2)_VERSION).tar.gz
  endif
 endif

package/pkg-utils.mk +1 -1

index 44bd2c9..c098b55 100644
@@ -83,7 +83,7 @@ INFLATE.tar  = cat
 suitable-extractor = $(INFLATE$(suffix $(1)))
 
 # MESSAGE Macro -- display a message in bold type
-MESSAGE = echo "$(TERM_BOLD)>>> $($(PKG)_NAME) $($(PKG)_VERSION) $(call qstrip,$(1))$(TERM_RESET)"
+MESSAGE = echo "$(TERM_BOLD)>>> $($(PKG)_BASE_NAME) $(call qstrip,$(1))$(TERM_RESET)"
 TERM_BOLD := $(shell tput smso)
 TERM_RESET := $(shell tput rmso)

Signature · 6 lines

2.0.3

_______________________________________________
buildroot mailing list
buildroot at busybox.net
http://lists.busybox.net/mailman/listinfo/buildroot

Signature · 6 lines

.-----------------.--------------------.------------------.--------------------.
|  Yann E. MORIN  | Real-Time Embedded | /"\ ASCII RIBBON | Erics' conspiracy: |
| +33 662 376 056 | Software  Designer | \ / CAMPAIGN     |  ___               |
| +33 223 225 172 `------------.-------:  X  AGAINST      |  \e/  There is no  |
| http://ymorin.is-a-geek.org/ | _/*\_ | / \ HTML MAIL    |   v   conspiracy.  |
'------------------------------^-------^------------------^--------------------'

Alex Suykov

author

2015-09-30 22:52 UTC #6 lore raw

RFC 1/2 non-versioned packages

Quoted 3 lines · Yann E. MORIN

And yes, I would like to keep being able to have packages without a version.Even if upstream are not playing by the rules, it is perfectly legit that we have support for that.

Sorry, bad wording in patch description. Packages with explicitly set (PKG)_SOURCE are ok, they only get build dir name change:
output/build/initscripts-undefined -> output/build/initscripts and so on.

The change is for packages that lack both _VERSION and _SOURCE. Current BR behavior is to assign (PKG)_SOURCE = (pkg)-undefined.tar.gz and try to download that, which I think is never correct. It's either explicitly-set custom SOURCE, or no sources at all.

Thomas Petazzoni

2015-10-15 20:41 UTC #7 lore raw

RFC 1/2 non-versioned packages

In reply to Alex Suykov

Peter, Arnout, Yann,

What do you think of this patch (and the following one) ? Even if the commit log isn't very clear, I believe the idea behind the patch is quite good. It is indeed a bit pointless to set <pkg>_VERSION to "undefined" when it is not defined.

Though I see that really only as a minor cosmetic improvement: besides changing the message and the name of the build directory, it doesn't change anything.

But let's take a decision on this, and not keep the patches around forever.

Thomas

Quoted 70 lines · Alex Suykov

In case (PKG)_VERSION is empty, do not attempt to download sources for the package. Attempts to download (package)-undefined.tar.gz are pointless, the file is not likely to exist.
Also, output/build/(package) looks better and makes more sense than output/build/(package)-undefined for such packages.
Non-versioned packages within buildroot: skeleton and mke2img.
Signed-off-by Alex Suykov <[email protected]>
---
 package/pkg-generic.mk | 6 ++----
 package/pkg-utils.mk   | 2 +-
 2 files changed, 3 insertions(+), 5 deletions(-)
The actual case the prompted me to do the change: I'm using buildroot as a test rig for building and running HEAD shapshots of a local git repo. "HEAD" is not a proper version, but without these changes I could only get "package-unknown" or "package-HEAD", and it took some effort to prevent download attempts.
I'm not sure how common this is, but it's possible. And since no proper downloadable package should ever come without a version, changing the way empty VERSION is handled should affect them.
package/pkg-generic.mk +2 -4
index 5201fca..cb5545b 100644
@@ -332,8 +332,6 @@ ifndef $(2)_VERSION
   $(2)_DL_VERSION := $$($(3)_DL_VERSION)
  else ifdef $(3)_VERSION
   $(2)_DL_VERSION := $$($(3)_VERSION)
- else
-  $(2)_DL_VERSION = undefined
  endif
 else
  $(2)_DL_VERSION := $$(strip $$($(2)_VERSION))
@@ -344,7 +342,7 @@ ifdef $(3)_OVERRIDE_SRCDIR
   $(2)_OVERRIDE_SRCDIR ?= $$($(3)_OVERRIDE_SRCDIR)
 endif
 
