From l at lrowe.co.uk  Mon May  2 17:39:43 2011
From: l at lrowe.co.uk (Laurence Rowe)
Date: Mon, 2 May 2011 18:39:43 +0100
Subject: [PATCH] vmod_digest
In-Reply-To: <4D42F382.5050402@schokola.de>
References: <4D42F382.5050402@schokola.de>
Message-ID: <BANLkTi=WhUy8SFqFRdcU_z8BE57YRWuHQg@mail.gmail.com>

On 28 January 2011 16:49, Nils Goroll <slink at schokola.de> wrote:
> Hi,
>
> here's an update to the vmod_digest code I initally posted on October 28th.
>
> phk had spotted a race condition in base64_init which I had overlooked (and even
> commented on sarcastically), so I've added code to pre-generate the base64
> lookup tables.
>
> All other changes are due to the changes to varnish code since then.

This looks really interesting. Do you have an up to date version of
this patch? It no longer applies and I was unable to find the base git
revision in the repository (maybe my lack of git-fu.)

Laurence



From MAILER-DAEMON at projects.linpro.no  Sun May  8 04:19:46 2011
From: MAILER-DAEMON at projects.linpro.no (Automatic Email Delivery Software)
Date: Sun,  8 May 2011 06:19:46 +0200 (CEST)
Subject: Returned mail: see transcript for details
Message-ID: <20110508041946.05EB713A4BA@projects.linpro.no>

Your mail 202.40.187.2:2809->87.238.46.8:25 contains contaminated file _From__Automatic_Email_Delivery_Software___MAILER_DAEMON_projects.lin__Date_8_May_2011_10:23:27__Subj_Returned_mail:_see_transcript_for_details_/mail.pif with virus Email-Worm.Win32.Mydoom.m,so it is dropped.




From slink at schokola.de  Wed May 11 12:31:11 2011
From: slink at schokola.de (Nils Goroll)
Date: Wed, 11 May 2011 14:31:11 +0200
Subject: still alive
Message-ID: <4DCA818F.3080200@schokola.de>

For those waiting for my feedback:

I am aware that there are a hand full of solaris bugs waiting for my attention
plus some TODOs regarding vmod_digest. Unfortunately, I can't spend much time
with Varnish ATM, so I need to ask for your patience.

Thanks, Nils



From martin at varnish-software.com  Sat May 14 12:26:41 2011
From: martin at varnish-software.com (Martin Blix Grydeland)
Date: Sat, 14 May 2011 14:26:41 +0200
Subject: VCL compiler creates non-compileable code when calling functions that
	takes zero arguments (patch attached)
Message-ID: <BANLkTi=ryksL6cDQL6vrQ7hGn6pb4V9iqA@mail.gmail.com>

The VCL compiler will produce code that does not compile when calling
functions that takes zero arguments (these would be imported from vmods as
none of the internal functions takes zero arguments). This because it will
always output a comma in preparation for the arguments. Attached is a patch
to correct this and only supply the comma when it is needed.

Regards
Martin Blix Grydeland

-- 
Martin Blix Grydeland
Varnish Software AS
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://www.varnish-cache.org/lists/pipermail/vinyl-dev/attachments/20110514/06f8efea/attachment.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: libvcl-zeroarguments.patch
Type: application/octet-stream
Size: 791 bytes
Desc: not available
URL: <https://www.varnish-cache.org/lists/pipermail/vinyl-dev/attachments/20110514/06f8efea/attachment.obj>

From varnish at bsdchicks.com  Sat May 14 16:42:45 2011
From: varnish at bsdchicks.com (Rogier R. Mulhuijzen)
Date: Sat, 14 May 2011 18:42:45 +0200 (CEST)
Subject: [PATCH] Director stacking
Message-ID: <20110514181539.P55234@ishtar.drwilco.net>


Hi,

I've implemented director stacking (using directors as backends in a 
director) in 3.0. Pretty much everything was already set up for it in VRT, 
only VCC needed some work, and thanks to phk's work there, it was pretty 
simple. And adding this feature made the hosts TAILQ superfluous, bonus!

https://github.com/drwilco/varnish-cache/compare/master...director-stacking

Backporting to 2.1 was a bit trickier, since VCC changed massively, but if 
you want to use it there:

https://github.com/drwilco/varnish-cache/compare/2.1...director-stacking-2.1

(For some reason github is showing tons of commits, but if you hit the 
files-changed tab, it does only show what it should.)

