headers
Ivan Shmakov | 1 Jun 2008 09:22
Picon
Gravatar

Re: DESTDIR support is broken (patch)

>>>>> Julien ÉLIE <julien <at> trigofacile.com> writes:

 >> One could still do relative links, they need not be in the same
 >> directory (think "..") - something along the lines of:

 >> - $(LN_S) $(MAN8)/docheckgroups.8 $D$(MAN5)/localgroups.5
 >> - $(LN_S) $(MAN8)/perl-nocem.8 $D$(MAN5)/nocem.ctl.5
 >> + $(LN_S) ../man8/docheckgroups.8 $D$(MAN5)/localgroups.5
 >> + $(LN_S) ../man8/perl-nocem.8 $D$(MAN5)/nocem.ctl.5

 > Thanks, Matija.  I have just changed that.  It is proper.

	But please note that the code above silently assumes that both
	$(MAN5) and $(MAN8) are subdirectories of a common directory and
	their ``base'' names are `man5' and `man8', respectively.  If
	for whatever reason the user building INN would set these any
	other way, the `make install' step would happily and without any
	error create a couple of dangling symlinks:

$ LC_ALL=C make MAN5=/foo/bar MAN8=/baz/qux install 2>&1 \
      | tee install.out 

	Thus, it's worthwhile to mention the assumption made at least in
	the code comments.  A better solution would be to check whether
	the assumption is actually satisfied and, if it isn't, to fall
	back creating absolute symbolic links (as was done before.)
Permalink | Reply | 5 comments |
headers
Alexander Bartolich | 1 Jun 2008 19:07
Picon
Picon

Re: innreport generates XHTML 1.1 with CSS

I just sent 600 lines to inn-patches <at> isc.org.
This is a patch against subversion trunk.

innreport now generates XHTML 1.1 using a CSS.
- checked standard conformance with http://validator.w3.org/ and http://jigsaw.w3.org/css-validator/
- checked visual appearance with Firefox 2.0.0.14

Added four optional parameters to innreport.conf
- html_xhtml11_icon (default: http://www.w3.org/Icons/valid-xhtml11
- html_vcss_icon (default: http://jigsaw.w3.org/css-validator/images/vcss)
- html_encoding (default: ISO-8859-1)
- html_css_url

If html_css_url is not specified then a style sheet is embedded into HTML.
This default CSS keeps the old arrangement of elements but gives tables
a cleaner appearance because of "border-collapse:collapse". Header cells,
line numbers and odd/even lines are formatted with different background.

I did not touch version number and copyright notice.
Under what license is innreport, anyway?

--

--
Permalink | Reply | 503 comments |
headers
David Hláčik | 3 Jun 2008 23:33
Picon
Favicon

INN Python support in Centos 5.1

Hello . i want to ask several questions :
1) Is INN on Centos5.1 compiled with python auth hooks support?
2) If i want my messages on news server to keep forever (history) , should i
change expire.ctl?
   What i need to set

3) Is there some configuration which i need to set to be able to use python
auth hooks? Or can i simple proceed with adding them to readers.conf ( I
allready tested my nnrpd_auth.py with nnrpd.py, which is working (nnrpd
module is imported ).

Thanks in advance!
Permalink | Reply | 3 comments |
headers
David Hláčik | 4 Jun 2008 00:15
Picon
Favicon

Re: INN Python support in Centos 5.1

N¬ŠË(}ªâ…«ÞŠzÞi׫³÷(ÿÚºØÿ¥Øÿþ,¶Ïÿÿúr¶'ý«­‡ÿÿ¢›¥Ø?žzéwö®¶éËÿÚqǬ³ú]ƒÿç{
®‹©³ÿᕧ"“ÿÿiDzÏÑ<ÀÛaŠÈ¬Â­‰© zÛbžë¢ihƒû2žÖ±zºè®)ÿz×?ì,þ·šuêìýÊ'ýxÓÿԝìiyË^vÚ$zÿ¢›¥Ø?žzéwö®¶éÈéÓ¹ïɺ}ÿÛM<jÝuÿ}Ï3ðÚ¾'G—ÿâ“÷Z¾'ÿ†VœŠOÞ»ü+¢×ÿüw¥–âÁ©í¶†¬’ǯz¶¥ªç¬¶*'³ÿÿ×ò, Óhœ'§¶‹9ÿW(š˜¥yÜ"¶r¶'jëa†Š$²Ë©¦Šíÿýÿ!ø°j{fÊg¬±¨²‰ç{,z»Þ®Ú$yê_¢·¯z¿áŠËh¯/ÿ².•ßâr§ç±¦*ÞýËeÿõ¡jاyçm¢Ç­ÿýÿ"Ëaz·¬¢gœ¢wâ‚êÚ¶*'Âœ†)ÞyÛh±ëm¡·šnW­¢ë§+a¢ÚºØa¢‰,üêÜjx¬ŠjezšèqçÂ+ai×bžazkh­æz»ÿr‰ßüÚ–ZÞiÜ­zË^vl§žº]ý«­‡úrÂ+ažzéwúrÿbr¬ÂŠäŠxÿžzéwù¨vé^ŠÈ¦¦Šíyßÿÿôájy,Šv½©Ü
Permalink | Reply | 0 comments |
headers
David Hláčik | 4 Jun 2008 21:20
Picon
Favicon

