All of lore.kernel.org
 help / color / mirror / Atom feed
* Checksums.ini again - new format this time
@ 2009-10-29 11:41 Marcin Juszkiewicz
  2009-10-29 11:43 ` Phil Blundell
                   ` (2 more replies)
  0 siblings, 3 replies; 15+ messages in thread
From: Marcin Juszkiewicz @ 2009-10-29 11:41 UTC (permalink / raw)
  To: openembedded-devel

[-- Attachment #1: Type: text/plain, Size: 1152 bytes --]


Hi

I know that checksums.ini returns many times on ML so many of you are bored 
with it. But I have some stuff to show/share about it.

Sometime ago new format was suggested for conf/checksums.ini and I implemented 
it today. It looks simple and is simple:

[archivename]
url0=url-to-sources
url1=alternative-url-to-sources
url2=another-alternative-url-to-sources
md5=md5sum
sha256=sha256sum

What differs from old implementation? Use of archive names instead of urls so 
no more "I fetch from my local Debian mirror and got hit by lack of checksums" 
etc problems. By default OE does not even care about urls - it just checks for 
section by archive name and use md5/sha256 sums to check does file is the 
proper one. Urls can be used by source mirror building tools.

How many changes are needed? Few:

1. 3 lines patch to base.bbclass
2. new checksums sorter
3. old checksums -> new checksums converter

I think that it would be nice to switch to that.

Ideas? Opinions?

Regards, 
-- 
JID:      hrw@jabber.org
Website:  http://marcin.juszkiewicz.com.pl/
LinkedIn: http://www.linkedin.com/in/marcinjuszkiewicz

[-- Attachment #2: base.bbclass.diff --]
[-- Type: text/x-patch, Size: 597 bytes --]

diff --git a/classes/base.bbclass b/classes/base.bbclass
index d29ba4b..bf7d49c 100644
--- a/classes/base.bbclass
+++ b/classes/base.bbclass
@@ -72,6 +72,9 @@ def base_chk_file(parser, pn, pv, src_uri, localpath, data):
     elif parser.has_section(src_uri):
         md5    = parser.get(src_uri, "md5")
         sha256 = parser.get(src_uri, "sha256")
+    elif parser.has_section(os.path.basename(src_uri)):
+        md5    = parser.get(os.path.basename(src_uri), "md5")
+        sha256 = parser.get(os.path.basename(src_uri), "sha256")
     else:
         no_checksum = True
 



[-- Attachment #3: oe-checksums-converter.py --]
[-- Type: text/x-python, Size: 3645 bytes --]

#!/usr/bin/env python
# ex:ts=4:sw=4:sts=4:et

# Copyright (C) 2007 OpenedHand
# Copyright (C) 2009 Marcin Juszkiewicz
#
# This program is free software; you can redistribute it and/or modify
# it under the terms of the GNU General Public License version 2 as
# published by the Free Software Foundation.
#
# This program is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
# GNU General Public License for more details.
#
# You should have received a copy of the GNU General Public License along
# with this program; if not, write to the Free Software Foundation, Inc.,
# 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.

#
# OpenEmbedded source checksums converter
#
# based on oe-checksums-sorter.py
#
# This script parse conf/checksums.ini and converts it to new format.
# Duplicate entries are removed.
#

import ConfigParser
import getopt
import os
import sys
import tempfile

def usage(rc):
    print """usage: %s [--inplace|-i] conf/checksums.ini

    --inplace, -i: update file in place (default is to write to stdout)