Cheers,

 	Doc

P.S.: Are github links like these fine, or should I also attach plain 
patch files?



From varnish at bsdchicks.com  Sat May 14 16:50:03 2011
From: varnish at bsdchicks.com (Rogier R. Mulhuijzen)
Date: Sat, 14 May 2011 18:50:03 +0200 (CEST)
Subject: [PATCH] Fallback director
Message-ID: <20110514184249.Y55234@ishtar.drwilco.net>


Hi,

The whole director stacking thing came out of wanting to implement a 
"priority director", where instead of weight, you could give priority 
value for each backend. I had the idea of balancing over all backends with 
the same priority, but then ran into how that should be done. Stacking 
made it easy, just one backend per priority level, but then one doesn't 
really need a value, just a list of backends which is checked in order.

And since the Round Robin Director really only needed a minute change for 
that logic, I folded the Fallback Director into its code:

https://github.com/drwilco/varnish-cache/compare/director-stacking...fallback-director

Cheers,

 	Doc



From phk at phk.freebsd.dk  Mon May 16 07:43:27 2011
From: phk at phk.freebsd.dk (Poul-Henning Kamp)
Date: Mon, 16 May 2011 07:43:27 +0000
Subject: VCL compiler creates non-compileable code when calling functions
	that takes zero arguments (patch attached) 
In-Reply-To: Your message of "Sat, 14 May 2011 14:26:41 +0200."
	<BANLkTi=ryksL6cDQL6vrQ7hGn6pb4V9iqA@mail.gmail.com> 
Message-ID: <34752.1305531807@critter.freebsd.dk>

In message <BANLkTi=ryksL6cDQL6vrQ7hGn6pb4V9iqA at mail.gmail.com>, Martin Blix Gr
ydeland writes:

>The VCL compiler will produce code that does not compile when calling
>functions that takes zero arguments (these would be imported from vmods as
>none of the internal functions takes zero arguments). This because it will
>always output a comma in preparation for the arguments. Attached is a patch
>to correct this and only supply the comma when it is needed.

This should be fixed now.  I didn't quite use your patch, because the
fix makes the 'q' variable pointless.

-- 
Poul-Henning Kamp       | UNIX since Zilog Zeus 3.20
phk at FreeBSD.ORG         | TCP/IP since RFC 956
FreeBSD committer       | BSD since 4.3-tahoe    
Never attribute to malice what can adequately be explained by incompetence.



From perbu at varnish-software.com  Fri May 20 12:39:46 2011
From: perbu at varnish-software.com (Per Buer)
Date: Fri, 20 May 2011 14:39:46 +0200
Subject: add vmod-std to the build
Message-ID: <BANLkTi=RsVQV=6-MgtdBtcFBT1YLarz-fg@mail.gmail.com>

Hi.

I've added a man page for the std module to the sphinx reference docs. If
someone could sprinkle some automake magic on top it would be great.

-- 
Per Buer, CEO
Phone: +47 21 98 92 61 / Mobile: +47 958 39 117 / Skype: per.buer
*Varnish makes websites fly!*
Whitepapers <http://www.varnish-software.com/whitepapers> |
Video<http://www.youtube.com/watch?v=x7t2Sp174eI> |
Twitter <https://twitter.com/varnishsoftware>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://www.varnish-cache.org/lists/pipermail/vinyl-dev/attachments/20110520/8cb45263/attachment.html>

From jbq at caraldi.com  Tue May 24 14:53:43 2011
From: jbq at caraldi.com (Jean-Baptiste Quenot)
Date: Tue, 24 May 2011 16:53:43 +0200
Subject: Spurious TODO in varnishncsa.c 3.0 beta1
Message-ID: <BANLkTim2B89tjenph43XzELK-owu=0FFnA@mail.gmail.com>

For the sake of completeness, there is a TODO in varnishncsa.c:

     * TODO:        - Log in any format one wants

But this is actually implemented with the -F option.