-$(2)_BASE_NAME	=  $(1)-$$($(2)_VERSION)
+$(2)_BASE_NAME	=  $$(if $$($(2)_VERSION),$(1)-$$($(2)_VERSION),$(1))
 $(2)_DL_DIR	=  $$(DL_DIR)/$$($(2)_BASE_NAME)
 $(2)_DIR	=  $$(BUILD_DIR)/$$($(2)_BASE_NAME)
 
@@ -374,7 +372,7 @@ endif
 ifndef $(2)_SOURCE
  ifdef $(3)_SOURCE
   $(2)_SOURCE = $$($(3)_SOURCE)
- else
+ else ifdef $(2)_VERSION
   $(2)_SOURCE			?= $$($(2)_RAWNAME)-$$($(2)_VERSION).tar.gz
  endif
 endif
package/pkg-utils.mk +1 -1
index 44bd2c9..c098b55 100644
@@ -83,7 +83,7 @@ INFLATE.tar  = cat
 suitable-extractor = $(INFLATE$(suffix $(1)))
 
 # MESSAGE Macro -- display a message in bold type
-MESSAGE = echo "$(TERM_BOLD)>>> $($(PKG)_NAME) $($(PKG)_VERSION) $(call qstrip,$(1))$(TERM_RESET)"
+MESSAGE = echo "$(TERM_BOLD)>>> $($(PKG)_BASE_NAME) $(call qstrip,$(1))$(TERM_RESET)"
 TERM_BOLD := $(shell tput smso)
 TERM_RESET := $(shell tput rmso)
 

Signature · 3 lines

Thomas Petazzoni, CTO, Free Electrons
Embedded Linux, Kernel and Android engineering
http://free-electrons.com

Arnout Vandecappelle

2015-10-15 22:31 UTC #8 lore raw

RFC 1/2 non-versioned packages

Quoted 6 lines · Thomas Petazzoni

Peter, Arnout, Yann,
What do you think of this patch (and the following one) ? Even if the commit log isn't very clear, I believe the idea behind the patch is quite good. It is indeed a bit pointless to set <pkg>_VERSION to "undefined" when it is not defined.

Well, I do like the current situation that in the build directory it's very
explicity that the package has no explicit version. But I don't find it very important so I can accept this change.

Quoted 3 lines

Though I see that really only as a minor cosmetic improvement: besides changing the message and the name of the build directory, it doesn't change anything.

It actually does change something: it avoids downloads of packages without
explicit _SOURCE or _VERSION. This sounds like a useful feature to me.

Regards,
Arnout

Quoted 79 lines

But let's take a decision on this, and not keep the patches around forever.
Thomas
Quoted 70 lines · Alex Suykov
In case (PKG)_VERSION is empty, do not attempt to download sources for the package. Attempts to download (package)-undefined.tar.gz are pointless, the file is not likely to exist.
Also, output/build/(package) looks better and makes more sense than output/build/(package)-undefined for such packages.
Non-versioned packages within buildroot: skeleton and mke2img.
Signed-off-by Alex Suykov <[email protected]>
---
 package/pkg-generic.mk | 6 ++----
 package/pkg-utils.mk   | 2 +-
 2 files changed, 3 insertions(+), 5 deletions(-)
The actual case the prompted me to do the change: I'm using buildroot as a test rig for building and running HEAD shapshots of a local git repo. "HEAD" is not a proper version, but without these changes I could only get "package-unknown" or "package-HEAD", and it took some effort to prevent download attempts.
I'm not sure how common this is, but it's possible. And since no proper downloadable package should ever come without a version, changing the way empty VERSION is handled should affect them.
package/pkg-generic.mk +2 -4
index 5201fca..cb5545b 100644
@@ -332,8 +332,6 @@ ifndef $(2)_VERSION
   $(2)_DL_VERSION := $$($(3)_DL_VERSION)
  else ifdef $(3)_VERSION
   $(2)_DL_VERSION := $$($(3)_VERSION)
- else
-  $(2)_DL_VERSION = undefined
  endif
 else
  $(2)_DL_VERSION := $$(strip $$($(2)_VERSION))
@@ -344,7 +342,7 @@ ifdef $(3)_OVERRIDE_SRCDIR
   $(2)_OVERRIDE_SRCDIR ?= $$($(3)_OVERRIDE_SRCDIR)
 endif
 