    If no input file is given, will read from standard input.
    """ % sys.argv[0]
    sys.exit(rc)

try:
    optlist, args = getopt.getopt(sys.argv[1:], "ih", ["inplace", "help"])
except getopt.GetoptError, e:
    print >> sys.stderr, "%s: %s" % (sys.argv[0], e)
    usage(1)

inplace = False
infp = sys.stdin
filename = None
for opt, val in optlist:
    if opt == '-i' or opt == '--inplace':
        inplace = True
    elif opt == 'h' or opt == '--help':
        usage(0)
    else:
        print >> sys.stderr, "%s: %s: invalid argument" % (sys.argv[0], opt)
        usage(1)

if len(args) == 0:
    if inplace:
        print >> sys.stderr, "%s: --inplace requires a filename" % sys.argv[0]
        usage(1)
elif len(args) == 1:
    filename = args[0]
    try:
        infp = open(filename, "r")
    except Exception, e:
        print >> sys.stderr, "%s: %s" % (sys.argv[0], e)
        sys.exit(1)
else:
    print >> sys.stderr, "%s: extra arguments" % sys.argv[0]
    usage(1)

out = sys.stdout
tmpfn = None
if inplace:
    outfd, tmpfn = tempfile.mkstemp(prefix='cksums',
                                    dir=os.path.dirname(filename) or '.')
    os.chmod(tmpfn, os.stat(filename).st_mode)
    out = os.fdopen(outfd, 'w')

checksums_parser = ConfigParser.SafeConfigParser()
checksums_parser.readfp(infp)

new_list = []
seen = {}

for source in checksums_parser.sections():
    archive = source.split("/")[-1]
    md5 = checksums_parser.get(source, "md5")
    sha = checksums_parser.get(source, "sha256")

    tup = (archive, source, md5, sha)
    if not seen.has_key(tup):
        new_list.append(tup)
        seen[tup] = 1

new_list.sort()

archive = ''
archiveid = 0   # for urlX entries
sums = {}       # to keep md5/sha256 in case of archive change

for entry in new_list:

    if archive != entry[0]:

        if archive != '': # not totally first entry
            print >> out, "md5=%s\nsha256=%s\n" % (sums[0], sums[1]) # print old sums
            sums = (entry[2], entry[3])

        print >> out, "[%s]\nurl0=%s" % (entry[0], entry[1])
        archive = entry[0]
        archiveid = 1

    else:
        print >> out, "url%d=%s" % (archiveid, entry[1])
        archiveid += 1

    sums = (entry[2], entry[3])

print >> out, "md5=%s\nsha256=%s\n" % (entry[2], entry[3])

if inplace:
    out.close()
    os.rename(tmpfn, filename)

[-- Attachment #4: oe-checksums-sorter.py --]
[-- Type: text/x-python, Size: 3438 bytes --]

#!/usr/bin/env python
# ex:ts=4:sw=4:sts=4:et

# Copyright (C) 2007 OpenedHand
# Copyright (C) 2009 Marcin Juszkiewicz
#
# This program is free software; you can redistribute it and/or modify
# it under the terms of the GNU General Public License version 2 as
# published by the Free Software Foundation.
#
# This program is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
# GNU General Public License for more details.
#
# You should have received a copy of the GNU General Public License along
# with this program; if not, write to the Free Software Foundation, Inc.,
# 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.

#
# OpenEmbedded source checksums sorter
#
# This script parse conf/checksums.ini and sorts it alphabetically by archive
# name source archive. Duplicate entries are removed.
#

import ConfigParser
import getopt
import os
import sys
import tempfile

def usage(rc):
    print """usage: %s [--inplace|-i] conf/checksums.ini

    --inplace, -i: update file in place (default is to write to stdout)

    If no input file is given, will read from standard input.
    """ % sys.argv[0]
    sys.exit(rc)

try:
    optlist, args = getopt.getopt(sys.argv[1:], "ih", ["inplace", "help"])
except getopt.GetoptError, e:
    print >> sys.stderr, "%s: %s" % (sys.argv[0], e)
    usage(1)

inplace = False
infp = sys.stdin
filename = None
for opt, val in optlist:
    if opt == '-i' or opt == '--inplace':
        inplace = True
    elif opt == 'h' or opt == '--help':
        usage(0)
    else:
        print >> sys.stderr, "%s: %s: invalid argument" % (sys.argv[0], opt)
        usage(1)

