Lele Gaifax | 1 Nov 2002 01:05
Picon

Re: Subversion 0.14.4 released

>>>>> On Thu, 31 Oct 2002 13:49:52 -0800, Greg Stein <gstein <at> lyra.org> said:

    GS> On Wed, Oct 30, 2002 at 02:06:39PM -0600, Karl Fogel wrote:
    >> Next time, let's please call it "0.14.4b" or "0.14.4.2" (just
    >> by tweaking the minor number in svn_version.h appropriately).

    GS> svn_version.h defines three *integer* values so that
    GS> versioning can be programmatically detected and managed. We
    GS> don't have a system for ".4b" or ".4.2" types of
    GS> numbering.

What about using a "virtual" fourth value that defaults to the
revision number? This, btw, augment the information kept in the
version number, and should solve the ambiguity noted by Karl. I gives
both the user and developer point of view of the sources in one shot. 

just my .01 euro,
thanx&bye,
lele.
--

-- 
nickname: Lele Gaifax	| Quando vivro' di quello che ho pensato ieri
real: Emanuele Gaifas	| comincero' ad aver paura di chi mi copia.
email: lele <at> seldati.it	|		-- Fortunato Depero, 1929.

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe <at> subversion.tigris.org
For additional commands, e-mail: dev-help <at> subversion.tigris.org

cmpilato | 1 Nov 2002 01:06
Favicon
Gravatar

Re: Memory leak in svn update?

Philip Martin <philip <at> codematters.co.uk> writes:

> John Barstow <John_Barstow <at> gfsg.co.nz> writes:
> 
> > On reviewing my logs, the 'special meaning' is running out of memory.
> > Specifically, adding 981 files that total about 10 MB on-disk resulted in
> > over 1.16GB of memory use when doing an update - versus about 15MB memory
> > use getting the same files as a checkout.
> > 
> > That's a pretty serious leak.
> > And, I might add, occurs in build 3597 on Win32.
> 
> Yup, I get the same leak on Unix.  An update of a working copy
> containing 400 small files in a single directory, where the update
> changes all 400 files, used hundreds of MB before I killed it.  The
> problem occurs when using ra_dav but not when using ra_local.  Total
> size of the working copy 6.5MB, that's including the text bases.

I think we actually lost some pool usage Goodness in ra_dav when I
finished the editor conversion.  The new editors don't manage their
own pools, and the checkouts and updates in ra_dav were sufficiently
complex that when I converted the editor drivers there, I punted
temporarily on implementing pool management (got a note in my personal
TODO file on this).  

In other words, we lost what management was in the old editor
interface, and didn't add any management in the new editor drivers.
As far as I know, this only happened in ra_dav, but would affect any
operations that use checkout or update-ish codepaths (diff, status,
switch, ...)
(Continue reading)

Philip Martin | 1 Nov 2002 01:10
Picon

Re: Memory leak in svn update?

cmpilato <at> collab.net writes:

> It would rock if someone (Philip?) would file an issue on this,
> perhaps even assigning it to me.

Doing now.

--

-- 
Philip Martin

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe <at> subversion.tigris.org
For additional commands, e-mail: dev-help <at> subversion.tigris.org

cmpilato | 1 Nov 2002 01:12
Favicon
Gravatar

Re: PATCH swig-1.3.16 and typemap

Jani Averbach <jaa <at> cc.jyu.fi> writes:

> 1) Swig complained that there is not anymore typemap(ignore)
>     (svn_types.i:33: Warning(119): %typemap(ignore) has been replaced by
>       %typemap(in,numinputs=0).) and compilation failed.
>    I fixed those instances and compilation continued.

Applied this part of the patch.  Thanks.

> However:
> 2) Swig tries link against apr-0 what I don't have that (I have
>    libapr.so  (apr from cvs)), I changed apr-0 to apr, and after that
>    linking also succeed

Sorry, we need to keep it as apr-0, since that is the way APR installs
itself these days.  How old is your APR?

> I tried the produced stuff by svnlook.py, and some of commands
> are working and some not (svnlook.py is out of date?):
> 
>   File "./svnlook.py", line 131, in _print_tree
>     wrap_editor, wrap_baton = delta.svn_delta_compat_wrap(e_ptr, e_baton,
> AttributeError: 'svn.delta' module has no attribute 'svn_delta_compat_wrap'

Yep, I noticed this, too.  And the fix was easy enough (commit this,
too).

