You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@subversion.apache.org by Philip Martin <ph...@wandisco.com> on 2010/11/01 09:30:00 UTC
Re: Issue 3727 - svn export into the current directory
Noorul Islam K M <no...@collab.net> writes:
> @@ -79,8 +84,17 @@
> {
> to = APR_ARRAY_IDX(targets, 1, const char *);
>
> + /* Get the RA connection. */
> + SVN_ERR(svn_client__ra_session_from_path(&ra_session, &revnum,
> + &url, truefrom, NULL,
> + &peg_revision,
> + &(opt_state->start_revision),
> + ctx, pool));
> +
> + SVN_ERR(svn_ra_check_path(ra_session, "", revnum, &kind, pool));
> +
> /* If given the cwd, pretend we weren't given anything. */
> - if (strcmp("", to) == 0)
> + if ((strcmp("", to) == 0) && (kind == svn_node_file))
> to = svn_path_uri_decode(svn_uri_basename(truefrom, pool), pool);
> else
> /* svn_cl__eat_peg_revisions() but only on one target */
svn_ra_check_path is expensive, as it involves a round trip to the
server, and output is not always used. Can we avoid the call when the
kind is not needed?
--
Philip
Re: Issue 3727 - svn export into the current directory
Posted by Philip Martin <ph...@wandisco.com>.
Noorul Islam K M <no...@collab.net> writes:
> [[[
> Fix issue 3727: Fix regression caused by r880559
>
> * subversion/tests/cmdline/export_tests.py
> (export_to_explicit_cwd): Remove XFail.
>
> * subversion/svn/export-cmd.c
> (svn_cl__export): Move logic into subversion/libsvn_client/export.c
>
> * subversion/libsvn_client/export.c
> (svn_client_export5): If the explicit TO path is '.', treat it as if it weren't
> given only if the source is a file.
>
> Patch by: Noorul Islam K M <noorul{_AT_}collab.net>
> ]]]
I tweaked it to use svn_path_is_empty, to pass a NULL pool where the
basename does not need to be allocated, and to update the
svn_client_export5 docstring. r1029751. Thanks!
--
Philip
Re: Issue 3727 - svn export into the current directory
Posted by Noorul Islam K M <no...@collab.net>.
Philip Martin <ph...@wandisco.com> writes:
> Noorul Islam K M <no...@collab.net> writes:
>
>> Philip Martin <ph...@wandisco.com> writes:
>>
>>> Noorul Islam K M <no...@collab.net> writes:
>>>
>>>> @@ -79,8 +84,17 @@
>>>> {
>>>> to = APR_ARRAY_IDX(targets, 1, const char *);
>>>>
>>>> + /* Get the RA connection. */
>>>> + SVN_ERR(svn_client__ra_session_from_path(&ra_session, &revnum,
>>>> + &url, truefrom, NULL,
>>>> + &peg_revision,
>>>> + &(opt_state->start_revision),
>>>> + ctx, pool));
>>>> +
>>>> + SVN_ERR(svn_ra_check_path(ra_session, "", revnum, &kind, pool));
>>>> +
>>>> /* If given the cwd, pretend we weren't given anything. */
>>>> - if (strcmp("", to) == 0)
>>>> + if ((strcmp("", to) == 0) && (kind == svn_node_file))
>>>> to = svn_path_uri_decode(svn_uri_basename(truefrom, pool), pool);
>>>> else
>>>> /* svn_cl__eat_peg_revisions() but only on one target */
>>>
>>> svn_ra_check_path is expensive, as it involves a round trip to the
>>> server, and output is not always used. Can we avoid the call when the
>>> kind is not needed?
>>
>> But I think here we need to check kind to branch. Do you mean to say
>> that we don't need kind in all cases?
>
> If the call strcpm("", to) returns non-zero then kind is never used.
>
> Including client.h is wrong as well.
Here is a modified patch. I moved the logic into export.c where we
already use ra_session.
[[[
Fix issue 3727: Fix regression caused by r880559
* subversion/tests/cmdline/export_tests.py
(export_to_explicit_cwd): Remove XFail.
* subversion/svn/export-cmd.c
(svn_cl__export): Move logic into subversion/libsvn_client/export.c
* subversion/libsvn_client/export.c
(svn_client_export5): If the explicit TO path is '.', treat it as if it weren't
given only if the source is a file.
Patch by: Noorul Islam K M <noorul{_AT_}collab.net>
]]]
Thanks and Regards
Noorul
Re: Issue 3727 - svn export into the current directory
Posted by Philip Martin <ph...@wandisco.com>.
Noorul Islam K M <no...@collab.net> writes:
> Philip Martin <ph...@wandisco.com> writes:
>
>> Noorul Islam K M <no...@collab.net> writes:
>>
>>> @@ -79,8 +84,17 @@
>>> {
>>> to = APR_ARRAY_IDX(targets, 1, const char *);
>>>
>>> + /* Get the RA connection. */
>>> + SVN_ERR(svn_client__ra_session_from_path(&ra_session, &revnum,
>>> + &url, truefrom, NULL,
>>> + &peg_revision,
>>> + &(opt_state->start_revision),
>>> + ctx, pool));
>>> +
>>> + SVN_ERR(svn_ra_check_path(ra_session, "", revnum, &kind, pool));
>>> +
>>> /* If given the cwd, pretend we weren't given anything. */
>>> - if (strcmp("", to) == 0)
>>> + if ((strcmp("", to) == 0) && (kind == svn_node_file))
>>> to = svn_path_uri_decode(svn_uri_basename(truefrom, pool), pool);
>>> else
>>> /* svn_cl__eat_peg_revisions() but only on one target */
>>
>> svn_ra_check_path is expensive, as it involves a round trip to the
>> server, and output is not always used. Can we avoid the call when the
>> kind is not needed?
>
> But I think here we need to check kind to branch. Do you mean to say
> that we don't need kind in all cases?
If the call strcpm("", to) returns non-zero then kind is never used.
Including client.h is wrong as well.
--
Philip
Re: Issue 3727 - svn export into the current directory
Posted by Noorul Islam K M <no...@collab.net>.
Philip Martin <ph...@wandisco.com> writes:
> Noorul Islam K M <no...@collab.net> writes:
>
>> @@ -79,8 +84,17 @@
>> {
>> to = APR_ARRAY_IDX(targets, 1, const char *);
>>
>> + /* Get the RA connection. */
>> + SVN_ERR(svn_client__ra_session_from_path(&ra_session, &revnum,
>> + &url, truefrom, NULL,
>> + &peg_revision,
>> + &(opt_state->start_revision),
>> + ctx, pool));
>> +
>> + SVN_ERR(svn_ra_check_path(ra_session, "", revnum, &kind, pool));
>> +
>> /* If given the cwd, pretend we weren't given anything. */
>> - if (strcmp("", to) == 0)
>> + if ((strcmp("", to) == 0) && (kind == svn_node_file))
>> to = svn_path_uri_decode(svn_uri_basename(truefrom, pool), pool);
>> else
>> /* svn_cl__eat_peg_revisions() but only on one target */
>
> svn_ra_check_path is expensive, as it involves a round trip to the
> server, and output is not always used. Can we avoid the call when the
> kind is not needed?
But I think here we need to check kind to branch. Do you mean to say
that we don't need kind in all cases?
Thanks and Regards
Noorul