You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@subversion.apache.org by Eric Gillespie <ep...@pretzelnet.org> on 2008/05/09 01:05:57 UTC
ra-serf bug
It's been a long time since I made the vague claim "serf breaks
the Python binding"; today I finally decided to try to narrow it
down. I have a tiny little C program that demonstrates the problem:
0 svn% ./serfbug svn://svn/r 84
fetched 118171
# neon
0 svn% ./serfbug https://svn/r 84
fetched 118171
# serf
0 svn% ./serfbug https://svn/r 84
subversion/libsvn_ra_serf/util.c:551: (apr_err=20014)
foo: Error running context: Internal error
*** glibc detected *** double free or corruption (!prev):
0x08098f68 ***
zsh: abort (core dumped) ./serfbug https://svn/r 84
This happens with both 1.5.x and trunk, fresh as of 5 minutes ago
(r31106). The error (but not the double free) goes away if you
pass revision instead of SVN_INVALID_REVNUM.
#include <errno.h>
#include <stdio.h>
#include <stdlib.h>
/* for svn initialization */
#include <apr_errno.h>
#include <apr_general.h>
#include <apr_strings.h>
#include <svn_auth.h>
#include <svn_cmdline.h>
#include <svn_config.h>
#include <svn_dso.h>
#include <svn_nls.h>
#include <svn_pools.h>
/* for svn_ra_get_dir2 */
#include <apr_hash.h>
#include <svn_error.h>
#include <svn_io.h>
#include <svn_path.h>
#include <svn_ra.h>
static svn_error_t *
open_tmp_file(apr_file_t **fp, void *baton, apr_pool_t *pool)
{
const char *base_fn, *base_path, *fn;
SVN_ERR(svn_io_temp_dir(&base_fn, pool));
base_path = svn_path_join(base_fn, "foo", pool);
return svn_io_open_unique_file2(fp, &fn, base_path, ".tmp",
svn_io_file_del_on_close, pool);
}
#define ERR(expr) \
do { \
svn_error_t *err = (expr); \
if (err) { \
fprintf(stderr, "%s:%ld: (apr_err=%d)\nfoo: %s\n", \
err->file, err->line, err->apr_err, \
err->message); \
return 1; \
} \
} while (0)
int
main(int argc, char *argv[])
{
apr_pool_t *pool;
apr_hash_t *cfg_hash;
long revision;
svn_ra_callbacks2_t *callbacks;
if (argc < 3 || (argv[1] && argv[1][0] == '-')) {
fputs("usage: foo URL REVISION\n", stderr);
if (argv[1] && argv[1][0] == '-') {
return 0;
}
return 2;
}
errno = 0;
revision = strtol(argv[2], NULL, 10);
if (errno != 0) {
fprintf(stderr, "foo: strtol(%s): ", argv[2]);
perror("");
return 3;
}
/* Initialize svn. */
{
apr_status_t apr_status;
char buf[1024];
if ((apr_status = apr_initialize()) != APR_SUCCESS) {
apr_strerror(apr_status, buf, sizeof(buf) - 1);
fprintf(stderr, "foo: cannot initialize APR: %s\n", buf);
return 1;
}
svn_dso_initialize();
if (atexit(apr_terminate) != 0) {
fputs("atexit registration failed\n", stderr);
return 1;
}
pool = svn_pool_create(NULL);
svn_utf_initialize(pool);
ERR(svn_nls_init());
ERR(svn_config_get_config(&cfg_hash, NULL, pool));
ERR(svn_ra_initialize(pool));
ERR(svn_ra_create_callbacks(&callbacks, pool));
callbacks->open_tmp_file = open_tmp_file;
ERR(svn_cmdline_setup_auth_baton(&(callbacks->auth_baton),
TRUE, /* non_interactive */
NULL, /* username */
NULL, /* password */
NULL, /* config_dir */
FALSE, /* no_auth_cache */
apr_hash_get(cfg_hash,
SVN_CONFIG_CATEGORY_CONFIG,
APR_HASH_KEY_STRING),
NULL, /* cancel_func */
NULL, /* cancel_baton */
pool));
}
/* Try something simple enough. */
{
svn_ra_session_t *ra;
apr_hash_t *dirents;
svn_revnum_t fetched_rev;
apr_hash_t *props;
ERR(svn_ra_open2(&ra, argv[1], callbacks, NULL, cfg_hash, pool));
ERR(svn_ra_get_dir2(ra, &dirents, &fetched_rev, &props,
"", /* path */
SVN_INVALID_REVNUM, /* revision */
0, /* dirent_fields */
pool));
printf("fetched %ld\n", fetched_rev);
}
return 0;
}
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: ra-serf bug
Posted by Justin Erenkrantz <ju...@erenkrantz.com>.