if len(args) == 0:
    if inplace:
        print >> sys.stderr, "%s: --inplace requires a filename" % sys.argv[0]
        usage(1)
elif len(args) == 1:
    filename = args[0]
    try:
        infp = open(filename, "r")
    except Exception, e:
        print >> sys.stderr, "%s: %s" % (sys.argv[0], e)
        sys.exit(1)
else:
    print >> sys.stderr, "%s: extra arguments" % sys.argv[0]
    usage(1)

out = sys.stdout
tmpfn = None
if inplace:
    outfd, tmpfn = tempfile.mkstemp(prefix='cksums',
                                    dir=os.path.dirname(filename) or '.')
    os.chmod(tmpfn, os.stat(filename).st_mode)
    out = os.fdopen(outfd, 'w')

checksums_parser = ConfigParser.ConfigParser()
checksums_parser.readfp(infp)

new_list = []
seen = {}

for archive in checksums_parser.sections():
    md5 = checksums_parser.get(archive, "md5")
    sha = checksums_parser.get(archive, "sha256")

    tup = (archive, md5, sha)

    urls = []
    for item in checksums_parser.items(archive):
        if item[0].startswith('url'):
            urls.append(item[1])

    urls.sort()
    for url in urls:
        tup += (url, '')

    if not seen.has_key(archive):
        new_list.append(tup)
        seen[archive] = 1

new_list.sort()

for entry in new_list:
    print >> out, "[%s]" % entry[0] 

    archiveid = 0
    for a in range(3, len(entry)):
        if entry[a]:
            print >> out, "url%d=%s" % (archiveid, entry[a])
            archiveid += 1
    print >> out, "md5=%s\nsha256=%s\n" % (entry[1], entry[2])

if inplace:
    out.close()
    os.rename(tmpfn, filename)

^ permalink raw reply related	[flat|nested] 15+ messages in thread

* Re: Checksums.ini again - new format this time
  2009-10-29 11:41 Checksums.ini again - new format this time Marcin Juszkiewicz
@ 2009-10-29 11:43 ` Phil Blundell
  2009-10-29 13:19   ` Koen Kooi
  2009-11-01 17:42 ` Koen Kooi
  2009-11-01 20:35 ` Leon Woestenberg
  2 siblings, 1 reply; 15+ messages in thread
From: Phil Blundell @ 2009-10-29 11:43 UTC (permalink / raw)
  To: openembedded-devel

On Thu, 2009-10-29 at 12:41 +0100, Marcin Juszkiewicz wrote:
> Ideas? Opinions?

There's already a checksums.ini session scheduled at OEDEM.  Let's talk
about it then.

p.




^ permalink raw reply	[flat|nested] 15+ messages in thread

* Re: Checksums.ini again - new format this time
  2009-10-29 11:43 ` Phil Blundell
@ 2009-10-29 13:19   ` Koen Kooi
  0 siblings, 0 replies; 15+ messages in thread
From: Koen Kooi @ 2009-10-29 13:19 UTC (permalink / raw)
  To: openembedded-devel

On 29-10-09 12:43, Phil Blundell wrote:
> On Thu, 2009-10-29 at 12:41 +0100, Marcin Juszkiewicz wrote:
>> Ideas? Opinions?
>
> There's already a checksums.ini session scheduled at OEDEM.  Let's talk
> about it then.

What's wrong with talking about it in the open right now?




^ permalink raw reply	[flat|nested] 15+ messages in thread

* Re: Checksums.ini again - new format this time
  2009-10-29 11:41 Checksums.ini again - new format this time Marcin Juszkiewicz
  2009-10-29 11:43 ` Phil Blundell
@ 2009-11-01 17:42 ` Koen Kooi
  2009-11-01 20:35 ` Leon Woestenberg
  2 siblings, 0 replies; 15+ messages in thread
From: Koen Kooi @ 2009-11-01 17:42 UTC (permalink / raw)
  To: openembedded-devel

