[OE-core] Reasons for reverting "openssl: prevent ABI break from earlier krogoth releases" ?

Matt Hoosier matt.hoosier at gmail.com
Thu Jul 13 13:04:39 UTC 2017


Back on the Krogoth branch, there was a changed applied briefly last year
to smooth over some symbol-versioning breakage that occurred when OpenSSL
got upgraded past version 1.0.2g:

    openssl: prevent ABI break from earlier krogoth releases

    The backported upgrade to 1.0.2h included an updated GNU LD
    version-script which results in an ABI change. In order to try and
    respect ABI for existing binaries built against fido this commit
    partially reverts the version-script to maintain the existing ABI
    and instead only add the new symbols required by 1.0.2h.

    Suggested-by: Martin Jansa <martin.jansa at gmail.com>
    (From OE-Core rev: 480db6b)

    Signed-off-by: Joshua Lock <joshua.g.lock at intel.com>
    Signed-off-by: Richard Purdie <richard.purdie at linuxfoundation.org>
    Signed-off-by: Armin Kuster <akuster at mvista.com>
    Signed-off-by: Armin Kuster <akuster808 at gmail.com>

However, it got reverted nearly immediately (in
08f85da10b3a7fc6165f163fd0f23784a2c9c8e4), but there wasn't much discussion
in the commit message about why the reversion happened.

I'm running into the same kinds of troubles that the original commit
message cited as justification for implementing backwards-compatibility
with the SSL symbols. Is there any memory about why the old
foo at OPENSSL_1.0.0 versioned symbols ultimately got dropped?

-Matt
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openembedded.org/pipermail/openembedded-core/attachments/20170713/e78759bc/attachment-0002.html>


More information about the Openembedded-core mailing list