On Thu, May 8, 2008 at 6:05 PM, Eric Gillespie <ep...@pretzelnet.org> wrote:
> It's been a long time since I made the vague claim "serf breaks
> the Python binding"; today I finally decided to try to narrow it
> down. I have a tiny little C program that demonstrates the problem:
Thanks for the repro. r31107 has the fix. -- justin
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: ra-serf bug
Posted by Eric Gillespie <ep...@pretzelnet.org>.
Lieven Govaerts <sv...@mobsol.be> writes:
> Can you check if r31234 fixes your last two problems?
All 3 problems are fixed now, thanks. The next problem is not so
simple to reproduce, though, so I'm just going to abandon serf
again. I think what I'm seeing is related to
http://subversion.tigris.org/issues/show_bug.cgi?id=2939
Thanks.
--
Eric Gillespie <*> epg@pretzelnet.org
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: ra-serf bug
Posted by Lieven Govaerts <sv...@mobsol.be>.
Eric Gillespie wrote:
> Lieven Govaerts <sv...@mobsol.be> writes:
>
>> Attached is a second and better version of the patch. The 404 is now
>> converted to an error deeper in ra_serf code, which made svn with last
>> patch bail out while finding out the vcc/relative urls.
>
> OK, this fixes that problem, +1 to commit.
>
> Next problem:
>
> svn_ra_check_path is supposed to set kind to svn_node_none if path does
> not exist, but ra-serf raises SVN_ERR_RA_DAV_PATH_NOT_FOUND instead.
>
It looks like it's previous patch that introduces this issue. It's the
one place related to property handling where a 404 is handled, and since
the patch converts a 404 now to an error that error is raised before the
node kind can be set to svn_node_none.
Can you check if r31234 fixes your last two problems?
Lieven
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: ra-serf bug
Posted by Eric Gillespie <ep...@pretzelnet.org>.
Lieven Govaerts <sv...@mobsol.be> writes:
> Attached is a second and better version of the patch. The 404 is now
> converted to an error deeper in ra_serf code, which made svn with last
> patch bail out while finding out the vcc/relative urls.
OK, this fixes that problem, +1 to commit.
Next problem:
svn_ra_check_path is supposed to set kind to svn_node_none if path does
not exist, but ra-serf raises SVN_ERR_RA_DAV_PATH_NOT_FOUND instead.
Demonstration:
#include <errno.h>
#include <stdio.h>
#include <stdlib.h>
/* for svn initialization */
#include <apr_errno.h>
#include <apr_general.h>
#include <apr_hash.h>
#include <apr_strings.h>
#include <svn_auth.h>
#include <svn_cmdline.h>
#include <svn_config.h>
#include <svn_dso.h>
#include <svn_nls.h>
#include <svn_pools.h>
/* for open_tmp_file */
#include <svn_io.h>
#include <svn_path.h>
/* for svn_ra_get_dir2 */
#include <svn_error.h>
#include <svn_ra.h>
#include <svn_types.h>
static svn_error_t *
open_tmp_file(apr_file_t **fp, void *baton, apr_pool_t *pool)
{
const char *base_fn, *base_path, *fn;
SVN_ERR(svn_io_temp_dir(&base_fn, pool));
base_path = svn_path_join(base_fn, "foo", pool);
return svn_io_open_unique_file2(fp, &fn, base_path, ".tmp",
svn_io_file_del_on_close, pool);
}
#define ERR(expr) \
do { \
svn_error_t *err = (expr); \
if (err) { \
fprintf(stderr, "%s:%ld: (apr_err=%d)\nfoo: %s\n", \
err->file, err->line, err->apr_err, \
err->message); \
return 1; \
} \
} while (0)
int
main(int argc, char *argv[])
{
apr_pool_t *pool;
apr_hash_t *cfg_hash;
long revision;
svn_ra_callbacks2_t *callbacks;
if (argc < 3 || (argv[1] && argv[1][0] == '-')) {
