-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Hi,
I've actually been behind-the-lines (if you will) helping test Pan and
reporting directly to the maintainer of the commits (almost always Heinrich
Müller as it turns out). (I thought a bugzilla report wouldn't be proper since
the recent patches are "work in progress". But again git.gnome.org doesn't
have a per-patch feedback form as does github for example. <heh>)
Right before Christmas in December 2012 I git-pull'd head/master at 4f681cd and
then ran into the mutex glitch:
> $ make -w
> […]
> ar cru libdata.a article.o article-cache.o encode-cache.o cert-store.o data.o
> parts.o xref.o
> ranlib libdata.a
> make[3]: Leaving directory
> `/Volumes/RamDisk/Safari/pan2_gnomeorg_master_try_Dec2012c/pan/data'
> Making all in tasks
> make[3]: Entering directory
> `/Volumes/RamDisk/Safari/pan2_gnomeorg_master_try_Dec2012c/pan/tasks'
> g++-4.2 -DHAVE_CONFIG_H -I. -I../.. -I../.. -D_REENTRANT
> -I/usr/local/include/gmime-2.6 -I/usr/local/include
> -I/usr/local/include/glib-2.0 -I/usr/local/lib/glib-2.0/include
> -D_REENTRANT -I/usr/local/include/glib-2.0 -I/usr/local/lib/glib-2.0/include
> -I/usr/local/include -I/usr/local/include/p11-kit-1 -D_REENTRANT
> -I/usr/local/include/gtk-2.0 -I/usr/local/lib/gtk-2.0/include
> -I/usr/local/include/atk-1.0 -I/usr/local/include/pango-1.0
> -I/usr/local/include/glib-2.0 -I/usr/local/lib/glib-2.0/include
> -I/opt/X11/include/cairo -I/opt/X11/include -I/opt/X11/include/pixman-1
> -I/opt/X11/include/freetype2 -I/opt/X11/include/libpng15
> -I/usr/local/include -I/usr/local/ssl/include -I/usr/X11/include
> -I/usr/local/include -I/WhichXcode/Headers/FlatCarbon -I/usr/include -Os
> -mtune=core2 -march=core2 -force_cpusubtype_ALL -arch i386 -MT decoder.o
> -MD -MP -MF .deps/decoder.Tpo -c -o decoder.o decoder.cc
> In file included from decoder.h:30,
> from decoder.cc:38:
> ../../pan/general/locking.h:39: error: field ‘mutex’ has incomplete type
> make[3]: *** [decoder.o] Error 1
> make[3]: Leaving directory
> `/Volumes/RamDisk/Safari/pan2_gnomeorg_master_try_Dec2012c/pan/tasks'
> make[2]: *** [all-recursive] Error 1
> make[2]: Leaving directory
> `/Volumes/RamDisk/Safari/pan2_gnomeorg_master_try_Dec2012c/pan'
> make[1]: *** [all-recursive] Error 1
> make[1]: Leaving directory
> `/Volumes/RamDisk/Safari/pan2_gnomeorg_master_try_Dec2012c'
> make: *** [all] Error 2
> $ _
Alongside that report I sent my config.log to be complete. Later I sent a
subsequent e-mail describing my own research on this:
> Did a bit of checking on this compiler glitch,
> it seems I still need the 'static' word
> which was rubbed-out by part of commit
> decb39289a79f2408297cd52359a68650d895a99
> labelled "Merge branch 'xzver_new'".
> Searching on 'GMutex' and reading various
> write-ups on it, my thinking went like this:
> either add back the word 'static' on that line
> or perhaps add asterisk [e.g.] 'GMutex* mutex;'
> so I thought it would be best to put what was
> there as originally written.
>
> === 8-K ===
> --- ./pan/general/locking.h_orig 2012-12-22 06:32:44.000000000 -0600
> +++ ./pan/general/locking.h 2012-12-23 23:23:09.000000000 -0600
> @@ -36,7 +36,7 @@
> class Mutex
> {
> private:
> - GMutex mutex;
> + static GMutex mutex;
> GMutex * m;
>
> public:
> === 8-K ===
>
> The compile went normal (as well as expected anyway).
