From: Martin Jansa <martin.jansa@gmail.com>
To: Saul Wold <sgw@linux.intel.com>
Cc: openembedded-core@lists.openembedded.org
Subject: Re: [RFC] buildhistory: Record size of installed package not compressed archive
Date: Thu, 4 Jul 2013 10:30:54 +0200 [thread overview]
Message-ID: <20130704083054.GA3259@jama> (raw)
In-Reply-To: <51D51BD2.20505@linux.intel.com>
[-- Attachment #1: Type: text/plain, Size: 6598 bytes --]
On Wed, Jul 03, 2013 at 11:53:06PM -0700, Saul Wold wrote:
> On 06/29/2013 06:36 PM, Martin Jansa wrote:
> > * usually it's more important to know how much space will each
> > package take on target device then size of compressed package
> >
> > Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
> > ---
> > meta/classes/buildhistory.bbclass | 11 ++++----
> > scripts/oe-pkgdata-util | 59 ++++++++++++++++++++++++++++++++++++---
> > 2 files changed, 61 insertions(+), 9 deletions(-)
> >
>
> I am not sure how, but this patch is causing some failures in poky with
> core-image-sato build, it seems that that the args coming from the
> rootfs_install_complementary get goofed up somehow, not sure but I
> bisected down to this patch.
>
> the -poky seems to have a space added so it gets -poky -linux on the
> command line for oe-pkgdata-utils.
I'm sorry I haven't replied to that patch yet, it was sent only as RFC
to see if people agree that we should just replace current sizes with
size when installed on target (I have different version which records
both, but is a bit slower and resulting file bigger without good
reasons)
I'm aware of the issue and will fix it soon.
> > diff --git a/meta/classes/buildhistory.bbclass b/meta/classes/buildhistory.bbclass
> > index 36e7fe1..1ebe68e 100644
> > --- a/meta/classes/buildhistory.bbclass
> > +++ b/meta/classes/buildhistory.bbclass
> > @@ -317,6 +317,8 @@ buildhistory_get_installed() {
> > # Get list of installed packages
> > pkgcache="$1/installed-packages.tmp"
> > list_installed_packages file | sort > $pkgcache
> > + pkgcachearch="$1/installed-packages-arch.tmp"
> > + list_installed_packages arch | sort > $pkgcachearch
> >
> > cat $pkgcache | awk '{ print $1 }' > $1/installed-package-names.txt
> > if [ -s $pkgcache ] ; then
> > @@ -338,18 +340,17 @@ buildhistory_get_installed() {
> >
> > # Produce installed package sizes list
> > printf "" > $1/installed-package-sizes.tmp
> > - cat $pkgcache | while read pkg pkgfile
> > + cat $pkgcachearch | while read pkg arch
> > do
> > - if [ -f $pkgfile ] ; then
> > - pkgsize=`du -k $pkgfile | head -n1 | awk '{ print $1 }'`
> > - echo $pkgsize $pkg >> $1/installed-package-sizes.tmp
> > - fi
> > + size=`oe-pkgdata-util read_values ${TMPDIR}/pkgdata ${TARGET_VENDOR}-${TARGET_OS} "PKGSIZE" ${pkg}_${arch}`
> > + echo "$size $pkg" >> $1/installed-package-sizes.tmp
> > done
> > cat $1/installed-package-sizes.tmp | sort -n -r | awk '{print $1 "\tKiB " $2}' > $1/installed-package-sizes.txt
> > rm $1/installed-package-sizes.tmp
> >
> > # We're now done with the cache, delete it
> > rm $pkgcache
> > + rm $pkgcachearch
> >
> > if [ "$2" != "sdk" ] ; then
> > # Produce some cut-down graphs (for readability)
> > diff --git a/scripts/oe-pkgdata-util b/scripts/oe-pkgdata-util
> > index 629b2d5..88b8122 100755
> > --- a/scripts/oe-pkgdata-util
> > +++ b/scripts/oe-pkgdata-util
> > @@ -20,9 +20,12 @@
> > # 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
> > #
> > #
> > -# Currently only has one function - mapping of packages to their dev/dbg/doc/locale etc.
> > -# counterparts ("glob" command). Could be extended in future to perform other useful querying
> > -# functions on the pkgdata though.