fputs("usage: foo URL REVISION\n", stderr);
if (argv[1] && argv[1][0] == '-') {
return 0;
}
return 2;
}
errno = 0;
revision = strtol(argv[2], NULL, 10);
if (errno != 0) {
fprintf(stderr, "foo: strtol(%s): ", argv[2]);
perror("");
return 3;
}
/* Initialize svn. */
{
apr_status_t apr_status;
char buf[1024];
if ((apr_status = apr_initialize()) != APR_SUCCESS) {
apr_strerror(apr_status, buf, sizeof(buf) - 1);
fprintf(stderr, "foo: cannot initialize APR: %s\n", buf);
return 1;
}
svn_dso_initialize();
if (atexit(apr_terminate) != 0) {
fputs("atexit registration failed\n", stderr);
return 1;
}
pool = svn_pool_create(NULL);
svn_utf_initialize(pool);
ERR(svn_nls_init());
ERR(svn_config_get_config(&cfg_hash, NULL, pool));
ERR(svn_ra_initialize(pool));
ERR(svn_ra_create_callbacks(&callbacks, pool));
callbacks->open_tmp_file = open_tmp_file;
ERR(svn_cmdline_setup_auth_baton(&(callbacks->auth_baton),
TRUE, /* non_interactive */
NULL, /* username */
NULL, /* password */
NULL, /* config_dir */
FALSE, /* no_auth_cache */
apr_hash_get(cfg_hash,
SVN_CONFIG_CATEGORY_CONFIG,
APR_HASH_KEY_STRING),
NULL, /* cancel_func */
NULL, /* cancel_baton */
pool));
}
{
svn_ra_session_t *ra;
svn_revnum_t head;
svn_node_kind_t kind;
ERR(svn_ra_open2(&ra, argv[1], callbacks, NULL, cfg_hash, pool));
ERR(svn_ra_get_latest_revnum(ra, &head, pool));
/* Try something simple enough. */
ERR(svn_ra_check_path(ra, "notexist", head, &kind, pool));
if (kind == svn_node_none)
puts("good, kind is none");
else
printf("bad, kind is %d\n", kind);
}
return 0;
}
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: ra-serf bug
Posted by Lieven Govaerts <sv...@mobsol.be>.
Eric Gillespie wrote:
> ping :)
>
> Eric Gillespie <ep...@pretzelnet.org> writes:
>
>> Lieven Govaerts <sv...@mobsol.be> writes:
>>
>>> Index: subversion/libsvn_ra_serf/property.c
>>> ===================================================================
>>> --- subversion/libsvn_ra_serf/property.c (revision 31082)
>>> +++ subversion/libsvn_ra_serf/property.c (working copy)
>>> @@ -612,6 +612,13 @@ svn_ra_serf__wait_for_props(svn_ra_serf__propfind_
>>> svn_error_clear(err);
>>> SVN_ERR(prop_ctx->parser_ctx->error);
>>> }
>>> + if (prop_ctx->parser_ctx->status_code == 404)
>> Looks like status_code is a pointer to the status code, so you
>> need to dereference it.
Right.
>> Even then, I still get the bogus dirent rather than NULL.
>
Attached is a second and better version of the patch. The 404 is now
converted to an error deeper in ra_serf code, which made svn with last
patch bail out while finding out the vcc/relative urls.
There might be other similar situations, so maybe this patch isn't
complete, but atleast it should solve your specific issue.
Lieven
Re: ra-serf bug
Posted by Eric Gillespie <ep...@pretzelnet.org>.
ping :)
Eric Gillespie <ep...@pretzelnet.org> writes:
> Lieven Govaerts <sv...@mobsol.be> writes:
>
> > Index: subversion/libsvn_ra_serf/property.c
> > ===================================================================
> > --- subversion/libsvn_ra_serf/property.c (revision 31082)
> > +++ subversion/libsvn_ra_serf/property.c (working copy)
> > @@ -612,6 +612,13 @@ svn_ra_serf__wait_for_props(svn_ra_serf__propfind_
> > svn_error_clear(err);
> > SVN_ERR(prop_ctx->parser_ctx->error);
> > }
> > + if (prop_ctx->parser_ctx->status_code == 404)
>
> Looks like status_code is a pointer to the status code, so you
> need to dereference it. Even then, I still get the bogus dirent
> rather than NULL.
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: ra-serf bug
Posted by Eric Gillespie <ep...@pretzelnet.org>.
Lieven Govaerts <sv...@mobsol.be> writes:
> Index: subversion/libsvn_ra_serf/property.c
> ===================================================================
> --- subversion/libsvn_ra_serf/property.c (revision 31082)
> +++ subversion/libsvn_ra_serf/property.c (working copy)
> @@ -612,6 +612,13 @@ svn_ra_serf__wait_for_props(svn_ra_serf__propfind_
> svn_error_clear(err);
> SVN_ERR(prop_ctx->parser_ctx->error);
> }
> + if (prop_ctx->parser_ctx->status_code == 404)
Looks like status_code is a pointer to the status code, so you
need to dereference it. Even then, I still get the bogus dirent
rather than NULL.
--
Eric Gillespie <*> epg@pretzelnet.org
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: ra-serf bug
Posted by Lieven Govaerts <sv...@mobsol.be>.
Karl Fogel wrote:
> Eric Gillespie <ep...@pretzelnet.org> writes:
>> Got another one for you: svn_ra_stat is supposed to return NULL
>> for the dirent for paths that do not exist; local/svn/neon do,
>> but serf just returns a bogus one:
>
> Just looking into this briefly between beer and bed (no, really):
>
> Would a non-existent path be indicated by a particular value in
> prop_ctx->status_code after svn_ra_serf__wait_for_props() is done? I'm
> talking about in serf.c:fetch_path_props(). The call to that function
> in svn_ra_serf__stat() seems like the only chance we'd have to discover
> the path does not exist.
>
The equivalent ra_neon code in svn_ra_neon__do_stat uses a 404 response
code when fetching the properties to set dirent to NULL.
Attached (untested, no log msg) patch should solve this problem.
Lieven
Re: ra-serf bug
Posted by Karl Fogel <kf...@red-bean.com>.
Eric Gillespie <ep...@pretzelnet.org> writes:
> Got another one for you: svn_ra_stat is supposed to return NULL
> for the dirent for paths that do not exist; local/svn/neon do,
> but serf just returns a bogus one:
Just looking into this briefly between beer and bed (no, really):
Would a non-existent path be indicated by a particular value in
prop_ctx->status_code after svn_ra_serf__wait_for_props() is done? I'm
talking about in serf.c:fetch_path_props(). The call to that function
in svn_ra_serf__stat() seems like the only chance we'd have to discover
the path does not exist.
Also, by the by: very few calls to svn_ra_serf__deliver_props() wrap it
in an SVN_ERR() check (though two calls do, oddly enough). It's just a
style issue right now, as that function in practice only returns
SVN_NO_ERROR anyway, but it does seem like the error checks ought to be
there, in case the function ever gets more sophisticated.
-Karl
> 0 svn% ./serfbug2 svn://svn/r 84
> dirent NULL
> # neon
> 0 svn% ./serfbug2 https://svn/r 84
> dirent NULL
> # serf
> 0 svn% ./serfbug2 https://svn/r 84
> dirent not NULL:
> kind 0
> size 0
> has_props 0
> created_rev 0
> time 0
> last_author (null)
> *** glibc detected *** double free or corruption (!prev):
> 0x08098f18 ***
> zsh: abort (core dumped) ./serfbug2 https://svn/r 84
>
>
> #include <errno.h>
> #include <stdio.h>
> #include <stdlib.h>
>
> /* for svn initialization */
> #include <apr_errno.h>
> #include <apr_general.h>
> #include <apr_hash.h>
> #include <apr_strings.h>
> #include <svn_auth.h>
> #include <svn_cmdline.h>
> #include <svn_config.h>
> #include <svn_dso.h>
> #include <svn_nls.h>
> #include <svn_pools.h>
> /* for svn_ra_stat */
> #include <svn_error.h>
> #include <svn_io.h>
> #include <svn_path.h>
> #include <svn_ra.h>
> #include <svn_types.h>
>
> static svn_error_t *
> open_tmp_file(apr_file_t **fp, void *baton, apr_pool_t *pool)
> {
> const char *base_fn, *base_path, *fn;
> SVN_ERR(svn_io_temp_dir(&base_fn, pool));
> base_path = svn_path_join(base_fn, "foo", pool);
> return svn_io_open_unique_file2(fp, &fn, base_path, ".tmp",
> svn_io_file_del_on_close, pool);
> }
>
> #define ERR(expr) \
> do { \
> svn_error_t *err = (expr); \
> if (err) { \
> fprintf(stderr, "%s:%ld: (apr_err=%d)\nfoo: %s\n", \
> err->file, err->line, err->apr_err, \
> err->message); \
> return 1; \
> } \
> } while (0)
>
> int
> main(int argc, char *argv[])
> {
> apr_pool_t *pool;
> apr_hash_t *cfg_hash;
> long revision;
> svn_ra_callbacks2_t *callbacks;
>
> if (argc < 3 || (argv[1] && argv[1][0] == '-')) {
> fputs("usage: foo URL REVISION\n", stderr);
> if (argv[1] && argv[1][0] == '-') {
> return 0;
> }
> return 2;
> }
>
> errno = 0;
> revision = strtol(argv[2], NULL, 10);
> if (errno != 0) {
> fprintf(stderr, "foo: strtol(%s): ", argv[2]);
> perror("");
> return 3;
> }
>
> /* Initialize svn. */
> {
> apr_status_t apr_status;
> char buf[1024];
>
> if ((apr_status = apr_initialize()) != APR_SUCCESS) {
> apr_strerror(apr_status, buf, sizeof(buf) - 1);
> fprintf(stderr, "foo: cannot initialize APR: %s\n", buf);
> return 1;
> }
>
> svn_dso_initialize();
>
> if (atexit(apr_terminate) != 0) {
> fputs("atexit registration failed\n", stderr);
> return 1;
> }
>
> pool = svn_pool_create(NULL);
> svn_utf_initialize(pool);
> ERR(svn_nls_init());
>
> ERR(svn_config_get_config(&cfg_hash, NULL, pool));
>
> ERR(svn_ra_initialize(pool));
>
> ERR(svn_ra_create_callbacks(&callbacks, pool));
> callbacks->open_tmp_file = open_tmp_file;
> ERR(svn_cmdline_setup_auth_baton(&(callbacks->auth_baton),
> TRUE, /* non_interactive */
> NULL, /* username */
> NULL, /* password */
> NULL, /* config_dir */
> FALSE, /* no_auth_cache */
> apr_hash_get(cfg_hash,
> SVN_CONFIG_CATEGORY_CONFIG,
> APR_HASH_KEY_STRING),
> NULL, /* cancel_func */
> NULL, /* cancel_baton */
> pool));
> }
>
> /* Try something simple enough. */
> {
> svn_ra_session_t *ra;
> svn_revnum_t head;
> svn_dirent_t *dirent;
> ERR(svn_ra_open2(&ra, argv[1], callbacks, NULL, cfg_hash, pool));
> ERR(svn_ra_get_latest_revnum(ra, &head, pool));
> ERR(svn_ra_stat(ra, "notexist", head, &dirent, pool));
> if (dirent == NULL)
> puts("dirent NULL");
> else
> {
> printf("dirent not NULL:\n"
> " kind %d\n"
> " size %lld\n"
> " has_props %d\n"
> " created_rev %lld\n"
> " time %lld\n"
> " last_author %s\n",
> dirent->kind, (long long)dirent->size, dirent->has_props,
> (long long)dirent->created_rev, (long long)dirent->time,
> dirent->last_author);
> }
> }
>
> return 0;
> }
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
> For additional commands, e-mail: dev-help@subversion.tigris.org
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: ra-serf bug
Posted by Eric Gillespie <ep...@pretzelnet.org>.
Got another one for you: svn_ra_stat is supposed to return NULL
for the dirent for paths that do not exist; local/svn/neon do,
but serf just returns a bogus one:
0 svn% ./serfbug2 svn://svn/r 84
dirent NULL
# neon
0 svn% ./serfbug2 https://svn/r 84
dirent NULL
# serf
0 svn% ./serfbug2 https://svn/r 84
dirent not NULL:
kind 0
size 0
has_props 0
created_rev 0
time 0
last_author (null)
*** glibc detected *** double free or corruption (!prev):
0x08098f18 ***
zsh: abort (core dumped) ./serfbug2 https://svn/r 84
#include <errno.h>
#include <stdio.h>
#include <stdlib.h>
/* for svn initialization */
#include <apr_errno.h>
#include <apr_general.h>
#include <apr_hash.h>
#include <apr_strings.h>
#include <svn_auth.h>
#include <svn_cmdline.h>
#include <svn_config.h>
#include <svn_dso.h>
#include <svn_nls.h>
#include <svn_pools.h>
/* for svn_ra_stat */
#include <svn_error.h>
#include <svn_io.h>
#include <svn_path.h>
#include <svn_ra.h>
#include <svn_types.h>
static svn_error_t *
open_tmp_file(apr_file_t **fp, void *baton, apr_pool_t *pool)
{
const char *base_fn, *base_path, *fn;
SVN_ERR(svn_io_temp_dir(&base_fn, pool));
base_path = svn_path_join(base_fn, "foo", pool);
return svn_io_open_unique_file2(fp, &fn, base_path, ".tmp",
svn_io_file_del_on_close, pool);
}
#define ERR(expr) \
do { \
svn_error_t *err = (expr); \
if (err) { \
fprintf(stderr, "%s:%ld: (apr_err=%d)\nfoo: %s\n", \
err->file, err->line, err->apr_err, \
err->message); \
return 1; \
} \
} while (0)
int
main(int argc, char *argv[])
{
apr_pool_t *pool;
apr_hash_t *cfg_hash;
long revision;
svn_ra_callbacks2_t *callbacks;
if (argc < 3 || (argv[1] && argv[1][0] == '-')) {
fputs("usage: foo URL REVISION\n", stderr);
if (argv[1] && argv[1][0] == '-') {
return 0;
}
return 2;
}
errno = 0;
revision = strtol(argv[2], NULL, 10);
if (errno != 0) {
fprintf(stderr, "foo: strtol(%s): ", argv[2]);
perror("");
return 3;
}
/* Initialize svn. */
{
apr_status_t apr_status;
char buf[1024];
if ((apr_status = apr_initialize()) != APR_SUCCESS) {
apr_strerror(apr_status, buf, sizeof(buf) - 1);
fprintf(stderr, "foo: cannot initialize APR: %s\n", buf);
return 1;
}
svn_dso_initialize();
if (atexit(apr_terminate) != 0) {
fputs("atexit registration failed\n", stderr);
return 1;
}
pool = svn_pool_create(NULL);
svn_utf_initialize(pool);
ERR(svn_nls_init());
ERR(svn_config_get_config(&cfg_hash, NULL, pool));
ERR(svn_ra_initialize(pool));
ERR(svn_ra_create_callbacks(&callbacks, pool));
callbacks->open_tmp_file = open_tmp_file;
ERR(svn_cmdline_setup_auth_baton(&(callbacks->auth_baton),
TRUE, /* non_interactive */
NULL, /* username */
NULL, /* password */
NULL, /* config_dir */
FALSE, /* no_auth_cache */
apr_hash_get(cfg_hash,
SVN_CONFIG_CATEGORY_CONFIG,
APR_HASH_KEY_STRING),
NULL, /* cancel_func */
NULL, /* cancel_baton */
pool));
}
/* Try something simple enough. */
{
svn_ra_session_t *ra;
svn_revnum_t head;
svn_dirent_t *dirent;
ERR(svn_ra_open2(&ra, argv[1], callbacks, NULL, cfg_hash, pool));
ERR(svn_ra_get_latest_revnum(ra, &head, pool));
ERR(svn_ra_stat(ra, "notexist", head, &dirent, pool));
if (dirent == NULL)
puts("dirent NULL");
else
{
printf("dirent not NULL:\n"
" kind %d\n"
" size %lld\n"
" has_props %d\n"
" created_rev %lld\n"
" time %lld\n"
" last_author %s\n",
dirent->kind, (long long)dirent->size, dirent->has_props,
(long long)dirent->created_rev, (long long)dirent->time,
dirent->last_author);
}
}
return 0;
}
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Re: ra-serf bug
Posted by Justin Erenkrantz <ju...@erenkrantz.com>.
On Thu, May 8, 2008 at 6:05 PM, Eric Gillespie <ep...@pretzelnet.org> wrote:
> This happens with both 1.5.x and trunk, fresh as of 5 minutes ago
> (r31106). The error (but not the double free) goes away if you
> pass revision instead of SVN_INVALID_REVNUM.
There's something goofy with the pool logic. If I add:
svn_pool_clear(pool);
to the end of your .c file, I don't get the double free error. I
don't have the time to chase it further - sorry. -- justin
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org