You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@subversion.apache.org by Noorul Islam K M <no...@collab.net> on 2011/01/05 15:17:25 UTC
Re: [PATCH] Inconsistent behavior in cat command
Noorul Islam K M <no...@collab.net> writes:
> Julian Foad <ju...@wandisco.com> writes:
>
>> On Thu, 2010-12-02, Noorul Islam K M wrote:
>>
>>> Noorul Islam K M <no...@collab.net> writes:
>>>
>>> > When I was trying to come up with a patch for issue 3713, I observed the
>>> > following.
>>> >
>>> > For example I have two files 1.txt and 2.txt in a repository located at
>>> > file:///tmp/testrepo
>>> >
>>> > svn cat behaves differently for local paths and URLs. See the
>>> > illustration below.
>>> >
>>> > noorul@noorul:/tmp/wc/testrepo$ svn cat 1.txt 2.txt
>>> > 1
>>> > 2
>>> >
>>> > A) Local non-existent target followed by existing target
>>> >
>>> > noorul@noorul:/tmp/wc/testrepo$ svn cat 3.txt 1.txt
>>> > svn: warning: '/tmp/wc/testrepo/3.txt' is not under version control
>>> > 1
>>> >
>>> > B) Non-existent URL followed by existing URL
>>> >
>>> > noorul@noorul:/tmp/wc/testrepo$ svn cat ^/3.txt ^/1.txt
>>> >
>>> > svn: File not found: revision 1, path '/3.txt'
>>> >
>>> >
>>> > In case A, even though the first target was non-existent it performs cat
>>> > operation on the second target but in the case of B, svn errors out at
>>> > the first failure itself.
>>> >
>>> > I am not sure about behavior of other svn commands which accepts
>>> > multiple targets. When I discussed this Julian in IRC, he said a
>>> > discussion is needed to come up with standardized behavior across svn
>>> > commands. Any thoughts?
>>> >
>>>
>>> Any updates?
>>
>> Hi Noorul.
>>
>> A good way to start a discussion of this sort is to:
>>
>> list the possible solutions;
>> find out and describe how the other subcommands behave;
>
> I checked 'svn info' and it behaves the same way for both wc and URL.
>
>> say what you think is good and bad about each possible solution;
>> say which solution you think we should choose.
>>
>> That will make it much easier for readers to respond.
>
> I figured out that for non-existent URL path, svn_client_cat2 returns
> SVN_ERR_FS_NOT_FOUND which needs to be caught by svn_cl__try. Therefore
> I passed this code as one the arguments.
>
> Attached is the patch. All tests pass with this patch.
>
> Log
>
> [[[
> Make 'svn cat' not to error out when one of the URL targets do not
> exist.
>
> * subversion/svn/cat-cmd.c
> (svn_cl__cat): Pass SVN_ERR_FS_NOT_FOUND to svn_cl__try in order to
> catch the error, print warning and proceed with other targets.
>
> Patch by: Noorul Islam K M <noorul{_AT_}collab.net>
> ]]]
>
> Thanks and Regards
> Noorul
>
> Index: subversion/svn/cat-cmd.c
> ===================================================================
> --- subversion/svn/cat-cmd.c (revision 1053010)
> +++ subversion/svn/cat-cmd.c (working copy)
> @@ -78,6 +78,7 @@
> SVN_ERR_UNVERSIONED_RESOURCE,
> SVN_ERR_ENTRY_NOT_FOUND,
> SVN_ERR_CLIENT_IS_DIRECTORY,
> + SVN_ERR_FS_NOT_FOUND,
> SVN_NO_ERROR));
> }
> svn_pool_destroy(subpool);
Initially this thread was monitored by Julian and now I found out that
he is busy. It will be great if someone else can take a look at this
one.
Thanks and Regards
Noorul
Re: [PATCH] Inconsistent behavior in cat command
Posted by Noorul Islam K M <no...@collab.net>.
Gavin Beau Baumanis <ga...@thespidernet.com> writes:
> Ping. This submission has received no new comments.
>
I will be re-submitting this patch once the discussion going on in the
following thread concludes.
http://svn.haxx.se/dev/archive-2011-01/0210.shtml
I think you can remove this patch from your watch list because this will
be combined with some more changes to cat command.
