You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@subversion.apache.org by Kannan <ka...@collab.net> on 2009/10/26 13:55:35 UTC
[PATCH] Fix a deprecation warning.
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Log:
Resolve a deprecation warning.
* subversion/svn/info-cmd.c
(svn_cl__info): Use `svn_client_info3()'.
Patch by: Kannan R <ka...@collab.net>
- --
Regards,
Kannan
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.6 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/
iQEVAwUBSuWqV3lTqcY7ytmIAQIYfQgAgy9fcixVKuyZRLPk3G6amY4UEdCFJxNm
wJ+ptmhrjLj1BmVspjPeWyVYlu+55WVel0S1SJQ7OugT6cHkhbt89UBGO0iajTPz
0lFP/IlOXLbmFs/0W2M0jKz1O8NGwq8vWwEO19ksdJGNPv2ePkuffqCV5BPeH9P9
MRrRI2l6DE701JTq2wTCyxKOLji+4Ym6mTLFgrdmzcOjxyiUNpCvKLjAW0B78DBs
znNC/C3mf5tib1oyHhul3yWA6lGNkjDhsKeLYEAcps8Mi2hwDICuva/B/lVZT9Yz
m8tJpDDt7ta+pUBf9PvTTx8MmzbwVjfbcbusHeYz4Qg14JiDBw7XZg==
=OU03
-----END PGP SIGNATURE-----
------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=2411390
Re: [PATCH] Fix a deprecation warning.
Posted by Kannan <ka...@collab.net>.
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Julian Foad wrote:
> On Wed, 2009-11-11, Kannan wrote:
>> Stefan Sperling wrote:
>>> Thanks. There are still more test failures caused by this patch.
>>> Please run all tests and make sure they pass. You can run 'make check'
>>> in the working copy root to run all tests, and it will tell you which
>>> tests are failing.
>> Should've run `make check` rather than testing info tests alone, sorry
>> for that. Two tests(basic_tests.py 26 and changelist_tests.py 4) failed
>> expecting relative paths for info, which is fixed with this patch to
>> expect absolute ones. Make check passes.
>>
>> [[[
>> Updated Log:
>> Resolve a deprecation warning.
>
> Hold on a minute... This doesn't just resolve a deprecation warning, it
> also changes the command-line output format. The log message should say
> so.
>
Oops, missed the whole point. Thank you for pointing out.
- --
Thanks & Regards,
Kannan
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.6 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/
iQEVAwUBSvwap3lTqcY7ytmIAQImawf9EEsl2FI3zVzSevxdg7e4fi6lFXGviqXY
grLXIkjSV0jl73+KU+biT6zPTkwyxKcBqM76c/cjUYiqrx5kgHAcOikJnWMJhFoS
E+SkUB0NMIFQzB/XLQ0rGnzuDJoTO0vOE5TtZ3D6IzIO+D9V+i6se/SjV8nWvk/h
AN79V0PHgxb5N9yNRxlZLDMrokQT7FqSEPgh1TMkkNu9cECAL/FgA+wTnj/kFIPm
jgt2eWHGKcKUdmgvaw9a9Gh6+JvSqIsIcSP1oE69YVVKQ6rmX4lukMv0jvu7T3Wo
aqxI+D871b+ItxB6C1Nsk8xD7iIy4vfS5hXxdhD10oD3fQr+DBzo/A==
=G1hs
-----END PGP SIGNATURE-----
------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=2417060
Re: [PATCH] Fix a deprecation warning.
Posted by Kannan <ka...@collab.net>.
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Julian Foad wrote:
>> If we're to use info_receiver_relpath_wrapper() and a baton to pass the
>> relative path and other info, won't it resemble the implementation(for
>> backward compatibility)which is at present?
>
> Kannan,
>
> I'm sorry, I don't understand that question. What will resemble what
> implementation, in what way?
