[gentoo-portage-dev] [PATCH 1/2] portdbapi.cp_list(): Don't write into xmatch's "match-all" cache

[gentoo-portage-dev] [PATCH 1/2] portdbapi.cp_list(): Don't write into xmatch's "match-all" cache

[SebastianLuther]

From: Sebastian Luther <SebastianLuther@gmx.de>

xmatch returns _pkg_str instances these days. They require metadata
access, which cp_list doesn't have. That means that writing cp_list
results into the xmatch cache breaks xmatch users that expect _pkg_str
instances with full metadata.
---
 pym/portage/dbapi/porttree.py | 7 -------
 1 file changed, 7 deletions(-)

diff --git a/pym/portage/dbapi/porttree.py b/pym/portage/dbapi/porttree.py
index fc3fc03..62ac75c 100644
--- a/pym/portage/dbapi/porttree.py
+++ b/pym/portage/dbapi/porttree.py
@@ -740,12 +740,6 @@ class portdbapi(dbapi):
 
 		if self.frozen and mytree is None:
 			cachelist = self.xcache["cp-list"].get(mycp)
-			if cachelist is not None:
-				# Try to propagate this to the match-all cache here for
-				# repoman since he uses separate match-all caches for each
-				# profile (due to differences in _get_implicit_iuse).
-				self.xcache["match-all"][(mycp, mycp)] = cachelist
-				return cachelist[:]
 		mysplit = mycp.split("/")
 		invalid_category = mysplit[0] not in self._categories
 		d={}
@@ -796,7 +790,6 @@ class portdbapi(dbapi):
 		if self.frozen and mytree is None:
 			cachelist = mylist[:]
 			self.xcache["cp-list"][mycp] = cachelist
-			self.xcache["match-all"][(mycp, mycp)] = cachelist
 		return mylist
 
 	def freeze(self):
-- 
1.8.3.2

[gentoo-portage-dev] [PATCH 2/2] repoman: Add check for missing slot operators (bug 493742)

[SebastianLuther]

From: Sebastian Luther <SebastianLuther@gmx.de>

---
 bin/repoman | 12 +++++++++++-
 1 file changed, 11 insertions(+), 1 deletion(-)

diff --git a/bin/repoman b/bin/repoman
index d1542e9..2a332a7 100755
--- a/bin/repoman
+++ b/bin/repoman
@@ -78,7 +78,7 @@ from portage.output import ConsoleStyleFile, StyleWriter
 from portage.util import writemsg_level
 from portage.util._argparse import ArgumentParser
 from portage.package.ebuild.digestgen import digestgen
-from portage.eapi import eapi_has_iuse_defaults, eapi_has_required_use
+from portage.eapi import eapi_has_iuse_defaults, eapi_has_required_use, eapi_has_slot_operator
 
 if sys.hexversion >= 0x3000000:
 	basestring = str
@@ -354,6 +354,7 @@ qahelp = {
 	"portage.internal": "The ebuild uses an internal Portage function or variable",
 	"repo.eapi.banned": "The ebuild uses an EAPI which is banned by the repository's metadata/layout.conf settings",
 	"repo.eapi.deprecated": "The ebuild uses an EAPI which is deprecated by the repository's metadata/layout.conf settings",
+	"slot.operator.missing": "The ebuild depends on package with several slots and/or sub slots without specifying a slot operator",
 	"virtual.oldstyle": "The ebuild PROVIDEs an old-style virtual (see GLEP 37)",
 	"virtual.suspect": "Ebuild contains a package that usually should be pulled via virtual/, not directly.",
 	"usage.obsolete": "The ebuild makes use of an obsolete construct",
@@ -399,6 +400,7 @@ qawarnings = set((
 "metadata.warning",
 "portage.internal",
 "repo.eapi.deprecated",
+"slot.operator.missing",
 "usage.obsolete",
 "upstream.workaround",
 "LIVEVCS.stable",
@@ -2070,6 +2072,14 @@ for x in effective_scanlist:
 							 " with a non-zero revision:" + \
 							 " '%s'") % (mytype, atom))
 
+					if not (atom.blocker or atom.slot or atom.slot_operator) and \
+						eapi_has_slot_operator(eapi):
+						child_slots = set((child.slot, child.sub_slot) for child in portdb.xmatch("match-all", atom))
+						if len(child_slots) > 1:
+							stats['slot.operator.missing'] += 1
+							fails['slot.operator.missing'].append(relative_path + \
+								": '%s' missing slot or slot operator" % atom)
+
 			type_list.extend([mytype] * (len(badsyntax) - len(type_list)))
 
 		for m, b in zip(type_list, badsyntax):
-- 
1.8.3.2

Re: [gentoo-portage-dev] [PATCH 2/2] repoman: Add check for missing slot operators (bug 493742)

[Brian Dolbec]

[-- Attachment #1: Type: text/plain, Size: 2495 bytes --]

