[
https://issues.apache.org/jira/browse/ETCH-172?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13120502#comment-13120502
]
Thomas Marsh commented on ETCH-172:
-----------------------------------
I noticed there are some 64 bit routines on Bob Jenkins' site. See
http://burtleburtle.net/bob/c/lookup8.c and
http://burtleburtle.net/bob/hash/evahash.html#hash64. I haven't had time to
further debug the problem to verify the exact problem, but if it is the hashing
routines, perhaps this is helpful.
> 64-bit platforms not supported in C binding
> -------------------------------------------
>
> Key: ETCH-172
> URL: https://issues.apache.org/jira/browse/ETCH-172
> Project: Etch
> Issue Type: Bug
> Components: c-binding
> Environment: Linux x86_64 (tested on CentOS 5.6)
> Reporter: Thomas Marsh
> Priority: Blocker
>
> It appears that there may be unsafe assumptions about bit size in the hash
> table implementation. All unit tests which require idname generation
> currently fail on 64 bit platforms. Rather than requiring that clients be
> compiled with "-m32" on Linux (as the CMakeList.txt explicitly adds), the
> platform should instead compile and run cleanly on 64 bit platforms and
> without requiring a 32bit library stack. The hashtable implementation should
> run natively in 64 bit space, or as a workaround, limit itself to 32bit hash
> space.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators:
https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira
