All of lore.kernel.org
 help / color / mirror / Atom feed
From: Robert Yang <liezhi.yang@windriver.com>
To: Richard Purdie <richard.purdie@linuxfoundation.org>
Cc: bitbake-devel@lists.openembedded.org
Subject: Re: [RFC PATCH 1/1] Fix bitbake-runtask
Date: Fri, 25 Nov 2011 17:06:56 +0800	[thread overview]
Message-ID: <4ECF5AB0.5020204@windriver.com> (raw)
In-Reply-To: <1322139798.24143.22.camel@ted>


Hi Richard,

Thank you very much, this is pretty good, I've tested it, it worked
well. What shall I do for it then? Should I send the pull request or
you will merge it from you branch?

// Robert

On 11/24/2011 09:03 PM, Richard Purdie wrote:
> Hi Robert,
>
> On Tue, 2011-11-15 at 21:45 +0800, Robert Yang wrote:
>> Fixes bug [YOCTO #1229]
>>
>> The bitbake doesn't use bitbake-runtask to excute the task any more,
>> but bitbake-runtask is a useful tool for excuting the task without
>> the scheduler of bitbake, so that the external tool can invoke it
>> easily. The main changes are:
>>
>> * It used 4 arguments in the past, but currently uses 2, the reasons:
>>    1) The argv[1] was the tmp/cache/hashdata.dat, but bitbake doesn't
>>       dump the hashdata.dat currently, and since the bitbake-runtask
>>       is only used by the external tool currently, so we don't have to
>>       dump the hashdata.dat, I had tried to add code in
>>       lib/bb/runqueue.py to dump it, but failed.
>>
>>    2) The argv[4] was used to check whether it is "True" or not, if true
>>       then don't excute the task, otherwise, excute it, I dont' know why
>>       we need this, so remove it, but it would be easy to add it back.
>>
>> * Since there is no hashdata.dat, there is no way to get the value of
>>    debug or debug_domains, so we don't use them any more.
>>
>> * Since there is no hashdata.dat, so we can't get the hashdata, so the
>>    function make_stamp doesn't work when run bitbake-runtask, the
>>    make_stamp is use for making stamps and dump sigdata to ${STAMP} (e.g,
>>    tmp/stamps/i586-poky-linux), making stamps are ok, but dump sigdata
>>    failed since there is no hashdata , so I added a variable BB_NO_DUMPSIG
>>    to prevent it dump the sigdata when run by btibake-runtask.
>>
>> * Add the log handler to bitbake-runtask, otherwise it would not print
>>    some useful information, for example, when run:
>>
>>    $ bitbake-runtask gzip do_fetch
>>
>>    There are more than one recipes of gzip, it just prints that
>>    "MultipleMatches", but doesn't print which recipes without the
>>     log handler.
>>
>> * It prints out some strange lines, for exmaple:
>> (clogging
>> LogRecord
>> p2
>> c__builtin__
>> object
>> p3
>> NtRp4
>> (dp5
>> S'taskpid'
>> p6
>> I6653
>> sS'threadName'
>> p7
>> S'MainThread'
>> p8
>>
>> It printed such lines in the past, but now it printed much more
>> lines (more than 2000 lines), though I think this is harmless,
>> I don't know how to reduce them.
>
> Sorry for the delay in getting back to you. This patch raised some
> questions I couldn't answer off the top of my head without experimenting
> with the codebase a bit.
>
> I've created an updated version of this patch which is at:
> http://git.yoctoproject.org/cgit.cgi/poky-contrib/commit/?h=rpurdie/useradd2&id=f11b9e26d806557e7f79538b87e20f5d63a11762
> and also pasted below.
>
> The changes:
>
> I reworked the arguments passed to the code to be cleaner and to make
> some of them optional. This makes it more intuitive to use.
>
> I changed the log handling in this code entirely. It now hooks into the
> log parsing code within bitbake so all log messages are nicely
> formatted. If someone did want the raw messages, we could add a mode
> which passed them to stdout like that but I think this is a good default
> mode.
>
> I also found a neater way to avoid the BB_NO_DUMPSIG option and retain
> some of the signature generation code.
>
> Can you have a look and see if this revised code does everything you
> need? Reading the patch, I need to clean up the mess I made of the
> BBConfiguration class but please see if there are any other problems.
>