Thanks and Regards
Noorul
>
> On 06/01/2011, at 1:17 AM, Noorul Islam K M wrote:
>
>> Noorul Islam K M <no...@collab.net> writes:
>>
>>> Julian Foad <ju...@wandisco.com> writes:
>>>
>>>> On Thu, 2010-12-02, Noorul Islam K M wrote:
>>>>
>>>>> Noorul Islam K M <no...@collab.net> writes:
>>>>>
>>>>>> When I was trying to come up with a patch for issue 3713, I observed the
>>>>>> following.
>>>>>>
>>>>>> For example I have two files 1.txt and 2.txt in a repository located at
>>>>>> file:///tmp/testrepo
>>>>>>
>>>>>> svn cat behaves differently for local paths and URLs. See the
>>>>>> illustration below.
>>>>>>
>>>>>> noorul@noorul:/tmp/wc/testrepo$ svn cat 1.txt 2.txt
>>>>>> 1
>>>>>> 2
>>>>>>
>>>>>> A) Local non-existent target followed by existing target
>>>>>>
>>>>>> noorul@noorul:/tmp/wc/testrepo$ svn cat 3.txt 1.txt
>>>>>> svn: warning: '/tmp/wc/testrepo/3.txt' is not under version control
>>>>>> 1
>>>>>>
>>>>>> B) Non-existent URL followed by existing URL
>>>>>>
>>>>>> noorul@noorul:/tmp/wc/testrepo$ svn cat ^/3.txt ^/1.txt
>>>>>>
>>>>>> svn: File not found: revision 1, path '/3.txt'
>>>>>>
>>>>>>
>>>>>> In case A, even though the first target was non-existent it performs cat
>>>>>> operation on the second target but in the case of B, svn errors out at
>>>>>> the first failure itself.
>>>>>>
>>>>>> I am not sure about behavior of other svn commands which accepts
>>>>>> multiple targets. When I discussed this Julian in IRC, he said a
>>>>>> discussion is needed to come up with standardized behavior across svn
>>>>>> commands. Any thoughts?
>>>>>>
>>>>>
>>>>> Any updates?
>>>>
>>>> Hi Noorul.
>>>>
>>>> A good way to start a discussion of this sort is to:
>>>>
>>>> list the possible solutions;
>>>> find out and describe how the other subcommands behave;
>>>
>>> I checked 'svn info' and it behaves the same way for both wc and URL.
>>>
>>>> say what you think is good and bad about each possible solution;
>>>> say which solution you think we should choose.
>>>>
>>>> That will make it much easier for readers to respond.
>>>
>>> I figured out that for non-existent URL path, svn_client_cat2 returns
>>> SVN_ERR_FS_NOT_FOUND which needs to be caught by svn_cl__try. Therefore
>>> I passed this code as one the arguments.
>>>
>>> Attached is the patch. All tests pass with this patch.
>>>
>>> Log
>>>
>>> [[[
>>> Make 'svn cat' not to error out when one of the URL targets do not
>>> exist.
>>>
>>> * subversion/svn/cat-cmd.c
>>> (svn_cl__cat): Pass SVN_ERR_FS_NOT_FOUND to svn_cl__try in order to
>>> catch the error, print warning and proceed with other targets.
>>>
>>> Patch by: Noorul Islam K M <noorul{_AT_}collab.net>
>>> ]]]
>>>
>>> Thanks and Regards
>>> Noorul
>>>
>>> Index: subversion/svn/cat-cmd.c
>>> ===================================================================
>>> --- subversion/svn/cat-cmd.c (revision 1053010)
>>> +++ subversion/svn/cat-cmd.c (working copy)
>>> @@ -78,6 +78,7 @@
>>> SVN_ERR_UNVERSIONED_RESOURCE,
>>> SVN_ERR_ENTRY_NOT_FOUND,
>>> SVN_ERR_CLIENT_IS_DIRECTORY,
>>> + SVN_ERR_FS_NOT_FOUND,
>>> SVN_NO_ERROR));
>>> }
>>> svn_pool_destroy(subpool);
>>
>> Initially this thread was monitored by Julian and now I found out that
>> he is busy. It will be great if someone else can take a look at this
>> one.