-$(2)_BASE_NAME	=  $(1)-$$($(2)_VERSION)
+$(2)_BASE_NAME	=  $$(if $$($(2)_VERSION),$(1)-$$($(2)_VERSION),$(1))
 $(2)_DL_DIR	=  $$(DL_DIR)/$$($(2)_BASE_NAME)
 $(2)_DIR	=  $$(BUILD_DIR)/$$($(2)_BASE_NAME)
 
@@ -374,7 +372,7 @@ endif
 ifndef $(2)_SOURCE
  ifdef $(3)_SOURCE
   $(2)_SOURCE = $$($(3)_SOURCE)
- else
+ else ifdef $(2)_VERSION
   $(2)_SOURCE			?= $$($(2)_RAWNAME)-$$($(2)_VERSION).tar.gz
  endif
 endif
package/pkg-utils.mk +1 -1
index 44bd2c9..c098b55 100644
@@ -83,7 +83,7 @@ INFLATE.tar  = cat
 suitable-extractor = $(INFLATE$(suffix $(1)))
 
 # MESSAGE Macro -- display a message in bold type
-MESSAGE = echo "$(TERM_BOLD)>>> $($(PKG)_NAME) $($(PKG)_VERSION) $(call qstrip,$(1))$(TERM_RESET)"
+MESSAGE = echo "$(TERM_BOLD)>>> $($(PKG)_BASE_NAME) $(call qstrip,$(1))$(TERM_RESET)"
 TERM_BOLD := $(shell tput smso)
 TERM_RESET := $(shell tput rmso)
 

Signature · 6 lines

Arnout Vandecappelle                          arnout at mind be
Senior Embedded Software Architect            +32-16-286500
Essensium/Mind                                http://www.mind.be
G.Geenslaan 9, 3001 Leuven, Belgium           BE 872 984 063 RPR Leuven
LinkedIn profile: http://www.linkedin.com/in/arnoutvandecappelle
GPG fingerprint:  7493 020B C7E3 8618 8DEC 222C 82EB F404 F9AC 0DDF

Yann E. MORIN

2016-01-19 22:32 UTC #9 lore raw

RFC 1/2 non-versioned packages

In reply to Alex Suykov

Alex, All,

On 2015-10-01 00:12 +0300, Alex Suykov spake thusly:

Quoted 10 lines

In case (PKG)_VERSION is empty, do not attempt to download sources for the package. Attempts to download (package)-undefined.tar.gz are pointless, the file is not likely to exist.
Also, output/build/(package) looks better and makes more sense than output/build/(package)-undefined for such packages.
Non-versioned packages within buildroot: skeleton and mke2img.
Signed-off-by Alex Suykov <[email protected]>

Tested-by "Yann E. MORIN" <[email protected]>

Reviewed-by "Yann E. MORIN" <[email protected]>

However, I'm still not convinced that the commit log is correct. What about something like;

core: do not attempt to download when no _VERSION and no _SOURCE

Currently, when the package defines no _SOURCE and no _VERSION variables, we end up trying to download <PACKAGE>-undefined.tar.gz.

This is utterly pointless.

Change the logic to not attempt a download when neither _SOURCE nor _VERSION is set:
- do not set an indefined _VERSION to "undefined"
- do not append an empty (or "undefined") version to the package rawname
- use the package rawname in the <ESSAGE macro

Consequently, for packages that have no _VERSION, the build directory will be just the package name (which to some may look nicer than the current "package-undefined").