> Cheers,
>
> Richard
>
>
>> From f11b9e26d806557e7f79538b87e20f5d63a11762 Mon Sep 17 00:00:00 2001
> From: Robert Yang<liezhi.yang@windriver.co>
> Date: Thu, 24 Nov 2011 12:55:14 +0000
> Subject: Fix bitbake-runtask
>
> Signed-off-by: Richard Purdie<richard.purdie@linuxfoundation.org>
> ---
> diff --git a/bitbake/bin/bitbake-runtask b/bitbake/bin/bitbake-runtask
> index bee0f42..7a3775c 100755
> --- a/bitbake/bin/bitbake-runtask
> +++ b/bitbake/bin/bitbake-runtask
> @@ -4,6 +4,10 @@ import os
>   import sys
>   import warnings
>   sys.path.insert(0, os.path.join(os.path.dirname(os.path.dirname(sys.argv[0])), 'lib'))
> +from bb import fetch2
> +import logging
> +
> +logger = logging.getLogger("BitBake")
>
>   try:
>       import cPickle as pickle
> @@ -16,13 +20,23 @@ class BBConfiguration(object):
>       Manages build options and configurations for one run
>       """
>
> -    def __init__(self, debug, debug_domains):
> -        setattr(self, "data", {})
> -        setattr(self, "file", [])
> -        setattr(self, "cmd", None)
> -        setattr(self, "dump_signatures", True)
> -        setattr(self, "debug", debug)
> -        setattr(self, "debug_domains", debug_domains)
> +    def __init__(self, **options):
> +    #def __init__(self, debug, debug_domains):
> +        self.data = {}
> +        self.file = []
> +        self.cmd = None
> +        self.dump_signatures = True
> +        #self.debug = debug
> +        #self.debug_domains", debug_domains)
> +        self.prefile = []
> +        self.postfile = []
> +        self.parse_only = True
> +
> +    def __getattr__(self, attribute):
> +        try:
> +            return super(BBConfiguration, self).__getattribute__(attribute)
> +        except AttributeError:
> +            return None
>
>   _warnings_showwarning = warnings.showwarning
>   def _showwarning(message, category, filename, lineno, file=None, line=None):
> @@ -39,82 +53,70 @@ warnings.showwarning = _showwarning
>   warnings.simplefilter("ignore", DeprecationWarning)
>
>   import bb.event
> -
> -# Need to map our I/O correctly. stdout is a pipe to the server expecting
> -# events. We save this and then map stdout to stderr.
> -
> -eventfd = os.dup(sys.stdout.fileno())
> -bb.event.worker_pipe = os.fdopen(eventfd, 'w', 0)
> -
> -# map stdout to stderr
> -os.dup2(sys.stderr.fileno(), sys.stdout.fileno())
> -
> -# Replace those fds with our own
> -#logout = data.expand("${TMPDIR}/log/stdout.%s" % os.getpid(), self.cfgData, True)
> -#mkdirhier(os.path.dirname(logout))
> -#newso = open("/tmp/stdout.%s" % os.getpid(), 'w')
> -#os.dup2(newso.fileno(), sys.stdout.fileno())
> -#os.dup2(newso.fileno(), sys.stderr.fileno())
> -
> -# Don't read from stdin from the parent
> -si = file("/dev/null", 'r')
> -os.dup2(si.fileno( ), sys.stdin.fileno( ))
> -
> -# We don't want to see signals to our parent, e.g. Ctrl+C
> -os.setpgrp()
> -
> -# Save out the PID so that the event can include it the
> -# events
> -bb.event.worker_pid = os.getpid()
> -bb.event.useStdout = False
> -
> -hashfile = sys.argv[1]
> -buildfile = sys.argv[2]
> -taskname = sys.argv[3]
> -
>   import bb.cooker
>
> -p = pickle.Unpickler(file(hashfile, "rb"))
> -hashdata = p.load()
> -
> -debug = hashdata["msg-debug"]
> -debug_domains = hashdata["msg-debug-domains"]
> -verbose = hashdata["verbose"]
> +buildfile = sys.argv[1]
> +taskname = sys.argv[2]
> +if len(sys.argv)>= 4:
> +    dryrun = sys.argv[3]
> +else:
> +    dryrun = False
> +if len(sys.argv)>= 5:
> +    hashfile = sys.argv[4]
> +    p = pickle.Unpickler(file(hashfile, "rb"))
> +    hashdata = p.load()
> +else:
> +    hashdata = None
> +
> +handler = bb.event.LogHandler()
> +logger.addHandler(handler)
> +
> +#An example to make debug log messages show up
> +#bb.msg.init_msgconfig(True, 3, [])
> +
> +console = logging.StreamHandler(sys.stdout)
> +format = bb.msg.BBLogFormatter("%(levelname)s: %(message)s")
> +bb.msg.addDefaultlogFilter(console)
> +console.setFormatter(format)
> +
> +def worker_fire(event, d):
> +    if isinstance(event, logging.LogRecord):
> +        console.handle(event)
