[gentoo-catalyst] rewrite-on-master branch

[gentoo-catalyst] rewrite-on-master branch

[Matt Turner]

I'm not okay with redoing the rename the branches trick. Let's use git
how it's supposed to be used.

I've rebased the origin/3.0 branch onto master. There are 81 patches in it.

All of the differences between origin/3.0 and rewrite-on-master seem
to be mistakes in the 3.0 branch.

I'm not advocating pushing this to master yet. The series needs review
and some clean up.

Re: [gentoo-catalyst] rewrite-on-master branch

[Brian Dolbec]

[-- Attachment #1: Type: text/plain, Size: 5790 bytes --]

On Fri, 2013-07-19 at 21:52 -0700, Matt Turner wrote:
> I'm not okay with redoing the rename the branches trick. Let's use git
> how it's supposed to be used.
> 
> I've rebased the origin/3.0 branch onto master. There are 81 patches in it.
> 
> All of the differences between origin/3.0 and rewrite-on-master seem
> to be mistakes in the 3.0 branch.
> 
> I'm not advocating pushing this to master yet. The series needs review
> and some clean up.
> 

No, there is no 3.0 mistakes in the diff,  a few minor differences and
the bindist and unpack_command commits that weren't in my 3.0 rewrite.

But I do agree that there is a little more to fix before pushing to
master.  I explained the differences below in the diffs.


diff --git a/catalyst/base/stagebase.py b/catalyst/base/stagebase.py
index b0144f6..8f380ab 100644
--- a/catalyst/base/stagebase.py
+++ b/catalyst/base/stagebase.py
@@ -683,7 +683,7 @@ class StageBase(TargetBase, ClearBase, GenBase):
 					unpack_cmd="tar -I lbzip2 -xpf "+self.settings["source_path"]+" -C "+\
 						self.settings["chroot_path"]
 				else:
-					unpack_cmd="tar xpf "+self.settings["source_path"]+" -C "+\
+					unpack_cmd="tar -I lbzip2 -xpf "+self.settings["source_path"]+" -C "+\
 						self.settings["chroot_path"]
 				error_msg="Tarball extraction of "+\
 					self.settings["source_path"]+" to "+\
@@ -698,7 +698,7 @@ class StageBase(TargetBase, ClearBase, GenBase):
 				unpack_cmd="tar -I lbzip2 -xpf "+self.settings["source_path"]+" -C "+\
 					self.settings["chroot_path"]
 			else:
-				unpack_cmd="tar xpf "+self.settings["source_path"]+" -C "+\
+				unpack_cmd="tar -I lbzip2 -xpf "+self.settings["source_path"]+" -C "+\
 					self.settings["chroot_path"]
 			error_msg="Tarball extraction of "+self.settings["source_path"]+\
 				" to "+self.settings["chroot_path"]+" failed."


This is from the 3 commits Zero_Chaos did which reduces the if: else: code block to the exact 
same unpack_command no matter what the if evaluates to.
No matter, the new compress.py code I have nearly done, completely rewrites this whole section.


diff --git a/catalyst/defaults.py b/catalyst/defaults.py
index d327a1b..d0def22 100644
--- a/catalyst/defaults.py
+++ b/catalyst/defaults.py
@@ -104,9 +104,9 @@ option_messages = {
 	"icecream": "Icecream compiler cluster support enabled.",
 	"kerncache": "Kernel cache support enabled.",
 	"pkgcache": "Package cache support enabled.",
+	"preserve_libs": "Preserving libs during unmerge.",
 	"purge": "Purge support enabled.",
 	"seedcache": "Seed cache support enabled.",
 	"snapcache": "Snapshot cache support enabled.",
-	'preserve_libs': "Preserving libs during unmerge.",
 	#"tarball": "Tarball creation enabled.",
 	}

This un-does a sort, where I moved 'preserve_libs' to a sorted order... meh.


diff --git a/catalyst/main.py b/catalyst/main.py
index 6f8bd8d..08c69ca 100644
--- a/catalyst/main.py
+++ b/catalyst/main.py
@@ -116,6 +116,14 @@ def parse_config(myconfig):
 	# add our python base directory to use for loading target arch's
 	conf_values["PythonDir"] = __selfpath__
 
+
+	if "bindist" in string.split(conf_values["options"]):
+		print "Binary redistribution enabled"
+	else:
+		print "Bindist is not enabled in catalyst.conf"
+		print "Binary redistribution of generated stages/isos may be prohibited by law."
+		print "Please see the use description for bindist on any package you are including."
+
 	# print out any options messages
 	for opt in conf_values['options']:
 		if opt in option_messages:

The bindist commit that was not in my branch yet.


diff --git a/etc/catalyst.conf b/etc/catalyst.conf
index a65413c..d52a5c8 100644
--- a/etc/catalyst.conf
+++ b/etc/catalyst.conf
@@ -60,7 +60,7 @@ hash_function="crc32"
 # pkgcache = keeps a tbz2 of every built package (useful if your build stops
 #	prematurely)
 # preserve_libs = enables portage to preserve used libs when unmerging packages
-#	(used on installcd-stage2 and stage4 targets)
+#   (used on installcd-stage2 and stage4 targets)
 # seedcache = use the build output of a previous target if it exists to speed up
 #	the copy
 # snapcache = cache the snapshot so that it can be bind-mounted into the chroot.


