Re: [blfs-dev] seeking advice about exim

2016-12-09 Thread Jeremy Henty

Tim Tassonis wrote:

> And there is no need to install exim if you don't want to run a mail
> server.

I disagree.  On my  home PC (which is not a mail  server) my Mail User
Agent (mutt)  uses exim to send  mail, and fcron uses  exim to deliver
the output  of cron jobs.  These  are not obscure edge  cases: this is
the way  commonly used tools are  designed to work by  default.  There
really are good reasons to install exim on a host, even if it is not a
mail server.

The fact is that many Unix programs  take it for granted that there is
a  sendmail-compatible  MTA  at /usr/sbin/sendmail  (or  some  similar
path).  It's not just for servers!

Regards,

Jeremy Henty
-- 
http://lists.linuxfromscratch.org/listinfo/blfs-dev
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page


Re: [blfs-dev] seeking advice about exim

2016-12-09 Thread Jeremy Henty

Ken Moffat wrote:

> It  sounds appropriate  to recommend  TLS if  it is  in the  default
> config file.

But it  is not!   The default  config file is  src/EDITME ,  and every
mention of TLS in that file is commented out.  BLFS makes some changes
to the default config but none of them affect TLS support.

If I  understand Pierre  Labastie's report  correctly, the  problem it
describes happens  because the  default configuration does  not enable
TLS.

Regards,

Jeremy Henty
-- 
http://lists.linuxfromscratch.org/listinfo/blfs-dev
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page


[blfs-dev] QtWebkit-5.7.0 fails to build with flex 2.6.2

2016-12-09 Thread Cyrillo Baggins
Cyrillo Baggins wrote:
>* Hello,
*>>* I had to downgrade to flex 2.6.1 to get QtWebkit-5.7.0 to build on my
*>* system, I found the tip on the gentoo bug tracker:
*>>* https://bugs.gentoo.org/show_bug.cgi?id=598248

*>>>* I propose adding a comment about the flex requirement in the BLFS book as
*>* LFS dev systemd book already updated to flex 2.6.2
*
>Did you apply flex-2.6.2-fixes-1.patch when you built flex?
>
>   -- Bruce

somehow missed the patch, I was now able to build QtWebkit with flex 2.6.2.
sorry for that.
-- 
http://lists.linuxfromscratch.org/listinfo/blfs-dev
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page


Re: [blfs-dev] seeking advice about exim

2016-12-09 Thread Pierre Labastie

On 09/12/2016 17:31, Jeremy Henty wrote:

Pierre Labastie wrote:


I should  have been clearer. It  is when sending local  mail through
smtp. I  think usual  MUA like heirloom  mailx use  batch submission
(exim -bm), which works with the default config file.

I use mutt, which defaults to using "/usr/sbin/sendmail -oem -oi" (and
sendmail is a symlink to exim),  so I am using batch submission rather
than SMTP.  That explains why I do not see your problem.


You can reproduce the problem with:
exim -bs -v
helo 
mail from: @
rcpt to: @
temporarily rejected RCPT @: syntax error in
"control=dkim_disable_verify"

Confirmed:  I tried  this and  reproduced the  problem exactly  as you
describe it.

I  still do  not understand  why this  reports a  syntax error.   What
happens when you run "/usr/sbin/exim  -bV"?  Does it report any syntax
errors in the configuration file?  I wonder if the configuration is OK
but something else is  feeding the "control=dkim_disable_verify" lines
into a context where they are not syntactically valid.

The man page has (emphasis by myself):
  -bV   This  option causes Exim to write the current version number,
 compilation number, and compilation date of the exim4  
binary
 to  the  standard output.  It also lists the DBM 
library that
 is being used, the optional modules (such as specific  
lookup
 types),  the drivers that are included in the binary, 
and the

 name of the run time configuration file that is in use.

 As part of its operation, -bV causes Exim to read and  
syntax
 check its configuration file. *However, this is a 
static check
 only. It cannot check values that are  to  be 
expanded.  For
 example,  although  a misspelt ACL verb is detected, 
an error
 in the verb's arguments is not.* You cannot rely on 
-bV  alone
 to discover (for example) all the typos in the 
configuration;
 some realistic testing is needed. The -bh and -N 
options pro‐

 vide more dynamic testing facilities.



[...]

I've tried to mail upstream, but I guess I need to subscribe, because my
mail does not appear in their archive.

I hope you  get a response!  It  would be a shame  if nothing happened
after you went to the trouble of bisecting the bug.

Thanks.

I've tried again, now after subscribing. According to the mailing list 
instructions, they will moderate my first message. So, wait and see.

Pierre
--
http://lists.linuxfromscratch.org/listinfo/blfs-dev
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page


Re: [blfs-dev] seeking advice about exim

2016-12-09 Thread Jeremy Henty

Pierre Labastie wrote:

> I should  have been clearer. It  is when sending local  mail through
> smtp. I  think usual  MUA like heirloom  mailx use  batch submission
> (exim -bm), which works with the default config file.

I use mutt, which defaults to using "/usr/sbin/sendmail -oem -oi" (and
sendmail is a symlink to exim),  so I am using batch submission rather
than SMTP.  That explains why I do not see your problem.

> You can reproduce the problem with:
> exim -bs -v
> helo 
> mail from: @
> rcpt to: @
> temporarily rejected RCPT @: syntax error in
> "control=dkim_disable_verify"

Confirmed:  I tried  this and  reproduced the  problem exactly  as you
describe it.