Fwd: python: error <type 'instance'>, ('No module named py',), No module named py

ÿÿÿÿÿÿÿñh¯«uçfzˁïÿÿÿÿÿÿÿº&ü6¯‰Ñåÿø¤ýÖ¯‰ßᕧ"“÷®ü6­{ý´ÓÏúÿ„®n7œ·ô^þœ­†‰ÿzºè¯ûr¥ïâžËZÇÿÿÿÿ6‰¨vé^©žvœÿÿÿÍ¢jºW§jg§$èþœ­†‰ÿ–+-þœ­†‰ÿ¢¸&¢Ëbšš+µ©í™ëa¡Ø§²'^ì,rV¬²+'{ºØp†'!µ©²ë®v¦{úZ²Ì(­Ö§w¥–·­º¹õ‰ù]j–®¶§¶'âq«b¢|°âÿiËnjYrŠyðŠY\jYfzØhu«­…éí‰Æ­{br"–W–[•ÿç{ºØZØ^­ç(™ë-…êk¡¹^›ÿâ½è(¶)ìµæ¡÷–[•ÿç{ºØ^®º+ž‰¨vé^©žvœÿ~º&jëaz{bq«^þÚòþ̬–ˆÿþz-‰ÇÿÿúÞ²émþÏÿ±é_þw°±«­‡ÿÛ¢ËÿÿöÛm¿ÿýì\z›DÅÇ©¶*'þk
þ̬–ˆÿþz-‰Çÿÿ÷«®Šÿ³ÿìÿûÿÿûr¥ïæ²ÿšÈÿj¸,þk ÿó§Yçÿ&éøÿm4ñ«|ÿO3ðÚ¾'G—ÿâ“÷Z¾'ÿ†VœŠOÞ»ü+¢×ÿüw¥–ð…«m†+y«­‰û¥–g¬j°†'!ŠÉž²È§‚g°†‰^u¬¦y©ìÿÿÿ§+a¢Þ®º+þÜ©{ø§²Ö§qïÿÿÿÍ¢jºW§jg§/ÿÿóhš‡n•éڙçÿ§/ÿý«­ë.–ÚrV¬²—`þœ°†'!ŠÇ–W~º&žzéwö®¶éËÿÿN†ÞuëZŠWÿÿç{,z»Þ®)çŠÇ–X§‚ØZ·^Ú"žںק¶'âq«b¢âµÆ¥–É算ÿjëaþœ°…êފ{-jw¡ù²rV¬²+ÿ†Š$yا¶ˆ§Ÿü!zv®¶§¶'âq«b¢vނ)ìŠ×–[&zØhwö®¶§¶'µïÚ®¦z{lýú蚗`ÿÿÈÂ)e¦º/‰×š²f§Ê)ߢ¹š¶*'jÉÞyם¶‹(–÷­†+&ŠË^¯/ÛyÊ.±ïâ­æ¥—)ÞyÛh²‰oz+ÿý8ZžKÿÿÿÿÿû¬¯öâŸ÷§¾œ­†‰ÿþ)©¢»euªÿ~º&žzév)©¢»lÊÉhƒ÷%jË'{ÿÿû®÷«
Permalink | Reply | 0 comments |
headers
David Hláčik | 5 Jun 2008 16:35
Picon
Favicon

investigate python auth problem

Hello ,
How to raise log verbosity, or where can i found python tracebacks, errors
when loading python_auth to readers.conf
All i get from logs is that python authenticate method returned wrong
result. I want to know more. When i run my nnrpd_auth.py and test it from my
test.py it is not returning errors, it is working as it should.
I have created own try, except part to see error, but all i got is : python:
Error: No module named py
I want to know more , i want to know why? There is no other info in logs.

Thanks in advance!

un  5 16:30:53 dev01 nnrpd[12054]: python method dynamic_close not found
Jun  5 16:30:53 dev01 nnrpd[12054]: python: n_a authenticate() invoked:
hostname david-nb.net.hlacik.eu, ipaddress 10.10.10.199, interface
10.10.10.183, user g
Jun  5 16:30:53 dev01 nnrpd[12054]: python: Error: No module named py
*Jun  5 16:30:53 dev01 nnrpd[12054]: python authenticate method returned
wrong result*
Jun  5 16:30:53 dev01 nnrpd[12054]: david-nb.net.hlacik.eu times user 0.016
system 0.020 idle 0.000 elapsed 0.100
Permalink | Reply | 2 comments |
headers
David Hláčik | 5 Jun 2008 16:48
Picon
Favicon

investigate python auth problem

Hello ,
How to raise log verbosity, or where can i found python tracebacks, errors
when loading python_auth to readers.conf
All i get from logs is that python authenticate method returned wrong
result. I want to know more. When i run my nnrpd_auth.py and test it from my
test.py it is not returning errors, it is working as it should.
I have created own try, except part to see error, but all i got is : python:
Error: No module named py
I want to know more , i want to know h