Currently libsvn_client/deprecated.c holds the .._relpath_wrapper() for
the deprecated implementation of `svn_client_info2()' where relative
paths are being passed through the baton. So, just had a doubt whether
implementing `svn_client_info3()' similar to this using a baton to pass
relative path, would resemble the deprecated implementation(or a copy of
code).
>>> FWIW, centralization and sanitization of relative path output in an
>> absolute path world is issue 3464:
>>> http://subversion.tigris.org/issues/show_bug.cgi?id=3464
>> Thank you Hyrum, in that case implementing using a separate wrapper(now)
>> for info alone becomes obsolete right?
>
> Do you mean "I can change the output format (from relative to absolute)
> now, because there is an issue filed for fixing that (changing it back
> to relative) later" ? If so, I disagree.
Again sorry for the brevity. Meant if I'm to write a wrapper for info
alone then won't it become obsolete as other subcommands too require the
same feature. Later I got clear from your clarifications that we need a
generic function(for all subcommands) to convert the format to relative
(if I'm right?).
> We COULD apply this patch now, if people think that is the best thing to
> do, but I am concerned that if we start making changes like this, we may
> not implement the necessary infrastructure to make it easy to "fix" it
> later, and also we may start to lose track of what user-visible changes
> we have made. If we do it properly straight away, then we will know we
> are on track.
>
> So if you can, please find a way to keep the output format the same, and
> you get bonus points if you can create a shared function to help do
> that.
>
Sure, will give it a shot. Thank you for the comments Julian.
- --
Thanks & Regards,
Kannan
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.6 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/
iQEVAwUBSwqadHlTqcY7ytmIAQLTQggAr3NvHSc5Z5SEJcX9yFTqFfXKQMu6j9bG
CWBW9IS2Q4olQSLfF7ar6YSw070SyyAaF59yc1wvV0MXshJMlfXRF3iHbzdcvVrM
lbtJfpUzwUQQk/0GF9p3Nsv2Oy3cKOh6rGnye7n6twMkr0uuOBDKCfNqhgbrwApb
Lc6KBNy4EYHMC/hzw7XZLiD6FOlIH2mma1KHWX/zKAj2GWXKyIXIcFNZ6YdX1uqm
jJ4FCJGeeUF7IgzOyeTHcrq5sgsOboboftmheKQ5JUPpJxX0mtfzxhal6+OjcxeZ
Yhdl1/s5nyht+pgAqLr8gEuGMR/v14MB3IvTvG/sC+dqE6aY9Q9Tdg==
=1JUD
-----END PGP SIGNATURE-----
Re: [PATCH] Fix a deprecation warning.
Posted by Julian Foad <ju...@btopenworld.com>.
On Thu, 2009-11-12, Kannan wrote:
> Hyrum K. Wright wrote:
> > On Nov 12, 2009, at 5:53 AM, Julian Foad wrote:
> >> Stefan Sperling wrote:
> >>> On Thu, Nov 12, 2009 at 11:01:10AM +0000, Julian Foad wrote:
> >>>> We try hard not to change the output format. Why do we need to change it
> >>>> here? The other subcommands manage to convert abs paths back to rel
> >>>> paths for display.
> >>> How? Can you show an example?
> >> I'm sure I recall somebody mentioning on dev@ that they did this for at
> >> least one other command earlier this year.
> >>
> >> A search for "to_rel" finds info_receiver_relpath_wrapper() which is
> >> undocumented, aargh, but looks like it does that sort of thing.
> >>
> >> Basically you just have to know what your base path is, transfer it
> >> (through a baton if necessary) to the output function, and strip it off
> >> the received abspath (making sure the base path is in canonical absolute
> >> form as well).
>
> If we're to use info_receiver_relpath_wrapper() and a baton to pass the
> relative path and other info, won't it resemble the implementation(for
> backward compatibility)which is at present?
Kannan,
I'm sorry, I don't understand that question. What will resemble what
implementation, in what way?
> > FWIW, centralization and sanitization of relative path output in an
> absolute path world is issue 3464:
> > http://subversion.tigris.org/issues/show_bug.cgi?id=3464
>
> Thank you Hyrum, in that case implementing using a separate wrapper(now)
> for info alone becomes obsolete right?
Do you mean "I can change the output format (from relative to absolute)
now, because there is an issue filed for fixing that (changing it back
to relative) later" ? If so, I disagree.
We COULD apply this patch now, if people think that is the best thing to
do, but I am concerned that if we start making changes like this, we may
not implement the necessary infrastructure to make it easy to "fix" it
later, and also we may start to lose track of what user-visible changes
we have made. If we do it properly straight away, then we will know we
are on track.
So if you can, please find a way to keep the output format the same, and
you get bonus points if you can create a shared function to help do
that.
Thanks.
- Julian
------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=2423354
Please start new threads on the <de...@subversion.apache.org> mailing list.
To subscribe to the new list, send an empty e-mail to <de...@subversion.apache.org>.
Re: [PATCH] Fix a deprecation warning.
Posted by Kannan <ka...@collab.net>.
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Hyrum K. Wright wrote:
> On Nov 12, 2009, at 5:53 AM, Julian Foad wrote:
>
>> Stefan Sperling wrote:
>>> On Thu, Nov 12, 2009 at 11:01:10AM +0000, Julian Foad wrote:
>>>> We try hard not to change the output format. Why do we need to change it
>>>> here? The other subcommands manage to convert abs paths back to rel
>>>> paths for display.
>>> How? Can you show an example?
>> I'm sure I recall somebody mentioning on dev@ that they did this for at
>> least one other command earlier this year.
>>
>> A search for "to_rel" finds info_receiver_relpath_wrapper() which is
>> undocumented, aargh, but looks like it does that sort of thing.
>>
>> Basically you just have to know what your base path is, transfer it
>> (through a baton if necessary) to the output function, and strip it off
>> the received abspath (making sure the base path is in canonical absolute
>> form as well).
If we're to use info_receiver_relpath_wrapper() and a baton to pass the
relative path and other info, won't it resemble the implementation(for
backward compatibility)which is at present?
> FWIW, centralization and sanitization of relative path output in an absolute path world is issue 3464:
> http://subversion.tigris.org/issues/show_bug.cgi?id=3464
Thank you Hyrum, in that case implementing using a separate wrapper(now)
for info alone becomes obsolete right?
- --
Thanks & Regards,
Kannan
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.6 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/
iQEVAwUBSvwcAHlTqcY7ytmIAQLucgf9GFhvDanmU77Ba+/JcrmnotsK30g0ak/J
z7gVNdTNnJ+coW5rYwcQX7mUvstBMDvFXhEPYsJ6AXlKSXdJqKt0F8zGFp4smbDS
rC+UzrILz7Lopuoo0RzBfd7E6kLFOa9u3lLKxrj1x4rsU+fgDXidXQNfo7JwGF9W
d845fi998fQrg0MHPEByAX7OivOpaZtb39AtvncxTmcgUzjHvs17OoGPz7MzgrzU
MVRyXZPPz3Wtpl2HxI8NnIKxlPx2xvjB7Fn8pTETZKq7yHlvaYtc/qzFOhVhtr7L
XrkgSJGtbz7m0XJkPIPssxAEMrIGtUVJxeRh6nwyrUZiN6oPnW/LRg==
=dUVf
-----END PGP SIGNATURE-----
------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=2417065
Re: [PATCH] Fix a deprecation warning.
Posted by "Hyrum K. Wright" <hy...@hyrumwright.org>.
On Nov 12, 2009, at 5:53 AM, Julian Foad wrote:
> Stefan Sperling wrote:
>> On Thu, Nov 12, 2009 at 11:01:10AM +0000, Julian Foad wrote:
>>> We try hard not to change the output format. Why do we need to change it
>>> here? The other subcommands manage to convert abs paths back to rel
>>> paths for display.
>>
>> How? Can you show an example?
>
> I'm sure I recall somebody mentioning on dev@ that they did this for at
> least one other command earlier this year.
>
> A search for "to_rel" finds info_receiver_relpath_wrapper() which is
> undocumented, aargh, but looks like it does that sort of thing.
>
> Basically you just have to know what your base path is, transfer it
> (through a baton if necessary) to the output function, and strip it off
> the received abspath (making sure the base path is in canonical absolute
> form as well).
FWIW, centralization and sanitization of relative path output in an absolute path world is issue 3464:
http://subversion.tigris.org/issues/show_bug.cgi?id=3464
-Hyrum
------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=2417033
Re: [PATCH] Fix a deprecation warning.
Posted by Julian Foad <ju...@wandisco.com>.
Stefan Sperling wrote:
> On Thu, Nov 12, 2009 at 11:01:10AM +0000, Julian Foad wrote:
> > We try hard not to change the output format. Why do we need to change it
> > here? The other subcommands manage to convert abs paths back to rel
> > paths for display.
>
> How? Can you show an example?
I'm sure I recall somebody mentioning on dev@ that they did this for at
least one other command earlier this year.
A search for "to_rel" finds info_receiver_relpath_wrapper() which is
undocumented, aargh, but looks like it does that sort of thing.
Basically you just have to know what your base path is, transfer it
(through a baton if necessary) to the output function, and strip it off
the received abspath (making sure the base path is in canonical absolute
form as well).
- Julian
------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=2416997
Re: [PATCH] Fix a deprecation warning.
Posted by Stefan Sperling <st...@elego.de>.
On Thu, Nov 12, 2009 at 11:01:10AM +0000, Julian Foad wrote:
> We try hard not to change the output format. Why do we need to change it
> here? The other subcommands manage to convert abs paths back to rel
> paths for display.
How? Can you show an example?
Stefan
------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=2416991
Re: [PATCH] Fix a deprecation warning.
Posted by Julian Foad <ju...@wandisco.com>.
On Wed, 2009-11-11, Kannan wrote:
> Stefan Sperling wrote:
> > Thanks. There are still more test failures caused by this patch.
> > Please run all tests and make sure they pass. You can run 'make check'
> > in the working copy root to run all tests, and it will tell you which
> > tests are failing.
>
> Should've run `make check` rather than testing info tests alone, sorry
> for that. Two tests(basic_tests.py 26 and changelist_tests.py 4) failed
> expecting relative paths for info, which is fixed with this patch to
> expect absolute ones. Make check passes.
>
> [[[
> Updated Log:
> Resolve a deprecation warning.
Hold on a minute... This doesn't just resolve a deprecation warning, it
also changes the command-line output format. The log message should say
so.
We try hard not to change the output format. Why do we need to change it
here? The other subcommands manage to convert abs paths back to rel
paths for display.
- Julian
> * subversion/svn/info-cmd.c
> (svn_cl__info): Make the path as absolute and pass to the
> new `svn_client_info3()' and update the receiver type to
> `svn_info_receiver2_t'.
>
> * subversion/libsvn_client/info.c
> (svn_client_info3, push_dir_info,
> crawl_entries, found_entry_baton): Update the receiver type to
> `svn_info_receiver2_t'.
>
> [in subversion/tests/cmdline]:
>
> * info-tests.py
> (info_on_added_file, info_on_mkdir): Expect absolute paths for `svn
> info'.
>
> * basic_tests.py
> (basic_info): Same.
>
> * changelist_tests.py
> (info_with_changelists): Same.
>
> ]]]
------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=2416970
Re: [PATCH] Fix a deprecation warning.
Posted by Kannan <ka...@collab.net>.