Furthermore, the message trace is also a bit leaner (to the same people that find "package" nicer than "package-undefined"). ;-)

Signed-off-by You

Tested-by "Yann E. MORIN" <[email protected]>

Reviewed-by "Yann E. MORIN" <[email protected]>

Regards,
Yann E. MORIN.

Quoted 67 lines

---
 package/pkg-generic.mk | 6 ++----
 package/pkg-utils.mk   | 2 +-
 2 files changed, 3 insertions(+), 5 deletions(-)

I'm not sure how common this is, but it's possible. And since no proper downloadable package should ever come without a version, changing the way empty VERSION is handled should affect them.

package/pkg-generic.mk +2 -4

index 5201fca..cb5545b 100644
@@ -332,8 +332,6 @@ ifndef $(2)_VERSION
   $(2)_DL_VERSION := $$($(3)_DL_VERSION)
  else ifdef $(3)_VERSION
   $(2)_DL_VERSION := $$($(3)_VERSION)
- else
-  $(2)_DL_VERSION = undefined
  endif
 else
  $(2)_DL_VERSION := $$(strip $$($(2)_VERSION))

@@ -344,7 +342,7 @@ ifdef $(3)_OVERRIDE_SRCDIR
   $(2)_OVERRIDE_SRCDIR ?= $$($(3)_OVERRIDE_SRCDIR)
 endif
 
-$(2)_BASE_NAME	=  $(1)-$$($(2)_VERSION)
+$(2)_BASE_NAME	=  $$(if $$($(2)_VERSION),$(1)-$$($(2)_VERSION),$(1))
 $(2)_DL_DIR	=  $$(DL_DIR)/$$($(2)_BASE_NAME)
 $(2)_DIR	=  $$(BUILD_DIR)/$$($(2)_BASE_NAME)

@@ -374,7 +372,7 @@ endif
 ifndef $(2)_SOURCE
  ifdef $(3)_SOURCE
   $(2)_SOURCE = $$($(3)_SOURCE)
- else
+ else ifdef $(2)_VERSION
   $(2)_SOURCE			?= $$($(2)_RAWNAME)-$$($(2)_VERSION).tar.gz
  endif
 endif

package/pkg-utils.mk +1 -1

index 44bd2c9..c098b55 100644
@@ -83,7 +83,7 @@ INFLATE.tar  = cat
 suitable-extractor = $(INFLATE$(suffix $(1)))
 
 # MESSAGE Macro -- display a message in bold type
-MESSAGE = echo "$(TERM_BOLD)>>> $($(PKG)_NAME) $($(PKG)_VERSION) $(call qstrip,$(1))$(TERM_RESET)"
+MESSAGE = echo "$(TERM_BOLD)>>> $($(PKG)_BASE_NAME) $(call qstrip,$(1))$(TERM_RESET)"
 TERM_BOLD := $(shell tput smso)
 TERM_RESET := $(shell tput rmso)

Signature · 6 lines

2.0.3

_______________________________________________
buildroot mailing list
buildroot at busybox.net
http://lists.busybox.net/mailman/listinfo/buildroot

Signature · 6 lines

.-----------------.--------------------.------------------.--------------------.
|  Yann E. MORIN  | Real-Time Embedded | /"\ ASCII RIBBON | Erics' conspiracy: |
| +33 662 376 056 | Software  Designer | \ / CAMPAIGN     |  ___               |
| +33 223 225 172 `------------.-------:  X  AGAINST      |  \e/  There is no  |
| http://ymorin.is-a-geek.org/ | _/*\_ | / \ HTML MAIL    |   v   conspiracy.  |
'------------------------------^-------^------------------^--------------------'

Alex Suykov

author

2016-01-24 22:33 UTC #10 lore raw

RFC 1/2 non-versioned packages

Quoted 3 lines · Yann E. MORIN

However, I'm still not convinced that the commit log is correct. What about something like;
(...)

No problem, text updated.
And the patch itself updated to apply cleanly onto current master (nothing significant, just line numbers).

I'm sending it as PATCH, but without v2 because the original message was RFC.