On Wed, 2014-01-01 at 23:14 +0100, SebastianLuther@gmx.de wrote:
> From: Sebastian Luther <SebastianLuther@gmx.de>
> 
> ---
>  bin/repoman | 12 +++++++++++-
>  1 file changed, 11 insertions(+), 1 deletion(-)
> 
> diff --git a/bin/repoman b/bin/repoman
> index d1542e9..2a332a7 100755
> --- a/bin/repoman
> +++ b/bin/repoman
> @@ -78,7 +78,7 @@ from portage.output import ConsoleStyleFile, StyleWriter
>  from portage.util import writemsg_level
>  from portage.util._argparse import ArgumentParser
>  from portage.package.ebuild.digestgen import digestgen
> -from portage.eapi import eapi_has_iuse_defaults, eapi_has_required_use
> +from portage.eapi import eapi_has_iuse_defaults, eapi_has_required_use, eapi_has_slot_operator
>  
>  if sys.hexversion >= 0x3000000:
>  	basestring = str
> @@ -354,6 +354,7 @@ qahelp = {
>  	"portage.internal": "The ebuild uses an internal Portage function or variable",
>  	"repo.eapi.banned": "The ebuild uses an EAPI which is banned by the repository's metadata/layout.conf settings",
>  	"repo.eapi.deprecated": "The ebuild uses an EAPI which is deprecated by the repository's metadata/layout.conf settings",
> +	"slot.operator.missing": "The ebuild depends on package with several slots and/or sub slots without specifying a slot operator",
>  	"virtual.oldstyle": "The ebuild PROVIDEs an old-style virtual (see GLEP 37)",
>  	"virtual.suspect": "Ebuild contains a package that usually should be pulled via virtual/, not directly.",
>  	"usage.obsolete": "The ebuild makes use of an obsolete construct",
> @@ -399,6 +400,7 @@ qawarnings = set((
>  "metadata.warning",
>  "portage.internal",
>  "repo.eapi.deprecated",
> +"slot.operator.missing",
>  "usage.obsolete",
>  "upstream.workaround",
>  "LIVEVCS.stable",
> @@ -2070,6 +2072,14 @@ for x in effective_scanlist:
>  							 " with a non-zero revision:" + \
>  							 " '%s'") % (mytype, atom))
>  
> +					if not (atom.blocker or atom.slot or atom.slot_operator) and \
> +						eapi_has_slot_operator(eapi):
> +						child_slots = set((child.slot, child.sub_slot) for child in portdb.xmatch("match-all", atom))
> +						if len(child_slots) > 1:
> +							stats['slot.operator.missing'] += 1
> +							fails['slot.operator.missing'].append(relative_path + \
> +								": '%s' missing slot or slot operator" % atom)
> +
>  			type_list.extend([mytype] * (len(badsyntax) - len(type_list)))
>  
>  		for m, b in zip(type_list, badsyntax):

Looks good

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

[gentoo-portage-dev] Re: [PATCH 2/2] repoman: Add check for missing slot operators (bug 493742)

[Ryan Hill]

[-- Attachment #1: Type: text/plain, Size: 366 bytes --]

On Wed,  1 Jan 2014 23:14:11 +0100
SebastianLuther@gmx.de wrote:

+	"slot.operator.missing": "The ebuild depends on package with several...
                                                       ^ a

-- 
Ryan Hill                        psn: dirtyepic_sk
   gcc-porting/toolchain/wxwidgets @ gentoo.org

47C3 6D62 4864 0E49 8E9E  7F92 ED38 BD49 957A 8463

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 490 bytes --]

[gentoo-portage-dev]

[SebastianLuther]

Changes:

* restrict check to runtime dependencies
* don't skip the check for atoms with slots, but only for slot+sub-slot
* fix typo found by Ryan

[gentoo-portage-dev] [PATCH] repoman: Add check for missing slot operators (bug 493742)

[SebastianLuther]

From: Sebastian Luther <SebastianLuther@gmx.de>

---
 bin/repoman | 14 +++++++++++++-
 1 file changed, 13 insertions(+), 1 deletion(-)

diff --git a/bin/repoman b/bin/repoman
index d1542e9..cb1d620 100755
--- a/bin/repoman
+++ b/bin/repoman
@@ -78,7 +78,7 @@ from portage.output import ConsoleStyleFile, StyleWriter
 from portage.util import writemsg_level
 from portage.util._argparse import ArgumentParser
 from portage.package.ebuild.digestgen import digestgen
-from portage.eapi import eapi_has_iuse_defaults, eapi_has_required_use
+from portage.eapi import eapi_has_iuse_defaults, eapi_has_required_use, eapi_has_slot_operator
 
 if sys.hexversion >= 0x3000000:
 	basestring = str
@@ -354,6 +354,7 @@ qahelp = {
 	"portage.internal": "The ebuild uses an internal Portage function or variable",
 	"repo.eapi.banned": "The ebuild uses an EAPI which is banned by the repository's metadata/layout.conf settings",
 	"repo.eapi.deprecated": "The ebuild uses an EAPI which is deprecated by the repository's metadata/layout.conf settings",
+	"slot.operator.missing": "The ebuild depends on a package with several slots and/or sub slots without specifying a slot operator",
 	"virtual.oldstyle": "The ebuild PROVIDEs an old-style virtual (see GLEP 37)",
 	"virtual.suspect": "Ebuild contains a package that usually should be pulled via virtual/, not directly.",
 	"usage.obsolete": "The ebuild makes use of an obsolete construct",
@@ -399,6 +400,7 @@ qawarnings = set((
 "metadata.warning",
 "portage.internal",
 "repo.eapi.deprecated",
+"slot.operator.missing",
 "usage.obsolete",
 "upstream.workaround",
 "LIVEVCS.stable",
@@ -2070,6 +2072,16 @@ for x in effective_scanlist:
 							 " with a non-zero revision:" + \
 							 " '%s'") % (mytype, atom))
 