> […]
And so that's where this particular glitch stands as I write/send this note.
;-)
As for the more-recent commits after that point --
I *luv* the extra boost we get when using the NSP's header-compression
facility. (Yet we still are suffering from poor speed while accessing the
locally-saved newsgroup headers. Hopefully the forthcoming[?] sql code should
help here further.)
I think Mr Müller is trying to fix the scenario I'm using: mixing two NSPs in
the single Pan task while using the primary/fallback settings. My main Pan
setup is using Giganews and Astraweb this way (BTW both are unlimited accounts
[for a long-long time; I'm luckily using discounted rates still with both]).
Rather often I am having some lousy network thru-put speeds with one NSP or the
other, so I will swap their Pan settings for primary vis-à-vis fallback. When
I do the swap, subsequent texts will become completely scrambled, even the
headers (inspecting the article-cache subdir). These scrambled articles won't
pass Pan's viewer logic, so that's why it Only Seems we get no "new" headers.
I think the crux of this problem is this: We know GN won't do the XZVER
method, while AW won't do GZIP, but I think it's the Pan primary/fallback logic
that itself is getting mixed-up on this, and thus the logic is using the
'wrong' compression method when I do the swap.
But the new code (starting at GIT 2c870a4 probably designed to partly 'fix'
this) hasn't worked well, still causing weird scrambled text arriving from the
NSPs. So I am 'stuck' at GIT level 88fe336 (see my User- line in the headers
for this post).
Another new problem with Pan: the "Go" -> "Next Unread Group" function is not
working for me. I'm certain the recent patches have introduced this, somehow,
since it was working before.
I _have_ opened bugzilla reports on other problems, since those have been
occurring for quite a while (a year or so) yet I've noticed some recent
discussion on those topics on this list. ;-)
On my personal projects, for those who are interested, I have gotten no-where
further on jumping off this fruity ship.
* For one thing, I needed to replace an ailing LCD TV; I got a LED model now,
and can say nearly every type of lighting is based on LEDs here. ;-D
* For another, I checked-out then decided to get Cox's 50/5 internet service,
plus most of their cable-TV options (eating more $fund$ each month); but I am
still keeping the DSL line also (in case we ever get Full U-verse here) [I'm
trying not to axe this merely for the $aving$].
* Also, I am running more-&-more tasks on this iMac (the only 'puter I have),
such that it's 'free' only during the wee hours of the mornings (BTW look for
certain numerous posts in a.b.s.radio.misc and/or a.b.conspiracy, also active
in other groups at times [blame TPTB for my sudden participation, it's my way
of 'protest']).
* Then I decided to get a Mede8er 500X2 box, so I could offload some work away
from the iMac etc.
* FWIW all of these have been on my Wish List for a long time, too, so I am
slowly getting them resolved.
* I need to clean-up the wiring etc. around here, it's becoming like a
lab-pit. <LOL>
At least this disabled/retired person is trying to stay active with his pet
hobbies. ;-)
p.s. I have also installed several GPG tools, which is why I have 'signed'
this post.
Thanks for reading.
:)
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.19
iQEcBAEBAgAGBQJRCvuyAAoJEKkeWNKet7+KgYcH/2U8me/LZAhJL+qhlBMbSbpH
edrsLb02BjH/PK0yuj5t6iO1Ikc7JqtoBx996z24B/WfvkiCr4ouKQXigHt/lt9I
l6dRgUEKjdniSdVDN8Bw1Jo5071XEblFqlIH43sC45gUBkUux1dMpGg7wX5Vez0M
bWAxOaZqg3ppKhngqnNCUqzPztELaiw0iodbZTMpXQM20tI6f/oQ4lB2etzyWj0M
pDWV9KfD/qn+WW9wz/OWbVvqOZFSLAZQZfIdvDV2HDdW9N7KWSxpkYcN8pzWtjLk
Jxu4+FyFT+dNgNN2qCXt9ssErgp7XfFN4oGY/53ZqnKphSO0bby0xKSGDwvU//w=
=tuXb
-----END PGP SIGNATURE-----
--
_______________________________________________
Pan-users mailing list
Pan-users@nongnu.org
https://lists.nongnu.org/mailman/listinfo/pan-users
