[arch-dev-public] boost 1.42 serialization issue
Hi, http://bugs.archlinux.org/task/18470#comment59302 there is a serious issue with boost 1.42 and upstream doesn't have a fix yet. An user suggested that is better to downgrade boost to 1.41. What do you think about this? Wait upstream to fix this and backport it or downgrade? If downgrade, who want to help on rebuilding? -- Ionut
Il 18/03/2010 11:18, Ionut Biru ha scritto:
Hi, http://bugs.archlinux.org/task/18470#comment59302
there is a serious issue with boost 1.42 and upstream doesn't have a fix yet. An user suggested that is better to downgrade boost to 1.41.
What do you think about this? Wait upstream to fix this and backport it or downgrade? If downgrade, who want to help on rebuilding?
-1 I am against the downgrade of boost. We have to find a solution for encfs, and we cannot downgrade and rebuild all of the packages just for encfs. -- Arch Linux Developer http://www.archlinux.org http://www.archlinux.it
On 03/18/2010 01:19 PM, Giovanni Scafora wrote:
Il 18/03/2010 11:18, Ionut Biru ha scritto:
Hi, http://bugs.archlinux.org/task/18470#comment59302
there is a serious issue with boost 1.42 and upstream doesn't have a fix yet. An user suggested that is better to downgrade boost to 1.41.
What do you think about this? Wait upstream to fix this and backport it or downgrade? If downgrade, who want to help on rebuilding?
-1 I am against the downgrade of boost. We have to find a solution for encfs, and we cannot downgrade and rebuild all of the packages just for encfs.
maybe i wasn't clear. encfs is a special case and is broken because of a bug in boost. pasting the comment here:
What distribution are you using? A ticket needs to be filed with that distribution that they should *not* upgrade to boost-1.42, as it can cause silent data lose for programs using boost serialization.
so what i'm understanding from here is that upstream doesn't want us to ship boost 1.42 -- Ionut
On Thu, 2010-03-18 at 12:19 +0100, Giovanni Scafora wrote:
Il 18/03/2010 11:18, Ionut Biru ha scritto:
Hi, http://bugs.archlinux.org/task/18470#comment59302
there is a serious issue with boost 1.42 and upstream doesn't have a fix yet. An user suggested that is better to downgrade boost to 1.41.
What do you think about this? Wait upstream to fix this and backport it or downgrade? If downgrade, who want to help on rebuilding?
-1 I am against the downgrade of boost. We have to find a solution for encfs, and we cannot downgrade and rebuild all of the packages just for encfs.
This is not "just for encfs", but for serious datacorruption. Encfs is an example of something that breaks, but if serialization and deserialization doesn't return the same data as the input data, then things can go seriously wrong.
On Thursday 18 March 2010 11:18:05 Ionut Biru wrote:
Hi, http://bugs.archlinux.org/task/18470#comment59302
there is a serious issue with boost 1.42 and upstream doesn't have a fix yet. An user suggested that is better to downgrade boost to 1.41.
What do you think about this? Wait upstream to fix this and backport it or downgrade? If downgrade, who want to help on rebuilding? Damn...we must proceed with the downgrade. I can rebuild mine and some others packages tomorrow.
-- Andrea
On 18/03/10 21:41, Andrea Scarpino wrote:
On Thursday 18 March 2010 11:18:05 Ionut Biru wrote:
Hi, http://bugs.archlinux.org/task/18470#comment59302
there is a serious issue with boost 1.42 and upstream doesn't have a fix yet. An user suggested that is better to downgrade boost to 1.41.
What do you think about this? Wait upstream to fix this and backport it or downgrade? If downgrade, who want to help on rebuilding? Damn...we must proceed with the downgrade. I can rebuild mine and some others packages tomorrow.
Go for it. It is a small rebuild list and can be moved from [testing] as soon as the build is finished (we know 1.41 works...). I'd appreciate if this was done quickly. I have a toolchain update and the start of the gmp rebuild sitting on my here waiting to upload. Allan
On 03/18/2010 02:42 PM, Allan McRae wrote:
On 18/03/10 21:41, Andrea Scarpino wrote:
On Thursday 18 March 2010 11:18:05 Ionut Biru wrote:
Hi, http://bugs.archlinux.org/task/18470#comment59302
there is a serious issue with boost 1.42 and upstream doesn't have a fix yet. An user suggested that is better to downgrade boost to 1.41.
What do you think about this? Wait upstream to fix this and backport it or downgrade? If downgrade, who want to help on rebuilding? Damn...we must proceed with the downgrade. I can rebuild mine and some others packages tomorrow.
Go for it. It is a small rebuild list and can be moved from [testing] as soon as the build is finished (we know 1.41 works...).
I'd appreciate if this was done quickly. I have a toolchain update and the start of the gmp rebuild sitting on my here waiting to upload.
Allan
todo: https://dev.archlinux.org/todo/29/ i'll upload boost 1.41 in couples of minutes. -- Ionut
Il 18/03/2010 11:18, Ionut Biru ha scritto:
Hi, http://bugs.archlinux.org/task/18470#comment59302
there is a serious issue with boost 1.42 and upstream doesn't have a fix yet. An user suggested that is better to downgrade boost to 1.41.
What do you think about this? Wait upstream to fix this and backport it or downgrade? If downgrade, who want to help on rebuilding?
I'm just rebuilding mkvtoolnix and source-highlight against testing/boost. -- Arch Linux Developer http://www.archlinux.org http://www.archlinux.it
participants (5)
-
Allan McRae
-
Andrea Scarpino
-
Giovanni Scafora
-
Ionut Biru
-
Jan de Groot