Date: Tue, 31 Jul 2018 12:27:32 +0200 From: Solar Designer <solar@...nwall.com> To: owl-dev@...ts.openwall.com Subject: Re: e2fsprogs package update On Sun, Jul 29, 2018 at 07:38:42PM +0300, Sergey Kurokhtin wrote: > I'm using VirtualBox VM with Owl current x86_64. > Sure i686 will be next step. OK. > Ok. I'll take a look to the -alt patch. Let me clarify. I didn't mean any pre-existing -alt patch. I meant combining all of the changes in ALT's package into a patch that we'll call -alt. We'd use this as a general approach to reduce maintenance costs of packages in Owl where the changes are based on ALT's. Right now, our GnuPG package is already in a similar state. > I investigated the failing tests and here some results: > > m_hugefile tries to create 4T file inside /tmp which is mounted into tmpfs. > So the following error occurs: > "m_hugefile/script: line 40: 104607 File size limit exceeded" > There are two possible solutions: > -- create test file outside /tmp; > -- skip this test if /tmp is mounted into tmpfs. > Something like this: > df | grep '/tmp' | grep 'tmpfs' > /dev/null 2>&1 && ( echo "Skipped"; exit > 0) Maybe something like: trap 'exit 0' XFSZ > Other tests could be fixed using corrections in the 'expect' files with > patch. > All calculations are correct but outcome from the e2fsprogs utilites > slightly different than it reflected in 'expect' files. Do you know or have a guess as to why it's slightly different? Is it desirable for us to expect these slightly different results, and fail if we start getting the originally expected results (perhaps so that we're reminded to revert the patch when that happens?) > Will proceed with creating patches. OK. Maybe just two: -alt (applied first as it'd be upstream for us) and -owl-tests. Alexander
Powered by blists - more mailing lists
Confused about mailing lists and their use? Read about mailing lists on Wikipedia and check out these guidelines on proper formatting of your messages.