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Gavin 'Beau' Baumanis wrote:
> Hi Kannan,
>
> Ping! (Just in case it fell off your radar)
> I can't see this patch has been submitted nor has any more comments been made to it since Julian last asked about the output format.
>
> Don't forget to ask for any help you might need.
Thank you Gavin, didn't receive the following message of Julian somehow!
>> We try not to change the output format without very good reason. Why
>> can't we keep the output format the same?
Thank you Julian, misunderstood that wc-ng needs absolute path format
even in output. With regard to my other queries, they are here:
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=2417065
- --
Thanks & Regards,
Kannan
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.6 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/
iQEVAwUBSwoq/3lTqcY7ytmIAQKsRwf/RJhAVSmK87lnVZoRMaBHdAkfegg4UW3M
tNmi8STP6LC0mOg5ivTsj0M4if7y6eYpfUZ/uKpnNbttTi7YATLhYtOKqzBIvbQR
MdQSlwyQHGLubA9RAkPwdrXW5ylgR/c4BkHfqbZ0ILxchdddxzYSUDKOqMxiI7BA
4h41Kuwm231CoUs5eEDorbgkJzsonebs90V9oADd2TfrVdPJ9Mnt4pI9naf0ybKE
BQ7yeFPdBANKNnmDW2+lVol924mFS48r38Nt7b4XZ6wp9yr9xcF9MtBToi5YWgds
Jy4PnSuFO2rf1+Rfs8Aoz2p9XlmTRIe3LkPUPMp6GK8Cs0gH8EOlaQ==
=+tOb
-----END PGP SIGNATURE-----
Re: [PATCH] Fix a deprecation warning.
Posted by Gavin Baumanis <ga...@thespidernet.com>.
Hi Kannan,
Ping! (Just in case it fell off your radar)
I can't see this patch has been submitted nor has any more comments been made to it since Julian last asked about the output format.
Don't forget to ask for any help you might need.
Gavin.
On 12/11/2009, at 21:00 , Julian Foad wrote:
> On Wed, 2009-11-11, Kannan wrote:
>> Stefan Sperling wrote:
>>> Thanks. There are still more test failures caused by this patch.
>>> Please run all tests and make sure they pass. You can run 'make check'
>>> in the working copy root to run all tests, and it will tell you which
>>> tests are failing.
>>
>> Should've run `make check` rather than testing info tests alone, sorry
>> for that. Two tests(basic_tests.py 26 and changelist_tests.py 4) failed
>> expecting relative paths for info, which is fixed with this patch to
>> expect absolute ones. Make check passes.
>>
>> [[[
>> Updated Log:
>> Resolve a deprecation warning.
>
> Hold on a minute... This patch doesn't just resolve a deprecation
> warning, it changes the command-line output format. You need to say
> things like that in the log message.
>
> We try not to change the output format without very good reason. Why
> can't we keep the output format the same?
>
> - Julian
>
>
>> * subversion/svn/info-cmd.c
>> (svn_cl__info): Make the path as absolute and pass to the
>> new `svn_client_info3()' and update the receiver type to
>> `svn_info_receiver2_t'.
>>
>> * subversion/libsvn_client/info.c
>> (svn_client_info3, push_dir_info,
>> crawl_entries, found_entry_baton): Update the receiver type to
>> `svn_info_receiver2_t'.
>>
>> [in subversion/tests/cmdline]:
>>
>> * info-tests.py
>> (info_on_added_file, info_on_mkdir): Expect absolute paths for `svn
>> info'.
>>
>> * basic_tests.py
>> (basic_info): Same.
>>
>> * changelist_tests.py
>> (info_with_changelists): Same.
>>
>> ]]]
>
> ------------------------------------------------------
> http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=2417506
------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=2420501
Re: [PATCH] Fix a deprecation warning.
Posted by Julian Foad <ju...@btopenworld.com>.