Best regards,
-- 
Jean-Baptiste Quenot



From martin at varnish-software.com  Wed May 25 08:26:34 2011
From: martin at varnish-software.com (Martin Blix Grydeland)
Date: Wed, 25 May 2011 10:26:34 +0200
Subject: [PATCH] Make varnishtest default to use path look up varnishd
Message-ID: <BANLkTimDz2u=_Xptm_y=umRRMsYJ2eCCOQ@mail.gmail.com>

Find attached patch for varnishtest to make it useable also after
installation/packaging. Any comments appreciated :)

Regards,
Martin Blix Grydeland

-- 
Martin Blix Grydeland
Varnish Software AS
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://www.varnish-cache.org/lists/pipermail/vinyl-dev/attachments/20110525/b5e846bd/attachment.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: 0001-Make-varnishtest-default-to-use-path-look-up-varnish.patch
Type: application/octet-stream
Size: 3718 bytes
Desc: not available
URL: <https://www.varnish-cache.org/lists/pipermail/vinyl-dev/attachments/20110525/b5e846bd/attachment.obj>

From tfheen at varnish-software.com  Thu May 26 06:49:51 2011
From: tfheen at varnish-software.com (Tollef Fog Heen)
Date: Thu, 26 May 2011 08:49:51 +0200
Subject: Spurious TODO in varnishncsa.c 3.0 beta1
In-Reply-To: <BANLkTim2B89tjenph43XzELK-owu=0FFnA@mail.gmail.com>
	(Jean-Baptiste Quenot's message of "Tue, 24 May 2011 16:53:43 +0200")
References: <BANLkTim2B89tjenph43XzELK-owu=0FFnA@mail.gmail.com>
Message-ID: <8739k2q7lc.fsf@qurzaw.varnish-software.com>

]] Jean-Baptiste Quenot 

| For the sake of completeness, there is a TODO in varnishncsa.c:
| 
|      * TODO:        - Log in any format one wants
| 
| But this is actually implemented with the -F option.

Most of it is implemented, yes.  I've updated the text now, thanks.

-- 
Tollef Fog Heen
Varnish Software
t: +47 21 98 92 64



From tfheen at varnish-software.com  Thu May 26 06:55:54 2011
From: tfheen at varnish-software.com (Tollef Fog Heen)
Date: Thu, 26 May 2011 08:55:54 +0200
Subject: add vmod-std to the build
In-Reply-To: <BANLkTi=RsVQV=6-MgtdBtcFBT1YLarz-fg@mail.gmail.com> (Per Buer's
	message of "Fri, 20 May 2011 14:39:46 +0200")
References: <BANLkTi=RsVQV=6-MgtdBtcFBT1YLarz-fg@mail.gmail.com>
Message-ID: <87y61uosqt.fsf@qurzaw.varnish-software.com>

]] Per Buer 

| I've added a man page for the std module to the sphinx reference docs. If
| someone could sprinkle some automake magic on top it would be great.

Seems like Martin did this yesterday.

-- 
Tollef Fog Heen
Varnish Software
t: +47 21 98 92 64



From andrea.campi at zephirworks.com  Thu May 26 07:28:37 2011
From: andrea.campi at zephirworks.com (Andrea Campi)
Date: Thu, 26 May 2011 09:28:37 +0200
Subject: add vmod-std to the build
In-Reply-To: <87y61uosqt.fsf@qurzaw.varnish-software.com>
References: <BANLkTi=RsVQV=6-MgtdBtcFBT1YLarz-fg@mail.gmail.com>
	<87y61uosqt.fsf@qurzaw.varnish-software.com>
Message-ID: <201DDA70-350F-48A6-9A1D-886C926A9E89@zephirworks.com>

It would be good to expose the same functionality for out-of-tree vmod so they could build and install man pages.