>>
>> Thanks and Regards
>> Noorul
Re: [PATCH] Inconsistent behavior in cat command
Posted by Gavin Beau Baumanis <ga...@thespidernet.com>.
Ping. This submission has received no new comments.
On 06/01/2011, at 1:17 AM, Noorul Islam K M wrote:
> Noorul Islam K M <no...@collab.net> writes:
>
>> Julian Foad <ju...@wandisco.com> writes:
>>
>>> On Thu, 2010-12-02, Noorul Islam K M wrote:
>>>
>>>> Noorul Islam K M <no...@collab.net> writes:
>>>>
>>>>> When I was trying to come up with a patch for issue 3713, I observed the
>>>>> following.
>>>>>
>>>>> For example I have two files 1.txt and 2.txt in a repository located at
>>>>> file:///tmp/testrepo
>>>>>
>>>>> svn cat behaves differently for local paths and URLs. See the
>>>>> illustration below.
>>>>>
>>>>> noorul@noorul:/tmp/wc/testrepo$ svn cat 1.txt 2.txt
>>>>> 1
>>>>> 2
>>>>>
>>>>> A) Local non-existent target followed by existing target
>>>>>
>>>>> noorul@noorul:/tmp/wc/testrepo$ svn cat 3.txt 1.txt
>>>>> svn: warning: '/tmp/wc/testrepo/3.txt' is not under version control
>>>>> 1
>>>>>
>>>>> B) Non-existent URL followed by existing URL
>>>>>
>>>>> noorul@noorul:/tmp/wc/testrepo$ svn cat ^/3.txt ^/1.txt
>>>>>
>>>>> svn: File not found: revision 1, path '/3.txt'
>>>>>
>>>>>
>>>>> In case A, even though the first target was non-existent it performs cat
>>>>> operation on the second target but in the case of B, svn errors out at
>>>>> the first failure itself.
>>>>>
>>>>> I am not sure about behavior of other svn commands which accepts
>>>>> multiple targets. When I discussed this Julian in IRC, he said a
>>>>> discussion is needed to come up with standardized behavior across svn
>>>>> commands. Any thoughts?
>>>>>
>>>>
>>>> Any updates?
>>>
>>> Hi Noorul.
>>>
>>> A good way to start a discussion of this sort is to:
>>>
>>> list the possible solutions;
>>> find out and describe how the other subcommands behave;
>>
>> I checked 'svn info' and it behaves the same way for both wc and URL.
>>
>>> say what you think is good and bad about each possible solution;
>>> say which solution you think we should choose.
>>>
>>> That will make it much easier for readers to respond.
>>
>> I figured out that for non-existent URL path, svn_client_cat2 returns
>> SVN_ERR_FS_NOT_FOUND which needs to be caught by svn_cl__try. Therefore
>> I passed this code as one the arguments.
>>
>> Attached is the patch. All tests pass with this patch.
>>
>> Log
>>
>> [[[
>> Make 'svn cat' not to error out when one of the URL targets do not
>> exist.
>>
>> * subversion/svn/cat-cmd.c
>> (svn_cl__cat): Pass SVN_ERR_FS_NOT_FOUND to svn_cl__try in order to
>> catch the error, print warning and proceed with other targets.
>>
>> Patch by: Noorul Islam K M <noorul{_AT_}collab.net>
>> ]]]
>>
>> Thanks and Regards
>> Noorul
>>
>> Index: subversion/svn/cat-cmd.c
>> ===================================================================
>> --- subversion/svn/cat-cmd.c (revision 1053010)
>> +++ subversion/svn/cat-cmd.c (working copy)
>> @@ -78,6 +78,7 @@
>> SVN_ERR_UNVERSIONED_RESOURCE,
>> SVN_ERR_ENTRY_NOT_FOUND,
>> SVN_ERR_CLIENT_IS_DIRECTORY,
>> + SVN_ERR_FS_NOT_FOUND,
>> SVN_NO_ERROR));
>> }
>> svn_pool_destroy(subpool);
>
> Initially this thread was monitored by Julian and now I found out that
> he is busy. It will be great if someone else can take a look at this
> one.
>
> Thanks and Regards
> Noorul