On Wed, 2009-11-11, Kannan wrote:
> Stefan Sperling wrote:
> > Thanks. There are still more test failures caused by this patch.
> > Please run all tests and make sure they pass. You can run 'make check'
> > in the working copy root to run all tests, and it will tell you which
> > tests are failing.
>
> Should've run `make check` rather than testing info tests alone, sorry
> for that. Two tests(basic_tests.py 26 and changelist_tests.py 4) failed
> expecting relative paths for info, which is fixed with this patch to
> expect absolute ones. Make check passes.
>
> [[[
> Updated Log:
> Resolve a deprecation warning.
Hold on a minute... This patch doesn't just resolve a deprecation
warning, it changes the command-line output format. You need to say
things like that in the log message.
We try not to change the output format without very good reason. Why
can't we keep the output format the same?
- Julian
> * subversion/svn/info-cmd.c
> (svn_cl__info): Make the path as absolute and pass to the
> new `svn_client_info3()' and update the receiver type to
> `svn_info_receiver2_t'.
>
> * subversion/libsvn_client/info.c
> (svn_client_info3, push_dir_info,
> crawl_entries, found_entry_baton): Update the receiver type to
> `svn_info_receiver2_t'.
>
> [in subversion/tests/cmdline]:
>
> * info-tests.py
> (info_on_added_file, info_on_mkdir): Expect absolute paths for `svn
> info'.
>
> * basic_tests.py
> (basic_info): Same.
>
> * changelist_tests.py
> (info_with_changelists): Same.
>
> ]]]
------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=2417506
Re: [PATCH] Fix a deprecation warning.
Posted by Kannan <ka...@collab.net>.
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Stefan Sperling wrote:
> Thanks. There are still more test failures caused by this patch.
> Please run all tests and make sure they pass. You can run 'make check'
> in the working copy root to run all tests, and it will tell you which
> tests are failing.
Should've run `make check` rather than testing info tests alone, sorry
for that. Two tests(basic_tests.py 26 and changelist_tests.py 4) failed
expecting relative paths for info, which is fixed with this patch to
expect absolute ones. Make check passes.
[[[
Updated Log:
Resolve a deprecation warning.
* subversion/svn/info-cmd.c
(svn_cl__info): Make the path as absolute and pass to the
new `svn_client_info3()' and update the receiver type to
`svn_info_receiver2_t'.
* subversion/libsvn_client/info.c
(svn_client_info3, push_dir_info,
crawl_entries, found_entry_baton): Update the receiver type to
`svn_info_receiver2_t'.
[in subversion/tests/cmdline]:
* info-tests.py
(info_on_added_file, info_on_mkdir): Expect absolute paths for `svn
info'.
* basic_tests.py
(basic_info): Same.
* changelist_tests.py
(info_with_changelists): Same.
]]]
>> Index: ../../../../subversion/svn/info-cmd.c
>> ===================================================================
>> --- ../../../../subversion/svn/info-cmd.c (revision 40430)
>> +++ ../../../../subversion/svn/info-cmd.c (working copy)
>
> Please don't send patches containing paths like this in the future,
> they are a bit combersome to apply.
> Just run 'svn diff' from the root of your svn trunk working copy,
> that's a bit easier for people consuming the patch (because they
> don't need to count how many path components to strip).
Thank you for pointing out, I'll follow hereafter.
- --
Thanks & Regards,
Kannan
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.6 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/
iQEVAwUBSvrT4XlTqcY7ytmIAQJu8Qf/YpocOAHaNcGJqLotSNHZx+Slp8/LnWUF
gI/5Je82vhQ5kVfsnvaMSKYkfiA9awG2t/fNWiZ9ISGn+fis1AzYXpQ7WuEjNrBs
0BLDb9qvFNhy/AmLSBlMJ69KcPYS4sLKYy0UfZFR6iTXW1esaKC2tqV4s4lW205v
npY/BgSG5bygjQB5iWYPQ5XoGkRR+D8ZsytjvgiHbJyHNWj7F0eUIL424/TLBZxI
Tbu14+1zIA1ex3bH+tHI8xplwcV4jIXPAwpbrYfPqOTD8UmrlI9v0SgO+aMtMugY
NdvFUCNed20UBu7xcs1+SzeBmS5C5J58neLPkwUO1hMNlt/bkEQ/4A==
=YctK
-----END PGP SIGNATURE-----
------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=2416605
Re: [PATCH] Fix a deprecation warning.