I  still do  not understand  why this  reports a  syntax error.   What
happens when you run "/usr/sbin/exim  -bV"?  Does it report any syntax
errors in the configuration file?  I wonder if the configuration is OK
but something else is  feeding the "control=dkim_disable_verify" lines
into a context where they are not syntactically valid.

> I have bisected through the source repository, and found the commit where
> the commands above started to generate the error...
> 
> commit f444c2c7e41c8a2b43b6270708e7eb88831f11e0
> Author: Jeremy Harris 
> Date:   Wed Jan 6 17:50:06 2016 +
> 
> DKIM: Remove embedded copy of PolarSSL and use OpenSSL/GnuTLS library.
>   Bug 1192
> 
> I've tried to mail upstream, but I guess I need to subscribe, because my
> mail does not appear in their archive.

I hope you  get a response!  It  would be a shame  if nothing happened
after you went to the trouble of bisecting the bug.

Regards,

Jeremy Henty
-- 
http://lists.linuxfromscratch.org/listinfo/blfs-dev
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page


Re: [blfs-dev] QtWebkit-5.7.0 fails to build with flex 2.6.2

2016-12-09 Thread Bruce Dubbs

Cyrillo Baggins wrote:

Hello,

I had to downgrade to flex 2.6.1 to get QtWebkit-5.7.0 to build on my
system, I found the tip on the gentoo bug tracker:

https://bugs.gentoo.org/show_bug.cgi?id=598248


I propose adding a comment about the flex requirement in the BLFS book as
LFS dev systemd book already updated to flex 2.6.2


Did you apply flex-2.6.2-fixes-1.patch when you built flex?

  -- Bruce


--
http://lists.linuxfromscratch.org/listinfo/blfs-dev
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page


Re: [blfs-dev] seeking advice about exim

2016-12-09 Thread akhiezer
> From: Pierre Labastie 
> Date: Fri, 9 Dec 2016 10:29:54 +0100
> Subject: Re: [blfs-dev] seeking advice about exim
>
.
.
> I think usual MUA like heirloom mailx use batch submission (exim -bm), 


(By default, yes; but can use smtp/tls/imap/pop/ no-problem.)


>
.
.
>



akh





--
-- 
http://lists.linuxfromscratch.org/listinfo/blfs-dev
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page


Re: [blfs-dev] seeking advice about exim

2016-12-09 Thread Pierre Labastie

On 09/12/2016 02:44, Jeremy Henty wrote:

Pierre Labastie wrote:


With version 4.87 of exim, the instructions in the book do not produce a
functional MTA. Namely sending local mails is not possible (I have not tried
to send mail to remote accounts anyway). The log shows:
---
2016-12-08 17:18:35 U=root F=<[obfuscated address]> temporarily rejected RCPT
[same address]: syntax error in "control=dkim_disable_verify"
---

The reason is that since version 4.87, DKIM support is no more built by
default, but only if TLS support is asked. But the default config file still
has "control = dkim_disable_verify", which is taken as syntax error since the
program does not know anything about DKIM.

I cannot reproduce this problem.

I have been using exim-4.87 since  May and it delivers local mail with
no problems.   I have searched (case-insensitively)  Local/Makefile in
the exim build  directory for all instances of "tls"  and they are all
commented out,  so I conclude  I have not  asked for TLS  support.  My
/etc/exim.conf   filecontains   3   instances   of"control   =
dkim_disable_verify", yet  this does not cause  any problems.  Running
"/usr/sbin/exim -bV" (which, according to  the docs, syntax checks the
configuration file) also reports no problems.

There is clearly something going on that we do not understand.  Either
I have missed something or the problem is not as you have reported it.
We should resolve this before changing the book.

Regards,

Jeremy Henty
I should have been clearer. It is when sending local mail through smtp. 
I think usual MUA like heirloom mailx use batch submission (exim -bm), 
which works with the default config file. But if you configure, for 
example, thunderbird, to use the local smtp server for delivery, it does 
not work.

You can reproduce the problem with:
exim -bs -v
helo 
mail from: @
rcpt to: @
temporarily rejected RCPT @: syntax error in 
"control=dkim_disable_verify"


I have bisected through the source repository, and found the commit 
where the commands above started to generate the error...


commit f444c2c7e41c8a2b43b6270708e7eb88831f11e0
Author: Jeremy Harris 
Date:   Wed Jan 6 17:50:06 2016 +

DKIM: Remove embedded copy of PolarSSL and use OpenSSL/GnuTLS library.
  Bug 1192

I've tried to mail upstream, but I guess I need to subscribe, because my 
mail does not appear in their archive.


Regards
Pierre

--
http://lists.linuxfromscratch.org/listinfo/blfs-dev
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page


[blfs-dev] QtWebkit-5.7.0 fails to build with flex 2.6.2

2016-12-09 Thread Cyrillo Baggins
Hello,

I had to downgrade to flex 2.6.1 to get QtWebkit-5.7.0 to build on my
system, I found the tip on the gentoo bug tracker:

https://bugs.gentoo.org/show_bug.cgi?id=598248


I propose adding a comment about the flex requirement in the BLFS book as
LFS dev systemd book already updated to flex 2.6.2
-- 
http://lists.linuxfromscratch.org/listinfo/blfs-dev
FAQ: http://www.linuxfromscratch.org/blfs/faq.html
Unsubscribe: See the above information page