> > +# Currently only has two functions
> > +# 1) glob - mapping of packages to their dev/dbg/doc/locale etc. counterparts.
> > +# 2) read_values - mapping of packagenames to their location in
> > +# pkgdata and then returns value of selected variable (e.g. PKGSIZE)
> > +# Could be extended in future to perform other useful querying functions on the
> > +# pkgdata though.
> > #
> >
> > import sys
> > @@ -32,7 +35,8 @@ import fnmatch
> > import re
> >
> > def usage():
> > - print("syntax: pkgdata-util glob [-d] <pkgdatadir> <vendor-os> <pkglist> \"<globs>\"");
> > + print("syntax: pkgdata-util glob [-d] <pkgdatadir> <vendor-os> <pkglist> \"<globs>\"\n \
> > + read_values [-d] <pkgdatadir> <vendor-os> <variable-name> \"<ipk filenames>\"");
> >
> >
> >
> > @@ -151,7 +155,52 @@ def glob(args):
> >
> > print("\n".join(mappedpkgs))
> >
> > +def read_values(args):
> > + if len(args) < 4:
> > + usage()
> > + sys.exit(1)
> > +
> > + pkgdata_dir = args[0]
> > + target_suffix = args[1]
> > + var = args[2]
> > + packages = args[3].split()
> >
> > + if target_suffix.startswith("-"):
> > + target_suffix = target_suffix[1:]
> > +
> > + def readvar(pkgdata_file, var):
> > + val = ""
> > + with open(pkgdata_file, 'r') as f:
> > + for line in f:
> > + if line.startswith(var + ":"):
> > + val = line.split(': ')[1].rstrip()
> > + return val
> > +
> > + if debug:
> > + print "read_values('%s', '%s', '%s' '%s'" % (pkgdata_dir, target_suffix, var, packages)
> > + for package in packages:
> > + pkg_split = package.split('_')
> > + pkg_name = pkg_split[0]
> > + pkg_arch = '_'.join(pkg_split[1:])
> > + if debug:
> > + print "package: name: '%s', arch: '%s'" % (pkg_name, pkg_arch)
> > + multimach_target_sys = "%s-%s" % (pkg_arch, target_suffix)
> > + revlink = os.path.join(pkgdata_dir, multimach_target_sys, "runtime-reverse", pkg_name)
> > + if debug:
> > + print(revlink)
> > + if not os.path.exists(revlink):
> > + # [YOCTO #4227] try to drop -gnueabi from TARGET_OS
> > + multimach_target_sys = '-'.join(multimach_target_sys.split('-')[:-1])
> > + revlink = os.path.join(pkgdata_dir, multimach_target_sys, "runtime-reverse", pkg_name)
> > + if debug:
> > + print(revlink)
> > + if os.path.exists(revlink):
> > + mappedpkg = os.path.basename(os.readlink(revlink))
> > + qvar = var
> > + if qvar == "PKGSIZE":
> > + # append packagename
> > + qvar = "%s_%s" % (var, mappedpkg)
> > + print(readvar(revlink, qvar))
> >
> > # Too lazy to use getopt
> > debug = False
> > @@ -173,6 +222,8 @@ if len(args) < 1:
> >
> > if args[0] == "glob":
> > glob(args[1:])
> > +if args[0] == "read_values":
> > + read_values(args[1:])
> > else:
> > usage()
> > sys.exit(1)
> >
--
Martin 'JaMa' Jansa jabber: Martin.Jansa@gmail.com
[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 205 bytes --]
next prev parent reply other threads:[~2013-07-04 8:30 UTC|newest]
Thread overview: 7+ messages / expand[flat|nested] mbox.gz Atom feed top
2013-06-30 1:36 [RFC] buildhistory: Record size of installed package not compressed archive Martin Jansa
2013-07-01 16:53 ` Otavio Salvador
2013-07-04 6:53 ` Saul Wold
2013-07-04 8:30 ` Martin Jansa [this message]
2013-07-06 10:04 ` [PATCHv2] " Martin Jansa
2013-07-08 15:47 ` Paul Eggleton
2013-07-08 16:40 ` Martin Jansa
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20130704083054.GA3259@jama \
--to=martin.jansa@gmail.com \
--cc=openembedded-core@lists.openembedded.org \
--cc=sgw@linux.intel.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox