I narrowed it down to the code in subversion/libsvn_subr/mergeinfo.c line 892-898 in adjust_remaining_ranges. At that point next_range actually starts before modified_range, so my guess is svn_sort__array_insert has some unexpected behaviour.
x892 svn_merge_range_t *new_modified_range = x
x893 apr_palloc(result_pool, sizeof(*new_modified_range)); x
x894 new_modified_range->start = next_range->end; x
x895 new_modified_range->end = modified_range->end; x
x896 new_modified_range->inheritable = FALSE; x
x897 modified_range->end = next_range->start; x
x898 (*range_index)+=2; x
>x899 svn_sort__array_insert(rangelist, &new_modified_range, x
*range_index); x
x901 /* Recurse with the new range. */ x
x902 adjust_remaining_ranges(rangelist, range_index, result_pool); x
Intuitively it seems to be awfully complicated to expand a range to the end of a change, and then cut it down recursively with adjust_remaining_ranges. My first thought would be to step through "rangelist" and "changes" side by side in svn_rangelist_merge2, and to modify, insert or skip a range in either list until you're at the end. Though obviously I have no idea about all the edge cases the current code most likely fixes...
So how should I proceed from here? Should I open a bug with my findings and the test case?
-----Original Message-----
From: Jens Restemeier [mailto:jens_at_playtonicgames.com]
Sent: 02 July 2017 23:44
To: Johan Corveleyn <jcorvel_at_gmail.com>
Cc: Stefan Sperling <stsp_at_elego.de>; users_at_subversion.apache.org
Subject: Re: "Unable to parse reversed revision range" when merging from trunk to branch
The problem seems to come from svn_rangelist_merge2.
This test program recreates the problem.
#include <svn_pools.h>
#include <svn_mergeinfo.h>
int main(int argc, char * argv[])
{
apr_pool_t *pool;
int exit_code = EXIT_SUCCESS;
svn_error_t *err;
if (svn_cmdline_init("svn", stderr) != EXIT_SUCCESS)
return EXIT_FAILURE;
pool = apr_allocator_owner_get(svn_pool_create_allocator(FALSE));
// 15014-19472,19473-19612*,19613-19614,19615-19630*,19631-19634,19635-20055*
svn_rangelist_t * rangelist = apr_array_make(pool, 1, sizeof(svn_merge_range_t *));
svn_merge_range_t *mrange = apr_pcalloc(pool, sizeof(*mrange));
mrange->start = 15013;
mrange->end = 19472;
mrange->inheritable = TRUE;
APR_ARRAY_PUSH(rangelist, svn_merge_range_t *) = mrange;
mrange = apr_pcalloc(pool, sizeof(*mrange));
mrange->start = 19472;
mrange->end = 19612;
mrange->inheritable = FALSE;
APR_ARRAY_PUSH(rangelist, svn_merge_range_t *) = mrange;
mrange = apr_pcalloc(pool, sizeof(*mrange));
mrange->start = 19612;
mrange->end = 19614;
mrange->inheritable = TRUE;
APR_ARRAY_PUSH(rangelist, svn_merge_range_t *) = mrange;
mrange = apr_pcalloc(pool, sizeof(*mrange));
mrange->start = 19614;
mrange->end = 19630;
mrange->inheritable = FALSE;
APR_ARRAY_PUSH(rangelist, svn_merge_range_t *) = mrange;
mrange = apr_pcalloc(pool, sizeof(*mrange));
mrange->start = 19630;
mrange->end = 19634;
mrange->inheritable = TRUE;
APR_ARRAY_PUSH(rangelist, svn_merge_range_t *) = mrange;
mrange = apr_pcalloc(pool, sizeof(*mrange));
mrange->start = 19634;
mrange->end = 20055;
mrange->inheritable = FALSE;
APR_ARRAY_PUSH(rangelist, svn_merge_range_t *) = mrange;
// 15014-20515*
svn_rangelist_t * changes = apr_array_make(pool, 1, sizeof(svn_merge_range_t *));
mrange = apr_pcalloc(pool, sizeof(*mrange));
mrange->start = 15013;
mrange->end = 20515;
mrange->inheritable = FALSE;
APR_ARRAY_PUSH(changes, svn_merge_range_t *) = mrange;
{
svn_string_t * tmpString;
svn_rangelist_to_string(&tmpString, rangelist, pool);
printf("rangelist %s\n", tmpString->data);
}
{
svn_string_t * tmpString;
svn_rangelist_to_string(&tmpString, changes, pool);
printf("changes %s\n", tmpString->data);
}
svn_rangelist_merge2(rangelist, changes, pool, pool);
{
svn_string_t * tmpString;
svn_rangelist_to_string(&tmpString, rangelist, pool);
printf("result %s\n", tmpString->data);
}
// wrong result
// result 15014-19472,19473-19612*,19613-19614,19615-19630*,19634-19631*,19631-19634,19635-20515*
svn_pool_destroy(pool);
return exit_code;
}
So while this correctly expands the last range from 20055 to 20515 it inserts a wrong inverse range. It’s a bit late, I’ll have a look tomorrow unless someone beats me to it. ;)
> Am 30.06.2017 um 20:50 schrieb Johan Corveleyn <jcorvel_at_gmail.com>:
>
> On Fri, Jun 30, 2017 at 6:10 PM, Jens Christian Restemeier
> <jens_at_playtonicgames.com> wrote:
>> I narrowed it down to somewhere in update_wc_mergeinfo. At the start
>> of the function where it gets the mergeinfo for the root directory it is:
>> /trunk:15014-19472,19473-19612*,19613-19614,19615-19630*,19631-19634,
>> 19635-2
>> 0055*
>>
>> A bit later where this gets parsed again from
>> svn_wc_canonicalize_svn_prop it is:
>> /trunk:15014-19472,19473-19612*,19613-19614,19615-19630*,19634-19631*
>> ,19631-
>> 19634,19635-20511*
>>
>> So somewhere in the mergeinfo update this gets added.
>>
>> Debugging this in gdb is not fun, though...
>
> Maybe it comes from a parent or child of your root directory (either
> in the working copy, or even part of the repository but outside of the
> scope of your working copy). Mergeinfo has some inheritance semantics,
> so it doesn't have to be directly defined on your root directory to be
> applicable.
>
> --
> Johan
Received on 2017-07-03 16:31:20 CEST