A minor comment indent change I made in 3.0 is lost.


diff --git a/targets/stage1/stage1-chroot.sh b/targets/stage1/stage1-chroot.sh
index a5d6a01..4ddb636 100755
--- a/targets/stage1/stage1-chroot.sh
+++ b/targets/stage1/stage1-chroot.sh
@@ -41,6 +41,9 @@ if [ -n "${clst_update_seed}" ]; then
 		echo "Invalid setting for update_seed: ${clst_update_seed}"
 		exit 1
 	fi
+
+	# reset emerge options for the target
+	clst_update_seed=no setup_myemergeopts
 else
 	echo "Skipping seed stage update..."
 fi


This one I want to check out more why it is different, but I suspect 
it is a difference in what Jorge committed to master which was a little different
than I had in 3.0.  I later rebased those changes a few times once we got the bugs
worked out.



diff --git a/targets/support/chroot-functions.sh b/targets/support/chroot-functions.sh
index 39bb1cd..6271dc4 100755
--- a/targets/support/chroot-functions.sh
+++ b/targets/support/chroot-functions.sh
@@ -133,6 +133,7 @@ setup_myemergeopts(){
 	then
 		export bootstrap_opts="${bootstrap_opts} -f"
 		export clst_myemergeopts="${clst_myemergeopts} -f"
+	# if we have PKGCACHE, and either update_seed is empty or 'no', make and use binpkgs
 	elif [ -n "${clst_PKGCACHE}" ] && [ -z "${clst_update_seed}" -o "${clst_update_seed}" = "no" ]
 	then
 		# if you add --usepkg, then also add --binpkg-respect-use=y

The difference in jmbvicetto's commit to master, as opposed to mine in 3.0, it's just a comment.

[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 620 bytes --]

Re: [gentoo-catalyst] rewrite-on-master branch

[Brian Dolbec]

[-- Attachment #1: Type: text/plain, Size: 3950 bytes --]

On Sat, 2013-07-20 at 02:39 -0700, Brian Dolbec wrote:
> On Fri, 2013-07-19 at 21:52 -0700, Matt Turner wrote:
> > I'm not okay with redoing the rename the branches trick. Let's use git
> > how it's supposed to be used.
> > 
> > I've rebased the origin/3.0 branch onto master. There are 81 patches in it.
> > 
> > All of the differences between origin/3.0 and rewrite-on-master seem
> > to be mistakes in the 3.0 branch.
> > 
> > I'm not advocating pushing this to master yet. The series needs review
> > and some clean up.
> > 
> 
> No, there is no 3.0 mistakes in the diff,  a few minor differences and
> the bindist and unpack_command commits that weren't in my 3.0 rewrite.
> 

> 
> diff --git a/targets/stage1/stage1-chroot.sh b/targets/stage1/stage1-chroot.sh
> index a5d6a01..4ddb636 100755
> --- a/targets/stage1/stage1-chroot.sh
> +++ b/targets/stage1/stage1-chroot.sh
> @@ -41,6 +41,9 @@ if [ -n "${clst_update_seed}" ]; then
>  		echo "Invalid setting for update_seed: ${clst_update_seed}"
>  		exit 1
>  	fi
> +
> +	# reset emerge options for the target
> +	clst_update_seed=no setup_myemergeopts
>  else
>  	echo "Skipping seed stage update..."
>  fi
> 
> 
> This one I want to check out more why it is different, but I suspect 
> it is a difference in what Jorge committed to master which was a little different
> than I had in 3.0.  I later rebased those changes a few times once we got the bugs
> worked out.

OK, this one is a rebase error from your cherrypicking.

Below is the code chunk from your rebase.  I had correctly moved the
code in question into the "if" where it should be.  That code was still
in the original location, so is now a duplicate of the code in the
correct location.  The mistake was/is present in current master.  The
3.0 version is the correct one.

 # Update stage3
  27 if [ -n "${clst_update_seed}" ]; then
  28         if [ "${clst_update_seed}" == "yes" ]; then
  29                 if [ -n "${clst_update_seed_command}" ]; then
  30                         echo "--- Updating seed stage with USER defined update_seed_command"
  31                         clst_root_path=/ run_merge "${clst_update_seed_command}"
  32                 else
  33                         echo "--- Updating seed stage with DEFAULT update_seed_command"
  34                         update_cmd="--update --deep --complete-graph --rebuild-if-new-ver gcc"
  35                         clst_root_path=/ run_merge ${update_cmd}
  36                 fi
  37 
  38                 # reset the emerge options for the target
  39                 clst_update_seed=no setup_myemergeopts
  40         elif [ "${clst_update_seed}" != "no" ]; then
  41                 echo "Invalid setting for update_seed: ${clst_update_seed}"
  42                 exit 1
  43         fi
  44 
  45         # reset emerge options for the target
  46         clst_update_seed=no setup_myemergeopts
  47 else
  48         echo "Skipping seed stage update..."
  49 fi


> 
> diff --git a/targets/support/chroot-functions.sh b/targets/support/chroot-functions.sh
> index 39bb1cd..6271dc4 100755
> --- a/targets/support/chroot-functions.sh
> +++ b/targets/support/chroot-functions.sh
> @@ -133,6 +133,7 @@ setup_myemergeopts(){
>  	then
>  		export bootstrap_opts="${bootstrap_opts} -f"
>  		export clst_myemergeopts="${clst_myemergeopts} -f"
> +	# if we have PKGCACHE, and either update_seed is empty or 'no', make and use binpkgs
>  	elif [ -n "${clst_PKGCACHE}" ] && [ -z "${clst_update_seed}" -o "${clst_update_seed}" = "no" ]
>  	then
>  		# if you add --usepkg, then also add --binpkg-respect-use=y
> 
> The difference in jmbvicetto's commit to master, as opposed to mine in 3.0, it's just a comment.

I was wrong, it was the difference in dwfreed's commit for master that
was different than his fix that I applied to 3.0 before it was applied
to master.


[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 620 bytes --]