On 29-10-09 12:41, Marcin Juszkiewicz wrote:
>
> Hi
>
> I know that checksums.ini returns many times on ML so many of you are bored
> with it. But I have some stuff to show/share about it.
>
> Sometime ago new format was suggested for conf/checksums.ini and I implemented
> it today. It looks simple and is simple:
>
> [archivename]
> url0=url-to-sources
> url1=alternative-url-to-sources
> url2=another-alternative-url-to-sources
> md5=md5sum
> sha256=sha256sum
>
> What differs from old implementation? Use of archive names instead of urls so
> no more "I fetch from my local Debian mirror and got hit by lack of checksums"
> etc problems. By default OE does not even care about urls - it just checks for
> section by archive name and use md5/sha256 sums to check does file is the
> proper one. Urls can be used by source mirror building tools.
>
> How many changes are needed? Few:
>
> 1. 3 lines patch to base.bbclass
> 2. new checksums sorter
> 3. old checksums ->  new checksums converter
>
> I think that it would be nice to switch to that.
>
> Ideas? Opinions?

Even though the cabal apparantly doesn't want people to respond to your 
idea, I'm going to respond anyway.

I like your idea and think we should implement it. I'm not sure how the 
urlX entries would benefit my usecase, but I suppose it doesn't hurt to 
have them.

regards,

Koen




^ permalink raw reply	[flat|nested] 15+ messages in thread

* Re: Checksums.ini again - new format this time
  2009-10-29 11:41 Checksums.ini again - new format this time Marcin Juszkiewicz
  2009-10-29 11:43 ` Phil Blundell
  2009-11-01 17:42 ` Koen Kooi
@ 2009-11-01 20:35 ` Leon Woestenberg
  2009-11-01 22:59   ` Andrea Adami
                     ` (2 more replies)
  2 siblings, 3 replies; 15+ messages in thread
From: Leon Woestenberg @ 2009-11-01 20:35 UTC (permalink / raw)
  To: openembedded-devel

Hello,

On Thu, Oct 29, 2009 at 12:41 PM, Marcin Juszkiewicz
<marcin@juszkiewicz.com.pl> wrote:
> [archivename]
> url0=url-to-sources
> url1=alternative-url-to-sources
> url2=another-alternative-url-to-sources
> md5=md5sum
> sha256=sha256sum
>

Can we simplify this a bit? i.e.

archivename
url=url-to-sources
url=alternative-url-to-sources
url=another-alternative-url-to-sources
md5=md5sum
sha256=sha256sum

Why not just url instead of url0, url1, url2, ... ?

If you need ordering, just use the order they appear in the file.

+1 for the idea,

Regards,
-- 
Leon



^ permalink raw reply	[flat|nested] 15+ messages in thread

* Re: Checksums.ini again - new format this time
  2009-11-01 20:35 ` Leon Woestenberg
@ 2009-11-01 22:59   ` Andrea Adami
  2009-11-02  2:09     ` Holger Hans Peter Freyther
  2009-11-02  2:06   ` Holger Hans Peter Freyther
  2009-11-02  9:18   ` Marcin Juszkiewicz
  2 siblings, 1 reply; 15+ messages in thread
From: Andrea Adami @ 2009-11-01 22:59 UTC (permalink / raw)
  To: openembedded-devel

And what about having a checksum file in each dir of  /recipes ?
If the checksum changes the PR should be changed too.

While Gentoo's guys started from the opposite situation (one checksum
each recipe), this could be an interesting read:
http://www.gentoo.org/news/20080204-digest-files-removed.xml

Specifically the bits about GLEP44.

My 2 (euro)cents

Andrea



^ permalink raw reply	[flat|nested] 15+ messages in thread

* Re: Checksums.ini again - new format this time
  2009-11-01 20:35 ` Leon Woestenberg
  2009-11-01 22:59   ` Andrea Adami
@ 2009-11-02  2:06   ` Holger Hans Peter Freyther
  2009-11-02  8:41     ` Koen Kooi
  2009-11-02  9:19     ` Marcin Juszkiewicz
  2009-11-02  9:18   ` Marcin Juszkiewicz
  2 siblings, 2 replies; 15+ messages in thread
From: Holger Hans Peter Freyther @ 2009-11-02  2:06 UTC (permalink / raw)
  To: openembedded-devel