Ditto for tests.



On May 26, 2011, at 8:55 AM, Tollef Fog Heen <tfheen at varnish-software.com> wrote:

> ]] Per Buer 
> 
> | I've added a man page for the std module to the sphinx reference docs. If
> | someone could sprinkle some automake magic on top it would be great.
> 
> Seems like Martin did this yesterday.
> 
> -- 
> Tollef Fog Heen
> Varnish Software
> t: +47 21 98 92 64
> 
> _______________________________________________
> varnish-dev mailing list
> varnish-dev at varnish-cache.org
> http://www.varnish-cache.org/lists/mailman/listinfo/varnish-dev



From andrea.campi at zephirworks.com  Thu May 26 08:20:32 2011
From: andrea.campi at zephirworks.com (Andrea Campi)
Date: Thu, 26 May 2011 10:20:32 +0200
Subject: add vmod-std to the build
In-Reply-To: <201DDA70-350F-48A6-9A1D-886C926A9E89@zephirworks.com>
References: <BANLkTi=RsVQV=6-MgtdBtcFBT1YLarz-fg@mail.gmail.com>
	<87y61uosqt.fsf@qurzaw.varnish-software.com>
	<201DDA70-350F-48A6-9A1D-886C926A9E89@zephirworks.com>
Message-ID: <BANLkTinstRcOQ6jMXCZiKyCsoEYEpD1uAQ@mail.gmail.com>

Also, eventually it would be nice not to require a checked out tree.
The tricky part is doing it in such a way that e.g. vmod.py emits correct
includes when building in-tree (e.g. #include "bin/varnishd/cache.h") and
out (e.g. #include <varnish/cache.h> or include <varnish/private/cache.h>
or something like that).
I guess it all depends on how private you want to keep some of the header
files.

Andrea
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://www.varnish-cache.org/lists/pipermail/vinyl-dev/attachments/20110526/18ebd0b2/attachment.html>

From martin at varnish-software.com  Thu May 26 15:57:15 2011
From: martin at varnish-software.com (Martin Blix Grydeland)
Date: Thu, 26 May 2011 17:57:15 +0200
Subject: varnishadm fails on long responses from varnish (patch attached)
Message-ID: <BANLkTi=C98U0o8jFB0e7kWNeNr184_weDw@mail.gmail.com>

Hi,

I've found that varnishadm will sometimes exit thinking it has lost the CLI
connection to varnish when the result is long (exceeding the TCP MTU I
guess). This is because the socket it is listening on is non-blocking, and
the read_tmo() function will not restart a read operation that failed with
EAGAIN. Patch attached will change this function into restarting the poll
with timeout on EAGAIN.

Please review - any comments appreciated.

Regards,
Martin Blix Grydeland

-- 
Martin Blix Grydeland
Varnish Software AS
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://www.varnish-cache.org/lists/pipermail/vinyl-dev/attachments/20110526/5ff95e57/attachment.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: 0001-Allow-restart-of-timeout-in-read_tmo-when-doing-non-.patch
Type: application/octet-stream
Size: 1010 bytes
Desc: not available
URL: <https://www.varnish-cache.org/lists/pipermail/vinyl-dev/attachments/20110526/5ff95e57/attachment.obj>

From lucas.brasilino at gmail.com  Thu May 26 20:47:38 2011
From: lucas.brasilino at gmail.com (Lucas Brasilino)
Date: Thu, 26 May 2011 17:47:38 -0300
Subject: Should manpages be mandatory for building varnish 3.0.0-beta1 ?
Message-ID: <BANLkTi=fMdUKBk=UVWFB2dhibkcz88OTWA@mail.gmail.com>

Hi,

As I tried to build varnish 3.0.0-beta1 in a Debian Lenny system without
'rst2man', which is not present in Lenny repo. Running 'configure' I've got:

[...]
checking for rst2man... no
checking for rst2man.py... no
configure: WARNING: rst2man not found ? not building man pages
[...]

