From mboxrd@z Thu Jan 1 00:00:00 1970 From: Kip Macy Subject: Re: [Xen-devel] hg 0.7 released Date: Tue, 20 Sep 2005 18:19:57 -0700 Message-ID: References: <200509161628.35324.hollisb@us.ibm.com> <20050921011435.GA21777@waste.org> Reply-To: kip.macy@gmail.com Mime-Version: 1.0 Content-Type: multipart/mixed; boundary="===============0267907029==" Return-path: In-Reply-To: <20050921011435.GA21777@waste.org> List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Mime-version: 1.0 Sender: mercurial-bounces@selenic.com Errors-To: mercurial-bounces@selenic.com To: Matt Mackall Cc: xen-devel@lists.xensource.com, mercurial@selenic.com, Hollis Blanchard List-Id: xen-devel@lists.xenproject.org --===============0267907029== Content-Type: multipart/alternative; boundary="----=_Part_11863_12682300.1127265597911" ------=_Part_11863_12682300.1127265597911 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: quoted-printable Content-Disposition: inline Hmm -=20 What are you doing differently in 0.7 that is causing the server to use mor= e=20 memory? -Kip On 9/20/05, Matt Mackall wrote: >=20 > On Tue, Sep 20, 2005 at 06:07:34PM -0700, Kip Macy wrote: > > nifty - but 0.7 is not backward compatible with 0.6b: > > %hg clone http://xenbits.xensource.com/xen-unstable.hg > > requesting all changes > > adding changesets > > adding manifests > > adding file changes > > added 6917 changesets with 0 changes to 0 files > > ** unknown exception encountered, details follow > > ** report bug details to mercurial@selenic.com >=20 > Not so hasty. The bug is in fact that xenbits is sometimes running out > of memory. Other times it works just fine. Here are the relevant bits > from my diagnosis this morning: >=20 > ---- >=20 > And then I added exception handling to the pull code, so that it now > does this: >=20 > abort: premature EOF reading chunk (got 75478 bytes, expected > 1651992944)! >=20 > That expected number says there's probably something else in the > stream. Firing up Ethereal and tracing a connection shows: >=20 > >=20 > So the server ran out of memory trying to decompress a manifest. >=20 > -- > Mathematics is the supreme nostalgia of our time. > ------=_Part_11863_12682300.1127265597911 Content-Type: text/html; charset=ISO-8859-1 Content-Transfer-Encoding: quoted-printable Content-Disposition: inline Hmm -
What are you doing differently in 0.7 that is causing the server to use mor= e memory?

      -Kip

On 9/20/05, Matt Mackall <mpm@selenic.com= > wrote:
On Tue, Sep 20, 2005 at 06:07:34PM -0700, Kip Macy wrote:
> nifty - b= ut 0.7 is not backward compatible with 0.6b:
> %hg clone http://xenbits.xensource.com/x= en-unstable.hg
> requesting all changes
> adding changesets
> addin= g manifests
> adding file changes
> added 6917 changesets with = 0 changes to 0 files
> ** unknown exception encountered, details foll= ow
> ** report bug details to = mercurial@selenic.com

Not so hasty. The bug is in fact that xenb= its is sometimes running out
of memory. Other times it works just fine. = Here are the relevant bits
from my diagnosis this morning:

----

And then I added exc= eption handling to the pull code, so that it now
does this:

abort= : premature EOF reading chunk (got 75478 bytes, expected
1651992944)!

That expected number says there's probably something else in thestream. Firing up Ethereal and tracing a connection shows:

<!-- = The above is a description of an error in a Python program,
formatted
     for a Web browser because the 'cgitb' module w= as enabled.  In
     case you
 &nb= sp;   are not reading this in a Web browser, here is the original=
     traceback:

Traceback (most recent call = last):
  File "/var/www/html/xen- unstable.hg", line 9, in ?
    h.run()
 = ; File "/usr/local/hg-0.6c/lib/python/mercurial/hgweb.py", l= ine 701,
  in run
    chunk =3D f.read(= 4096)
  File "/usr/local/hg-0.6c/lib/python/mercurial/hg.= py", line 1442, in
  read
    self.buf +=3D self.g.next()=
  File "/usr/local/hg-0.6c/lib/python/mercurial/hg.py&qu= ot;, line 1469, in
  gengroup
    for y= in self.manifest.group(linkmap): yield y
  File "/usr/lo= cal/hg- 0.6c/lib/python/mercurial/revlog.py", line 433,
  in grou= p
    chunks[r] =3D decompress(data[pos: pos + l])  File "/usr/local/hg-0.6c/lib/python/mercurial/revlog.py&= quot;, line 33,
  in decompress
    if t =3D=3D 'x': return zlib.decompress(bin)
Mem= oryError

-->

So the server ran out of memory trying to dec= ompress a manifest.

--
Mathematics is the supreme nostalgia of ou= r time.

------=_Part_11863_12682300.1127265597911-- --===============0267907029== Content-Type: text/plain; charset="us-ascii" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit Content-Disposition: inline _______________________________________________ Mercurial mailing list Mercurial@selenic.com http://selenic.com/mailman/listinfo/mercurial --===============0267907029==--