> Summa summarum, I don't have the faintest idea about swig, so please take
> deep look at this. Also there is need for change in ac-helper/swig.m4, but
> that is trivial. I don't know if these typemap changes are backward
(Continue reading)

cmpilato | 1 Nov 2002 01:14
Favicon
Gravatar

Re: [PATCH] links to issue database broken (was: svn commit: rev 3556 - trunk/www)

Benjamin Pflugmann <benjamin-svn-dev <at> pflugmann.de> writes:

> The URL points to the 0.14.4 milestone instead of the 0.14.5 one
> (verified on the web page). 
> 
> Patch against r3576 follows (although it will be outdated as soon as
> the 0.14.5 release will be made).
> 
> Regards,
> 
> 	Benjamin.
> 
> Patch from Benjamin Pflugmann <benjamin-svn-log <at> pflugmann.de>
> 
> * www/project_status.html: Correct links to open issues for 0.14.4 release.

Applied, thanks.

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe <at> subversion.tigris.org
For additional commands, e-mail: dev-help <at> subversion.tigris.org

Scott Lamb | 1 Nov 2002 01:17

401 Authorization Required

I updated my client's neon to 0.23.5 and apr/apr-util/subversion to the 
latest (3598) and now I get this error message on checkin:

     subversion/libsvn_client/commit.c:628: (apr_err=175002, src_err=0)
     svn: RA layer request failed
     svn: Commit failed (details follow):
     subversion/libsvn_ra_dav/util.c:412: (apr_err=175002, src_err=0)
     svn: MKACTIVITY of 
/svn/repos/!svn/act/0de13410-56ae-0310-9a19-a8544f16a50c returned status 
code 401 (Authorization Required)

The server does require authorization on checkin, but that's not new. I 
haven't changed the Apache config, upgraded the server, or anything.

So why is it erroring out here instead of proceeding to password auth, 
as it has before?

Thanks,
Scott

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe <at> subversion.tigris.org
For additional commands, e-mail: dev-help <at> subversion.tigris.org

Scott Lamb | 1 Nov 2002 01:20

Re: 401 Authorization Required

Scott Lamb wrote:
> I updated my client's neon to 0.23.5 and apr/apr-util/subversion to the 
> latest (3598) and now I get this error message on checkin:

and by 3598, I mean 3597. I guess commits happened while I was fiddling...

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe <at> subversion.tigris.org
For additional commands, e-mail: dev-help <at> subversion.tigris.org

cmpilato | 1 Nov 2002 01:23
Favicon
Gravatar

Re: [PATCH] invert backwards test

Karl Fogel <kfogel <at> newton.ch.collab.net> writes:

> SWIG binding folks: it looks like this never got applied.  I'd apply
> it myself, but I'm in the middle of testing other stuff and don't know
> enough to judge it correct just by looking.  So forwarding it
> here...

Applied.  Thanks, Matt (and Karl for the unearthing).

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe <at> subversion.tigris.org
For additional commands, e-mail: dev-help <at> subversion.tigris.org

cmpilato | 1 Nov 2002 01:32
Favicon
Gravatar

Re: [PATCH] DESTDIR support for python swig targets

"Craig Peterein" <craig <at> peterein.org> writes:

> Pretty self-explainatory.
> 
> Craig
> 
> * Makefile.in
>   (swig-py-ext): Support DESTDIR for compiling.
>   (install-swig-py-ext): Support DESTDIR for installation.

Aplicado.  Gracias.

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe <at> subversion.tigris.org
For additional commands, e-mail: dev-help <at> subversion.tigris.org

cmpilato | 1 Nov 2002 01:35
Favicon
Gravatar

Re: [PATCH] DESTDIR support for python swig targets

cmpilato <at> collab.net writes:

> "Craig Peterein" <craig <at> peterein.org> writes:
> 
> > Pretty self-explainatory.
> > 
> > Craig
> > 
> > * Makefile.in
> >   (swig-py-ext): Support DESTDIR for compiling.
> >   (install-swig-py-ext): Support DESTDIR for installation.
> 
> Aplicado.  Gracias.

Thousands of all those who actually *know* Spanish for whatever that
really translates to.  Something like "I am a graceful avacado" ...

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe <at> subversion.tigris.org
For additional commands, e-mail: dev-help <at> subversion.tigris.org


Gmane