On Sunday 01 November 2009 21:35:32 Leon Woestenberg wrote:
> Hello,
> 
> On Thu, Oct 29, 2009 at 12:41 PM, Marcin Juszkiewicz
> 
> <marcin@juszkiewicz.com.pl> wrote:
> > [archivename]
> > url0=url-to-sources
> > url1=alternative-url-to-sources
> > url2=another-alternative-url-to-sources
> > md5=md5sum
> > sha256=sha256sum
> 
> Can we simplify this a bit? i.e.
> 
> archivename
> url=url-to-sources
> url=alternative-url-to-sources
> url=another-alternative-url-to-sources
> md5=md5sum
> sha256=sha256sum
> 
> Why not just url instead of url0, url1, url2, ... ?

the problem with archivename there is no gurantee that the name will be 
globally unique. E.g. with Open Palmtop Integreated Environment (OPIE) One 
Password.. (OPIE) there was a clash and we might or might not have had a clash 
in release versions too...

z.



^ permalink raw reply	[flat|nested] 15+ messages in thread

* Re: Checksums.ini again - new format this time
  2009-11-01 22:59   ` Andrea Adami
@ 2009-11-02  2:09     ` Holger Hans Peter Freyther
  2009-11-09  6:53       ` Petr Štetiar
  0 siblings, 1 reply; 15+ messages in thread
From: Holger Hans Peter Freyther @ 2009-11-02  2:09 UTC (permalink / raw)
  To: openembedded-devel

On Sunday 01 November 2009 23:59:18 Andrea Adami wrote:
> And what about having a checksum file in each dir of  /recipes ?
> If the checksum changes the PR should be changed too.


I think that is the best way in terms of scalability and BB collections. We 
have some weird cases where we might do require ../ but these are broken or 
need to duplicate the checksums...

z.



^ permalink raw reply	[flat|nested] 15+ messages in thread

* Re: Checksums.ini again - new format this time
  2009-11-02  2:06   ` Holger Hans Peter Freyther
@ 2009-11-02  8:41     ` Koen Kooi
  2009-11-02  9:07       ` Holger Hans Peter Freyther
  2009-11-02  9:19     ` Marcin Juszkiewicz
  1 sibling, 1 reply; 15+ messages in thread
From: Koen Kooi @ 2009-11-02  8:41 UTC (permalink / raw)
  To: openembedded-devel

On 02-11-09 03:06, Holger Hans Peter Freyther wrote:
> On Sunday 01 November 2009 21:35:32 Leon Woestenberg wrote:
>> Hello,
>>
>> On Thu, Oct 29, 2009 at 12:41 PM, Marcin Juszkiewicz
>>
>> <marcin@juszkiewicz.com.pl>  wrote:
>>> [archivename]
>>> url0=url-to-sources
>>> url1=alternative-url-to-sources
>>> url2=another-alternative-url-to-sources
>>> md5=md5sum
>>> sha256=sha256sum
>>
>> Can we simplify this a bit? i.e.
>>
>> archivename
>> url=url-to-sources
>> url=alternative-url-to-sources
>> url=another-alternative-url-to-sources
>> md5=md5sum
>> sha256=sha256sum
>>
>> Why not just url instead of url0, url1, url2, ... ?
>
> the problem with archivename there is no gurantee that the name will be
> globally unique. E.g. with Open Palmtop Integreated Environment (OPIE) One
> Password.. (OPIE) there was a clash and we might or might not have had a clash
> in release versions too...

Archives get stored by archive name in DL_DIR already, so that's a moot 
point.

regards,

Koen




^ permalink raw reply	[flat|nested] 15+ messages in thread

* Re: Checksums.ini again - new format this time
  2009-11-02  8:41     ` Koen Kooi
@ 2009-11-02  9:07       ` Holger Hans Peter Freyther
  2009-11-02  9:26         ` Leon Woestenberg
  0 siblings, 1 reply; 15+ messages in thread
From: Holger Hans Peter Freyther @ 2009-11-02  9:07 UTC (permalink / raw)
  To: openembedded-devel

On Monday 02 November 2009 09:41:42 Koen Kooi wrote:

> > the problem with archivename there is no gurantee that the name will be
> > globally unique. E.g. with Open Palmtop Integreated Environment (OPIE)
> > One Password.. (OPIE) there was a clash and we might or might not have
> > had a clash in release versions too...
> 
> Archives get stored by archive name in DL_DIR already, so that's a moot
> point.