Posted by Stefan Sperling <st...@elego.de>.
On Mon, Nov 09, 2009 at 03:34:18PM +0530, Kannan wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> Stefan Sperling wrote:
> > It's not clear to me yet how we're going to translate absolute paths
> > to relative paths in all cases.
> >
> > I think for now it's fine to adjust the tests to expect absolute paths.
> > See patch_unidiff_absolute_paths in patch_tests.py for an example of
> > how we can deal with absolute paths in tests.
>
> Thank you Stefan. As per your comments, updated the tests to expect
> absolute paths and they pass now. Attaching the updated patch herewith.
Thanks. There are still more test failures caused by this patch.
Please run all tests and make sure they pass. You can run 'make check'
in the working copy root to run all tests, and it will tell you which
tests are failing.
> Index: ../../../../subversion/svn/info-cmd.c
> ===================================================================
> --- ../../../../subversion/svn/info-cmd.c (revision 40430)
> +++ ../../../../subversion/svn/info-cmd.c (working copy)
Please don't send patches containing paths like this in the future,
they are a bit combersome to apply.
Just run 'svn diff' from the root of your svn trunk working copy,
that's a bit easier for people consuming the patch (because they
don't need to count how many path components to strip).
Stefan
------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=2416169
Re: [PATCH] Fix a deprecation warning.
Posted by Kannan <ka...@collab.net>.
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Stefan Sperling wrote:
> It's not clear to me yet how we're going to translate absolute paths
> to relative paths in all cases.
>
> I think for now it's fine to adjust the tests to expect absolute paths.
> See patch_unidiff_absolute_paths in patch_tests.py for an example of
> how we can deal with absolute paths in tests.
Thank you Stefan. As per your comments, updated the tests to expect
absolute paths and they pass now. Attaching the updated patch herewith.
Updated Log:
Resolve a deprecation warning.
* subversion/svn/info-cmd.c
(svn_cl__info): Make the path as absolute and pass to the
new `svn_client_info3()' and update the receiver type to
`svn_info_receiver2_t'.
* subversion/libsvn_client/info.c
(svn_client_info3, push_dir_info,
crawl_entries, found_entry_baton): Update the receiver type to
`svn_info_receiver2_t'.
* subversion/tests/cmdline/info-tests.py
(info_on_added_file, info_on_mkdir): Expect absolute paths for `svn
info'.
Patch by: Kannan R <ka...@collab.net>
- --
Thanks & Regards,
Kannan
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.6 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/
iQEVAwUBSvfpIXlTqcY7ytmIAQJDLwf9ED7wF0ypMYcF3cFV2Dje4xJKNiQQoGtI
QQHpFZTM/dcl09DWbPInoW6//1yQ2+1Htm+an6d9f8C/dHZe0wRE4UrjnInpE2sD
eupFw23AErv4o65F1u2PdjyKiAFNcF25m74dduyVpGBeOU7jJF3P6hy8CLvhwNzj
WCzo25SO7+QgyaP0KRdDDGRY8+IAdaugGM/HHMAg5udcGZHyULS/TAZo1RZciK2X
pwD/D317vM+SncwKMQcs5FH6vxvXIGsbKG1kq+UtxIpyXznan7pl5Adtp/1Gnb8K
TCinhblpOjA2sZ+S7ayo6buQmWVlAbEQ6NJxFFLShlWVmPiwxSy5Fg==
=wfNS
-----END PGP SIGNATURE-----
------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=2415724
Re: [PATCH] Fix a deprecation warning.