un  5 16:30:53 dev01 nnrpd[12054]: python method dynamic_close not found
Jun  5 16:30:53 dev01 nnrpd[12054]: python: n_a authenticate() invoked:
hostname david-nb.net.hlacik.eu, ipaddress 10.10.10.199, interface
10.10.10.183, user g
Jun  5 16:30:53 dev01 nnrpd[12054]: python: Error: No module named py
*Jun  5 16:30:53 dev01 nnrpd[12054]: python authenticate method returned
wrong result*
Jun  5 16:30:53 dev01 nnrpd[12054]: david-nb.net.hlacik.eu times user 0.016
system 0.020 idle 0.000 elapsed 0.100
Permalink | Reply | 0 comments |
headers
Julien ÉLIE | 5 Jun 2008 20:55
Favicon

Re: Resend: perl filter & "standard" headers

Hi Matija,

First of all, sorry for not having answered sooner...

> Strange - It seems the original message did not reach inn-patches, hmmm.

The original message did not reach inn-patches but the second one (this one
in fact, to which I am answering) reached it.

> My smarthost looked satisfied with delivering both to inn-patches and
> inn-workers (times are CEST), but it seems it only reached inn-workers (with
> attachments stripped) and not inn-patches.

Well, inn-* mailing-lists are quite eerie :-)

> If it happens not to pass again, patches can be found at:
> http://linux.voyager.hr/inn/inn-2.4.4_filter_innd_new_headers.diff
> http://linux.voyager.hr/inn/inn-CURRENT_filter_innd_new_headers.diff

Thanks a lot for having provided the two versions.
I sorted alphabetically the headers in the POD documentation (but not in
innd.c -- I felt lazy...) and I saw that the following five headers
were missing, so I added them:

    Date-Received
    Posted
    Posting-Version
    Received
    Relay-Version
(Continue reading)

Permalink | Reply | 0 comments |
headers
Julien ÉLIE | 5 Jun 2008 21:14
Favicon

Re: DESTDIR support is broken (patch)

Hi Ivan,

> But please note that the code above silently assumes that both
> $(MAN5) and $(MAN8) are subdirectories of a common directory and
> their ``base'' names are `man5' and `man8', respectively.

You're right.

> Thus, it's worthwhile to mention the assumption made at least in
> the code comments.  A better solution would be to check whether
> the assumption is actually satisfied and, if it isn't, to fall
> back creating absolute symbolic links (as was done before.)

Does this patch suit you?

===================================================================
--- doc/man/Makefile    (révision 7854)
+++ doc/man/Makefile    (copie de travail)
 <at>  <at>  -53,8 +53,13  <at>  <at> 
            $(CP_MAN) $$M $D$(MAN5)/$$M ; \
        done
        rm -f $D$(MAN5)/localgroups.5 $D$(MAN5)/nocem.ctl.5
-       $(LN_S) ../man8/docheckgroups.8 $D$(MAN5)/localgroups.5
-       $(LN_S) ../man8/perl-nocem.8 $D$(MAN5)/nocem.ctl.5
+       if [ -r $D$(MAN5)/../man8 ] ; then \
+           $(LN_S) ../man8/docheckgroups.8 $D$(MAN5)/localgroups.5 ; \
+           $(LN_S) ../man8/perl-nocem.8 $D$(MAN5)/nocem.ctl.5 ; \
+       else \
+           $(LN_S) $(MAN8)/docheckgroups.8 $D$(MAN5)/localgroups.5 ; \
+           $(LN_S) $(MAN8)/perl-nocem.8 $D$(MAN5)/nocem.ctl.5 ; \
(Continue reading)

Permalink | Reply | 4 comments |
headers
Julien ÉLIE | 5 Jun 2008 21:24
Favicon

Re: Upgrade to INN 2.5 and history problem

Hi Russ,

>> He successfully managed to upgrade two news servers.  But the third does
>> not work:  he tried twice and when he starts rc.news, he has:
>>
>>     May 15 02:43:19 news innd: can't ftello /var/news/db/history Value too large for defined data type
>
> This is the generic error message returned by the dbz layer when it fails
> to open a dbz database.  It may mean what it says, but it's also quite
> possible that it's some entirely different problem.

It was in fact a Perl problem because if he compiled without Perl support
his INN, there was no problem at all.

> A debugging build will enable a bunch of debugging output that will help
> track down the problem.  Figuring it out short of that tends to be
> annoying.

Sure.  INN 2.5 did not even compile with "make warnings" (which also does
generates a debugging build, thanks to "-DDEBUG=1").  And we found out
the problem with that:

perl.c: In function `HandleHeaders':
perl.c:135: warning: comparison of unsigned expression >= 0 is always true
make[2]: *** [perl.o] Error 1
make[1]: *** [all-nnrpd] Error 2
make: *** [warnings] Error 2

The problem was solved with the last sources of Perl 5.10.  Well, it appears
he was using a broken version of Perl.
(Continue reading)

Permalink | Reply | 0 comments |

Gmane