No, it means we have another bug somewhere else. :)



^ permalink raw reply	[flat|nested] 15+ messages in thread

* Re: Checksums.ini again - new format this time
  2009-11-01 20:35 ` Leon Woestenberg
  2009-11-01 22:59   ` Andrea Adami
  2009-11-02  2:06   ` Holger Hans Peter Freyther
@ 2009-11-02  9:18   ` Marcin Juszkiewicz
  2 siblings, 0 replies; 15+ messages in thread
From: Marcin Juszkiewicz @ 2009-11-02  9:18 UTC (permalink / raw)
  To: openembedded-devel

Dnia niedziela, 1 listopada 2009 o 21:35:32 Leon Woestenberg napisał(a):
> On Thu, Oct 29, 2009 at 12:41 PM, Marcin Juszkiewicz
> > [archivename]
> > url0=url-to-sources
> > url1=alternative-url-to-sources
> > url2=another-alternative-url-to-sources
> > md5=md5sum
> > sha256=sha256sum
> 
> Can we simplify this a bit? i.e.
> 
> archivename
> url=url-to-sources
> url=alternative-url-to-sources
> url=another-alternative-url-to-sources
> md5=md5sum
> sha256=sha256sum
> 
> Why not just url instead of url0, url1, url2, ... ?
> 
> If you need ordering, just use the order they appear in the file.

ConfigParser class which we use will not behave properly I think (not tested 
that). And urlX lines are not used by OE anyway (but are usable for mirror 
scripts).

Regards, 
-- 
JID:      hrw@jabber.org
Website:  http://marcin.juszkiewicz.com.pl/
LinkedIn: http://www.linkedin.com/in/marcinjuszkiewicz





^ permalink raw reply	[flat|nested] 15+ messages in thread

* Re: Checksums.ini again - new format this time
  2009-11-02  2:06   ` Holger Hans Peter Freyther
  2009-11-02  8:41     ` Koen Kooi
@ 2009-11-02  9:19     ` Marcin Juszkiewicz
  1 sibling, 0 replies; 15+ messages in thread
From: Marcin Juszkiewicz @ 2009-11-02  9:19 UTC (permalink / raw)
  To: openembedded-devel

Dnia poniedziałek, 2 listopada 2009 o 03:06:33 Holger Hans Peter Freyther 
napisał(a):

> the problem with archivename there is no gurantee that the name will be
> globally unique. E.g. with Open Palmtop Integreated Environment (OPIE) One
> Password.. (OPIE) there was a clash and we might or might not have had a
>  clash in release versions too...

So far we have nearly 9k entries and no repeated names. When we will get such 
one then we can keep URLs for them as keys instead of archive names. OE 
supports both situations (my scripts not).

Regards, 
-- 
JID:      hrw@jabber.org
Website:  http://marcin.juszkiewicz.com.pl/
LinkedIn: http://www.linkedin.com/in/marcinjuszkiewicz





^ permalink raw reply	[flat|nested] 15+ messages in thread

* Re: Checksums.ini again - new format this time
  2009-11-02  9:07       ` Holger Hans Peter Freyther
@ 2009-11-02  9:26         ` Leon Woestenberg
  0 siblings, 0 replies; 15+ messages in thread
From: Leon Woestenberg @ 2009-11-02  9:26 UTC (permalink / raw)
  To: openembedded-devel

Hello,

On Mon, Nov 2, 2009 at 10:07 AM, Holger Hans Peter Freyther
<holger+oe@freyther.de> wrote:
> On Monday 02 November 2009 09:41:42 Koen Kooi wrote:
>
> No, it means we have another bug somewhere else. :)
>
We must define a key for finding the correct entry. I think the
package name should be that key value and has always been?

(Or did we use the full original URL as the lookup key for the package
in the checksum file?)