> +bb.event.worker_fire = worker_fire
> +bb.event.worker_pid = os.getpid()
>
> -bb.utils.init_logger(bb.msg, verbose, debug, debug_domains)
> +initialenv = os.environ.copy()
> +config = BBConfiguration()
>
> -cooker = bb.cooker.BBCooker(BBConfiguration(debug, debug_domains), None)
> -cooker.parseConfiguration()
> +def register_idle_function(self, function, data):
> +    pass
>
> -cooker.bb_cache = bb.cache.init(cooker)
> -cooker.status = bb.cache.CacheData()
> +cooker = bb.cooker.BBCooker(config, register_idle_function, initialenv)
> +config_data = cooker.configuration.data
> +cooker.status = config_data
> +cooker.handleCollections(bb.data.getVar("BBFILE_COLLECTIONS", config_data, 1))
>
> -(fn, cls) = cooker.bb_cache.virtualfn2realfn(buildfile)
> +fn, cls = bb.cache.Cache.virtualfn2realfn(buildfile)
>   buildfile = cooker.matchFile(fn)
> -fn = cooker.bb_cache.realfn2virtual(buildfile, cls)
> +fn = bb.cache.Cache.realfn2virtual(buildfile, cls)
>
>   cooker.buildSetVars()
>
>   # Load data into the cache for fn and parse the loaded cache data
> -the_data = cooker.bb_cache.loadDataFull(fn, cooker.get_file_appends(fn), cooker.configuration.data)
> -cooker.bb_cache.setData(fn, buildfile, the_data)
> -cooker.bb_cache.handle_data(fn, cooker.status)
> -
> -#exportlist = bb.utils.preserved_envvars_export_list()
> -#bb.utils.filter_environment(exportlist)
> +the_data = bb.cache.Cache.loadDataFull(fn, cooker.get_file_appends(fn), cooker.configuration.data)
>
>   if taskname.endswith("_setscene"):
>       the_data.setVarFlag(taskname, "quieterrors", "1")
>
> -bb.parse.siggen.set_taskdata(hashdata["hashes"], hashdata["deps"])
> -
> -for h in hashdata["hashes"]:
> -    bb.data.setVar("BBHASH_%s" % h, hashdata["hashes"][h], the_data)
> -for h in hashdata["deps"]:
> -    bb.data.setVar("BBHASHDEPS_%s" % h, hashdata["deps"][h], the_data)
> +if hashdata:
> +    bb.parse.siggen.set_taskdata(hashdata["hashes"], hashdata["deps"])
> +    for h in hashdata["hashes"]:
> +        bb.data.setVar("BBHASH_%s" % h, hashdata["hashes"][h], the_data)
> +    for h in hashdata["deps"]:
> +        bb.data.setVar("BBHASHDEPS_%s" % h, hashdata["deps"][h], the_data)
>
>   ret = 0
> -if sys.argv[4] != "True":
> +if dryrun != "True":
>       ret = bb.build.exec_task(fn, taskname, the_data)
>   sys.exit(ret)
>
> diff --git a/bitbake/lib/bb/siggen.py b/bitbake/lib/bb/siggen.py
> index 9231291..3aa3fe1 100644
> --- a/bitbake/lib/bb/siggen.py
> +++ b/bitbake/lib/bb/siggen.py
> @@ -160,7 +160,7 @@ class SignatureGeneratorBasic(SignatureGenerator):
>           k = fn + "." + task
>           if runtime == "customfile":
>               sigfile = stampbase
> -        elif runtime:
> +        elif runtime and k in self.taskhash:
>               sigfile = stampbase + "." + task + ".sigdata" + "." + self.taskhash[k]
>           else:
>               sigfile = stampbase + "." + task + ".sigbasedata" + "." + self.basehash[k]
> @@ -181,7 +181,7 @@ class SignatureGeneratorBasic(SignatureGenerator):
>               data['gendeps'][dep] = self.gendeps[fn][dep]
>               data['varvals'][dep] = self.lookupcache[fn][dep]
>
> -        if runtime:
> +        if runtime and k in self.taskhash:
>               data['runtaskdeps'] = self.runtaskdeps[k]
>               data['runtaskhashes'] = {}
>               for dep in data['runtaskdeps']:
>
>
>
>



  reply	other threads:[~2011-11-25  9:13 UTC|newest]

Thread overview: 6+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2011-11-15 13:45 [RFC PATCH 0/1] Fix bitbake-runtask Robert Yang
2011-11-15 13:45 ` [RFC PATCH 1/1] " Robert Yang
2011-11-24  2:06   ` Robert Yang
2011-11-24 13:03   ` Richard Purdie
2011-11-25  9:06     ` Robert Yang [this message]
2011-11-25 11:36       ` Richard Purdie

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=4ECF5AB0.5020204@windriver.com \
    --to=liezhi.yang@windriver.com \
    --cc=bitbake-devel@lists.openembedded.org \
    --cc=richard.purdie@linuxfoundation.org \
    /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 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.