Stabilize dev-libs/crypto++-5.6.1.
Created attachment 252667 [details] build.log Tests fail on amd64
(In reply to comment #1) > Created an attachment (id=252667) [details] > build.log > > Tests fail on amd64 confirming here on amd64, it eats all memory and fails.
x86 stable
(In reply to comment #1) > Created an attachment (id=252667) [details] > build.log > > Tests fail on amd64 What happened to filing new bugs? Stable for HPPA.
ppc64 stable
Tested again on new box. Same problem. Not stable for us. Please restrict test if you want to and call us again
Stable on alpha.
(In reply to comment #7) > Stable on alpha. You haven't committed it.
The pycryptopp project has an issue ticket that is tracking this issue: http://tahoe-lafs.org/trac/pycryptopp/ticket/65
AMD issues were fixed in -r1 since it changes installed binaries and it's only tests that fail. It will compile normally. I'd recommend that amd64 not stable this. That will keep stable users on a properly working version. The revbump will get people using ~amd64 onto a working version as well.
The changes in 5.6.1-r1 are amd64-specific, so I have copied non-amd64 keywords from 5.6.1 to 5.6.1-r1. Please continue stabilization of dev-libs/crypto++-5.6.1-r1.
(In reply to comment #11) > Please continue stabilization of dev-libs/crypto++-5.6.1-r1. Passed test. amd64 ok.
ppc stable
amd64 done. Thanks Agostino
(In reply to comment #9) > The pycryptopp project has an issue ticket that is tracking this issue: > http://tahoe-lafs.org/trac/pycryptopp/ticket/65 To avoid confusing anyone, that's almost certainly not the same bug. This ticket matches the symptoms of the bug in https://sourceforge.net/apps/trac/cryptopp/ticket/6 (which would make it a duplicate of gentoo bugs #322713 and #343169).
(In reply to comment #10) > AMD issues were fixed in -r1 since it changes installed binaries and it's only > tests that fail. It will compile normally. > > I'd recommend that amd64 not stable this. That will keep stable users on a > properly working version. The revbump will get people using ~amd64 onto a > working version as well. As far as I understand, this bug does not only affect tests, it just manifests most obviously in the SHA-1 test (although it is the Rijndael code that is wrong). The upstream fix at http://sourceforge.net/apps/trac/cryptopp/attachment/ticket/6/rijndael.patch should be applied.
sparc done