[aur-general] aur/acroread broken - request for help
hi list, the aur package for adobes pdf reader acroread [0] is broken since 2 weeks now. it ships an outdated version of this software, because adobe stopped building it for linux in may 2013. still some people are bound to this for several reasons, so please dont make this thread a discussion about their decision to use this piece of software. so 2 weeks ago it started do develop problems later identified as related to networking/ipv6. crashing silently after 15s while parsing /etc/hosts. some workarounds were found, none of which was very satisfying. since two days now it crashes immediately on startup, without any notice. the stack trace [1] does not make any sense to me, but gdb spit out some things about lib32-libxcb (running on an x86_64 machine) [2]. downgrading libxcp to version 1.11.1 seems to fix this temporarily, but doesnt provide a long term solution. for that i am stuck, so any help from the list is well appreciated. can someone make sense of that stack trace or the gdb output, and can we draw conclusions (like shipping outdated libraries with the package in an extra folder) to make this work again in long term (aka until the next breakage)? again, please contribute technical advice only, no discussion about free software here, please. best regards georg (maintainer of acroread) [0] https://aur.archlinux.org/packages/acroread [1] http://pastebin.com/FhQM3izE [2] http://pastebin.com/f3Py5LDn
On Fri, 3 Jun 2016 18:13:41 +0200 "G. Schlisio" <g.schlisio@dukun.de> wrote:
hi list,
the aur package for adobes pdf reader acroread [0] is broken since 2 weeks now. it ships an outdated version of this software, because adobe stopped building it for linux in may 2013.
Honestly, I gave up on keeping the Linux version of Adobe Reader running long since — but it's honestly unnecessary. Adobe Reader 11 works flawlessly under Wine, with a little assistance from Winetricks. Run the following commands: $ curl -O https://ardownload2.adobe.com/pub/adobe/reader/win/11.x/11.0.10/en_US/AdbeRd... $ export WINEDEBUG=-all WINEARCH=win32 WINEPREFIX=~/.local/share/wineprefixes/acroread/ $ winetricks prefix=acroread win7 atmlib riched20 wsh57 mspatcha $ wine AdbeRdr11010_en_US.exe # When prompted during installation, select “Manual” for updates. $ winetricks prefix=acroread winxp That sequence will install Adobe Reader 11 and all its necessary support DLLs into a dedicated Wine bottle, and Wine nowadays is smart enough to generate .desktop files that find it, use the correct prefix, and even associate PDFs and PDF forms with it. I hope that resolves your issues. ~Celti
Honestly, I gave up on keeping the Linux version of Adobe Reader running long since — but it's honestly unnecessary. Adobe Reader 11 works flawlessly under Wine, with a little assistance from Winetricks.
[…]
I hope that resolves your issues.
thank you, this is already one of the mentioned workarounds. i will reference your detailed instructions there. still, this does not fix the actual package, which is my ultimate goal. if it proves unfixable, it will be removed from AUR, but until then i will try to find a solution.
downgrading libxcp to version 1.11.1 seems to fix this temporarily Perhaps you could make a PKGBUILD for this exact version of libxcp, call it libxcp111 and make it a dependency for acroread? Seems that
On 06/03/16 10:13, G. Schlisio wrote: maintaining that PKGBUILD would be rather hands off, since it's not likely to have the 1.11 revision updated anymore. But it'd be in the same spirit as say libpng14 ( maybe might have to do a lib32- pkgbuild too? ) -- Thanks, John D Jones III UNIX Zealot; Perl Lover unixgeek1972@gmail.com jnbek1972@gmail.com http://zoelife4u.org/ Where Earth and Spirit Unite
Fri, 3 Jun 2016 18:13:41 +0200 "G. Schlisio" <g.schlisio@dukun.de>:
some things about lib32-libxcb (running on an x86_64 machine) [2]. downgrading libxcp to version 1.11.1 seems to fix this temporarily, but doesnt provide a long term solution. [2] http://pastebin.com/f3Py5LDn
Not exactly very helpful, but I wanted to let you know that I had the same problem with a very old/proprietary/binary 32-bit application I'm using. Running it on the laptop (CoreDuo, no x86_64 support) with Arch i686: no issues. On the x86_64 desktop Arch I had to downgrade lib32-libxcb to 1.11.1, like you. Any further insight appreciated. --byte
On Fri, Jun 03, 2016 at 09:28:48PM +0200, Jens Adam wrote:
Fri, 3 Jun 2016 18:13:41 +0200 "G. Schlisio" <g.schlisio@dukun.de>:
some things about lib32-libxcb (running on an x86_64 machine) [2]. downgrading libxcp to version 1.11.1 seems to fix this temporarily, but doesnt provide a long term solution. [2] http://pastebin.com/f3Py5LDn
Not exactly very helpful, but I wanted to let you know that I had the same problem with a very old/proprietary/binary 32-bit application I'm using.
Running it on the laptop (CoreDuo, no x86_64 support) with Arch i686: no issues. On the x86_64 desktop Arch I had to downgrade lib32-libxcb to 1.11.1, like you.
Any further insight appreciated.
--byte
Have you (both of you) raised a bug [1], sent a mail [2] [3] or dropped a line on IRC [4] ? You may well have found a regression, and it would be very helpful if you could give as much details as possible over the relevant channels :) [1] https://bugs.freedesktop.org/ [2] https://lists.freedesktop.org/mailman/listinfo/Xcb [3] https://lists.x.org/mailman/listinfo/xorg [4] irc://irc.freenode.net/xorg
Have you (both of you) raised a bug [1], sent a mail [2] [3] or dropped a line on IRC [4] ? You may well have found a regression, and it would be very helpful if you could give as much details as possible over the relevant channels :)
[1] https://bugs.freedesktop.org/ [2] https://lists.freedesktop.org/mailman/listinfo/Xcb [3] https://lists.x.org/mailman/listinfo/xorg [4] irc://irc.freenode.net/xorg
why would i file a bug at freedesktop.org? i honestly did not even think about it because in my opinion its adobes fault, and libxcb can most certainly do nothing about it. do you still think its a good idea? i'll try irc and ask for their opinion…
after intensive search it looks like gcc6 is at fault here. more details at the bug tracker [0]. [0] https://bugs.archlinux.org/task/49622
participants (5)
-
Eric Engestrom
-
G. Schlisio
-
Jens Adam
-
John D Jones III
-
Patrick Burroughs (Celti)