+					if runtime and not atom.blocker and not atom.slot_operator and \
+						eapi_has_slot_operator(eapi) and \
+						not (atom.slot and atom.sub_slot):
+						child_slots = set((child.slot, child.sub_slot) \
+							for child in portdb.xmatch("match-all", atom))
+						if len(child_slots) > 1:
+							stats['slot.operator.missing'] += 1
+							fails['slot.operator.missing'].append(relative_path + \
+								": '%s' in %s missing slot/sub-slot or slot operator" % (atom, mytype))
+
 			type_list.extend([mytype] * (len(badsyntax) - len(type_list)))
 
 		for m, b in zip(type_list, badsyntax):
-- 
1.8.3.2

[gentoo-portage-dev] [PATCH 1/2] portdbapi.cp_list(): Don't write into xmatch's "match-all" cache

[SebastianLuther]

From: Sebastian Luther <SebastianLuther@gmx.de>

xmatch returns _pkg_str instances these days. They require metadata
access, which cp_list doesn't have. That means that writing cp_list
results into the xmatch cache breaks xmatch users that expect _pkg_str
instances with full metadata.
---
 pym/portage/dbapi/porttree.py | 5 -----
 1 file changed, 5 deletions(-)

diff --git a/pym/portage/dbapi/porttree.py b/pym/portage/dbapi/porttree.py
index fc3fc03..53b9648 100644
--- a/pym/portage/dbapi/porttree.py
+++ b/pym/portage/dbapi/porttree.py
@@ -741,10 +741,6 @@ class portdbapi(dbapi):
 		if self.frozen and mytree is None:
 			cachelist = self.xcache["cp-list"].get(mycp)
 			if cachelist is not None:
-				# Try to propagate this to the match-all cache here for
-				# repoman since he uses separate match-all caches for each
-				# profile (due to differences in _get_implicit_iuse).
-				self.xcache["match-all"][(mycp, mycp)] = cachelist
 				return cachelist[:]
 		mysplit = mycp.split("/")
 		invalid_category = mysplit[0] not in self._categories
@@ -796,7 +792,6 @@ class portdbapi(dbapi):
 		if self.frozen and mytree is None:
 			cachelist = mylist[:]
 			self.xcache["cp-list"][mycp] = cachelist
-			self.xcache["match-all"][(mycp, mycp)] = cachelist
 		return mylist
 
 	def freeze(self):
-- 
1.8.3.2

Re: [gentoo-portage-dev] [PATCH 1/2] portdbapi.cp_list(): Don't write into xmatch's "match-all" cache

[Brian Dolbec]

[-- Attachment #1: Type: text/plain, Size: 1526 bytes --]

On Wed, 2014-01-01 at 23:46 +0100, SebastianLuther@gmx.de wrote:
> From: Sebastian Luther <SebastianLuther@gmx.de>
> 
> xmatch returns _pkg_str instances these days. They require metadata
> access, which cp_list doesn't have. That means that writing cp_list
> results into the xmatch cache breaks xmatch users that expect _pkg_str
> instances with full metadata.
> ---
>  pym/portage/dbapi/porttree.py | 5 -----
>  1 file changed, 5 deletions(-)
> 
> diff --git a/pym/portage/dbapi/porttree.py b/pym/portage/dbapi/porttree.py
> index fc3fc03..53b9648 100644
> --- a/pym/portage/dbapi/porttree.py
> +++ b/pym/portage/dbapi/porttree.py
> @@ -741,10 +741,6 @@ class portdbapi(dbapi):
>  		if self.frozen and mytree is None:
>  			cachelist = self.xcache["cp-list"].get(mycp)
>  			if cachelist is not None:
> -				# Try to propagate this to the match-all cache here for
> -				# repoman since he uses separate match-all caches for each
> -				# profile (due to differences in _get_implicit_iuse).
> -				self.xcache["match-all"][(mycp, mycp)] = cachelist
>  				return cachelist[:]
>  		mysplit = mycp.split("/")
>  		invalid_category = mysplit[0] not in self._categories
> @@ -796,7 +792,6 @@ class portdbapi(dbapi):
>  		if self.frozen and mytree is None:
>  			cachelist = mylist[:]
>  			self.xcache["cp-list"][mycp] = cachelist
> -			self.xcache["match-all"][(mycp, mycp)] = cachelist
>  		return mylist
>  
>  	def freeze(self):

Sounds correct to me.  Anyone object to committing this?

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