[jira] [Comment Edited] (SVN-4840) Merge assertion failure in svn_sort__array_insert

["Julian Foad (Jira)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/SVN-4840?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17025804#comment-17025804 ] 

Julian Foad edited comment on SVN-4840 at 1/29/20 10:54 AM:
------------------------------------------------------------

TODO: Backport to 1.10.x.

TODO: Consider optimization: investigate whether it is worth adding a shortcut code path for non-overlapping non-adjacent ranges. I believe this should not be necessary, as the algorithm has linear performance and is reasonably efficient, but it would be prudent to test and confirm this.


was (Author: julianfoad):
TODO: Backport to 1.10.x.

TODO: Optimization: investigate whether it is worth adding a shortcut code path for non-overlapping non-adjacent ranges.

> Merge assertion failure in svn_sort__array_insert
> -------------------------------------------------
>
>                 Key: SVN-4840
>                 URL: https://issues.apache.org/jira/browse/SVN-4840
>             Project: Subversion
>          Issue Type: Bug
>    Affects Versions: trunk, 1.10.6
>            Reporter: Julian Foad
>            Priority: Major
>
> I am investigating an assertion failure. I received a stack trace without parameter values, from which I quote the relevant part:
>  
> {code:java}
> #4 svn_sort__array_insert (array=?, new_element=?, insert_index=?)
>  at subversion/libsvn_subr/sorts.c:310
>  assert(0 <= insert_index && insert_index <= array->nelts);
> #5 svn_rangelist_merge2 (rangelist=?, chg=?,)
>  at subversion/libsvn_subr/mergeinfo.c:1231
>  svn_sort__array_insert(rangelist, &range_copy, i++);
> #6 svn_mergeinfo_merge2 (mergeinfo=?, changes=?,)
>  at subversion/libsvn_subr/mergeinfo.c:1838
> #7 combine_forked_mergeinfo_props (...)
>  at subversion/libsvn_wc/props.c:138
> #8 apply_single_mergeinfo_prop_change (...)
>  at subversion/libsvn_wc/props.c:1087
> {code}
> This was reported in Subversion 1.10.4. I have reproduced with 1.10.6 and trunk@1871000.
>  



--
This message was sent by Atlassian Jira
(v8.3.4#803005)