When I ran 'make', building fails with an error:

You need rst2man installed to make dist

I've made a patch that makes autotools to not generate manpage building
procedures when rst2man isn't present. So I managed to install varnish with
no problem.

If this patch interests you guys, I can issue a ticket and submit the patch. But
my doubt is if manpages are mandatory.

regards
Lucas Brasilino



From fil at rezo.net  Thu May 26 20:58:45 2011
From: fil at rezo.net (Fil)
Date: Thu, 26 May 2011 22:58:45 +0200
Subject: Should manpages be mandatory for building varnish 3.0.0-beta1 ?
In-Reply-To: <BANLkTi=fMdUKBk=UVWFB2dhibkcz88OTWA@mail.gmail.com>
References: <BANLkTi=fMdUKBk=UVWFB2dhibkcz88OTWA@mail.gmail.com>
Message-ID: <BANLkTik+R_5R_geLpyAkvHH1ymverezEXg@mail.gmail.com>

I've been plagued by this too an lost lots of time. Thanks!

On Thu, May 26, 2011 at 10:47 PM, Lucas Brasilino <lucas.brasilino at gmail.com
> wrote:

> You need rst2man installed to make dist
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://www.varnish-cache.org/lists/pipermail/vinyl-dev/attachments/20110526/e0d4da54/attachment.html>

From tfheen at varnish-software.com  Fri May 27 07:24:39 2011
From: tfheen at varnish-software.com (Tollef Fog Heen)
Date: Fri, 27 May 2011 09:24:39 +0200
Subject: add vmod-std to the build
In-Reply-To: <BANLkTinstRcOQ6jMXCZiKyCsoEYEpD1uAQ@mail.gmail.com> (Andrea
	Campi's message of "Thu, 26 May 2011 10:20:32 +0200")
References: <BANLkTi=RsVQV=6-MgtdBtcFBT1YLarz-fg@mail.gmail.com>
	<87y61uosqt.fsf@qurzaw.varnish-software.com>
	<201DDA70-350F-48A6-9A1D-886C926A9E89@zephirworks.com>
	<BANLkTinstRcOQ6jMXCZiKyCsoEYEpD1uAQ@mail.gmail.com>
Message-ID: <87k4dcobbc.fsf@qurzaw.varnish-software.com>

]] Andrea Campi 

| Also, eventually it would be nice not to require a checked out tree.

This is currently because we don't promise anything wrt API and ABI
stability.  Eventually, I'm hoping we can have a stable API and ABI, but
given how new the vmod mechanism is, we don't know if we've gotten it
right or not yet.

| The tricky part is doing it in such a way that e.g. vmod.py emits correct
| includes when building in-tree (e.g. #include "bin/varnishd/cache.h") and
| out (e.g. #include <varnish/cache.h> or include <varnish/private/cache.h>
| or something like that).

That's one of the easier problems to solve, really.

| I guess it all depends on how private you want to keep some of the header
| files.

Yup, basically.

Cheers,
-- 
Tollef Fog Heen
Varnish Software
t: +47 21 98 92 64



From tfheen at varnish-software.com  Fri May 27 07:26:59 2011
From: tfheen at varnish-software.com (Tollef Fog Heen)
Date: Fri, 27 May 2011 09:26:59 +0200
Subject: Should manpages be mandatory for building varnish 3.0.0-beta1 ?
In-Reply-To: <BANLkTi=fMdUKBk=UVWFB2dhibkcz88OTWA@mail.gmail.com> (Lucas
	Brasilino's message of "Thu, 26 May 2011 17:47:38 -0300")
References: <BANLkTi=fMdUKBk=UVWFB2dhibkcz88OTWA@mail.gmail.com>
Message-ID: <87fwo0ob7g.fsf@qurzaw.varnish-software.com>

]] Lucas Brasilino 

| As I tried to build varnish 3.0.0-beta1 in a Debian Lenny system without
| 'rst2man', which is not present in Lenny repo. Running 'configure' I've got:

[...]

| If this patch interests you guys, I can issue a ticket and submit the patch. But
| my doubt is if manpages are mandatory.

I'd rather suggest you use the tarballs we have.

http://repo.varnish-cache.org/source/ has tarballs for 2.1.x and 3.0
beta1, http://repo.varnish-cache.org/snapshots/ has snapshots for git
master.

-- 
Tollef Fog Heen
Varnish Software
t: +47 21 98 92 64



From lucas.brasilino at gmail.com  Fri May 27 21:25:20 2011
From: lucas.brasilino at gmail.com (Lucas Brasilino)
Date: Fri, 27 May 2011 18:25:20 -0300
Subject: Should manpages be mandatory for building varnish 3.0.0-beta1 ?
In-Reply-To: <87fwo0ob7g.fsf@qurzaw.varnish-software.com>
References: <BANLkTi=fMdUKBk=UVWFB2dhibkcz88OTWA@mail.gmail.com>
	<87fwo0ob7g.fsf@qurzaw.varnish-software.com>
Message-ID: <BANLkTimnzYaAK0+2nQ=W1vy1Tvx8e6fkSw@mail.gmail.com>

> | If this patch interests you guys, I can issue a ticket and submit the patch. But
> | my doubt is if manpages are mandatory.
>
> I'd rather suggest you use the tarballs we have.
>
> http://repo.varnish-cache.org/source/ has tarballs for 2.1.x and 3.0
> beta1, http://repo.varnish-cache.org/snapshots/ has snapshots for git
> master.

It worked. Thanks a lot.

regards
Lucas Brasilino



From phk at phk.freebsd.dk  Mon May 30 06:40:05 2011
From: phk at phk.freebsd.dk (Poul-Henning Kamp)
Date: Mon, 30 May 2011 06:40:05 +0000
Subject: add vmod-std to the build 
In-Reply-To: Your message of "Thu, 26 May 2011 10:20:32 +0200."
	<BANLkTinstRcOQ6jMXCZiKyCsoEYEpD1uAQ@mail.gmail.com> 
Message-ID: <7541.1306737605@critter.freebsd.dk>

In message <BANLkTinstRcOQ6jMXCZiKyCsoEYEpD1uAQ at mail.gmail.com>, Andrea Campi w
rites:

>Also, eventually it would be nice not to require a checked out tree.

That is guaranteed not happening for the 3.x series and it is at
best very speculative for later varnish releases.

Doing it, would require us to nail down a stable API across releases
for some of the most central bits in Varnish, and that would just
be an empty illusion.

-- 
Poul-Henning Kamp       | UNIX since Zilog Zeus 3.20
phk at FreeBSD.ORG         | TCP/IP since RFC 956
FreeBSD committer       | BSD since 4.3-tahoe    
Never attribute to malice what can adequately be explained by incompetence.



From martin at varnish-software.com  Tue May 31 08:38:30 2011
From: martin at varnish-software.com (Martin Blix Grydeland)
Date: Tue, 31 May 2011 10:38:30 +0200
Subject: -i option to varnishtest
Message-ID: <BANLkTinqmSdQFPf+JT-2A9tgmaNg9O6x1A@mail.gmail.com>

Some changes has been commited to varnishtest. It will now use the macro
definition of varnishd in order to find the varnishd binary at run-time.
This macro defaults to using path look up. It has also learned the -i
(in-tree) command line option that will set the varnishd macro definition to
find the varnishd binary relative to the varnishtest directory. So specify
-i to varnishtest if you run it manually from your build tree.

Regards,
Martin Blix Grydeland

-- 
Martin Blix Grydeland
Varnish Software AS
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://www.varnish-cache.org/lists/pipermail/vinyl-dev/attachments/20110531/0bf95449/attachment.html>