BTW, googling for the MD5/SHA256 is a very good cheat for finding
alternative locations, as I have learned during development of
'witpa'.

Regards,
-- 
Leon



^ permalink raw reply	[flat|nested] 15+ messages in thread

* Re: Checksums.ini again - new format this time
  2009-11-02  2:09     ` Holger Hans Peter Freyther
@ 2009-11-09  6:53       ` Petr Štetiar
  2009-11-10 11:39         ` Denys Dmytriyenko
  0 siblings, 1 reply; 15+ messages in thread
From: Petr Štetiar @ 2009-11-09  6:53 UTC (permalink / raw)
  To: openembedded-devel

Holger Hans Peter Freyther <holger+oe@freyther.de> [2009-11-02 03:09:56]:

> On Sunday 01 November 2009 23:59:18 Andrea Adami wrote:
> > And what about having a checksum file in each dir of  /recipes ?
> > If the checksum changes the PR should be changed too.
> 
> 
> I think that is the best way in terms of scalability and BB collections. We 
> have some weird cases where we might do require ../ but these are broken or 
> need to duplicate the checksums...

I like this idea more than one monolitic checksum file. checksums.ini is place
for storing of a duplicate information anyway. You need to enter URL in checksums,
you need to enter URL again into recipe.

Wouldn't it be better to write URL with checksum once at one place and reuse
it than in recipe as some kind of variable or something like that?

Well, at least that hrw's modification avoids sorting process...

-- ynezz



^ permalink raw reply	[flat|nested] 15+ messages in thread

* Re: Checksums.ini again - new format this time
  2009-11-09  6:53       ` Petr Štetiar
@ 2009-11-10 11:39         ` Denys Dmytriyenko
  0 siblings, 0 replies; 15+ messages in thread
From: Denys Dmytriyenko @ 2009-11-10 11:39 UTC (permalink / raw)
  To: openembedded-devel

On Mon, Nov 09, 2009 at 07:53:22AM +0100, Petr ?tetiar wrote:
> Holger Hans Peter Freyther <holger+oe@freyther.de> [2009-11-02 03:09:56]:
> 
> > On Sunday 01 November 2009 23:59:18 Andrea Adami wrote:
> > > And what about having a checksum file in each dir of  /recipes ?
> > > If the checksum changes the PR should be changed too.
> > 
> > 
> > I think that is the best way in terms of scalability and BB collections. We 
> > have some weird cases where we might do require ../ but these are broken or 
> > need to duplicate the checksums...
> 
> I like this idea more than one monolitic checksum file. checksums.ini is place
> for storing of a duplicate information anyway. You need to enter URL in checksums,
> you need to enter URL again into recipe.
> 
> Wouldn't it be better to write URL with checksum once at one place and reuse
> it than in recipe as some kind of variable or something like that?

It was discussed during OEDEM and the conclusion was to switch to using 
checksums in SRC_URIs of corresponding recipes.

--
Denys



^ permalink raw reply	[flat|nested] 15+ messages in thread

end of thread, other threads:[~2009-11-10 12:41 UTC | newest]

Thread overview: 15+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2009-10-29 11:41 Checksums.ini again - new format this time Marcin Juszkiewicz
2009-10-29 11:43 ` Phil Blundell
2009-10-29 13:19   ` Koen Kooi
2009-11-01 17:42 ` Koen Kooi
2009-11-01 20:35 ` Leon Woestenberg
2009-11-01 22:59   ` Andrea Adami
2009-11-02  2:09     ` Holger Hans Peter Freyther
2009-11-09  6:53       ` Petr Štetiar
2009-11-10 11:39         ` Denys Dmytriyenko
2009-11-02  2:06   ` Holger Hans Peter Freyther
2009-11-02  8:41     ` Koen Kooi
2009-11-02  9:07       ` Holger Hans Peter Freyther
2009-11-02  9:26         ` Leon Woestenberg
2009-11-02  9:19     ` Marcin Juszkiewicz
2009-11-02  9:18   ` Marcin Juszkiewicz

This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.