Posted by Stefan Sperling <st...@elego.de>.
On Fri, Nov 06, 2009 at 12:59:02PM +0530, Kannan wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> Stefan Sperling wrote:
> > Sorry, this got lost.
> >
> > I've just tested your patch. It changes output -- svn info shows
> > absolute paths instead of relative paths. This in itself is not
> > a huge problem, we know that it happens as part of the wc-ng transition,
> > and we may have to figure out a way to make absolute paths relative
> > for command line client output if possible.
> >
> > But the problem is that the changed output breaks tests.
> > I've only tested info_tests with your patch, and info tests 2 and 3
> > are failing. Please run all tests and make sure they pass before sending
> > patches. If you cannot get a test to pass, mention this in your patch
> > submission and ask for help. It's important to run *all* tests on a
> > patch, because breakage sometimes shows up in unexpected places. This
> > is quite time consuming but necessary.
>
> Thank you for the comments Stefan. I could see the info tests 2 and 3
> fail, as the path returned is an absolute one. Should we make a
> transition again to relative paths for the output?
It's not clear to me yet how we're going to translate absolute paths
to relative paths in all cases.
I think for now it's fine to adjust the tests to expect absolute paths.
See patch_unidiff_absolute_paths in patch_tests.py for an example of
how we can deal with absolute paths in tests.
Stefan
------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=2415706
Re: [PATCH] Fix a deprecation warning.
Posted by Kannan <ka...@collab.net>.
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Stefan Sperling wrote:
> Sorry, this got lost.
>
> I've just tested your patch. It changes output -- svn info shows
> absolute paths instead of relative paths. This in itself is not
> a huge problem, we know that it happens as part of the wc-ng transition,
> and we may have to figure out a way to make absolute paths relative
> for command line client output if possible.
>
> But the problem is that the changed output breaks tests.
> I've only tested info_tests with your patch, and info tests 2 and 3
> are failing. Please run all tests and make sure they pass before sending
> patches. If you cannot get a test to pass, mention this in your patch
> submission and ask for help. It's important to run *all* tests on a
> patch, because breakage sometimes shows up in unexpected places. This
> is quite time consuming but necessary.
Thank you for the comments Stefan. I could see the info tests 2 and 3
fail, as the path returned is an absolute one. Should we make a
transition again to relative paths for the output?
- --
Thanks & Regards,
Kannan
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.6 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/
iQEVAwUBSvPQPnlTqcY7ytmIAQLfBgf/TVbgMFL/dt45WD4K4aZ4gOZqMAznyveu
gsPczy6mucAbZZi6gGgbu1WNhsYfMANVgmvujLhcsWi8zm9qtcq6PNdXSyC4aw2D
iX5FhXwJAoZnRK7Kigg0bYwKsGb6WT7WmCY1mZRVb+Z71h/GGe0h80MUBP9PuqbL
TAftSZqkjLcz2Mo2Zs3jK9Mk+vZDyuqgG0yOPW3qKIg7P7vtUx1x1LFyv7yZYJ+U
Lfua6hDwaCEIcpwl7+E2S4vJZmlhpEdkuPf8MQtrDRotFarLpR64nxeX+/MDm/B4
IBPRSRErEaPVBd+UgB1p9MmoKdkJBiwD3WSZxkpOiwOQwd1dg543qg==
=gvRC
-----END PGP SIGNATURE-----
------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=2415007
Re: [PATCH] Fix a deprecation warning.
Posted by Stefan Sperling <st...@elego.de>.
On Tue, Nov 03, 2009 at 10:10:50PM +0530, Kannan wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> > Stefan Sperling wrote:
> >> On Mon, Oct 26, 2009 at 07:25:35PM +0530, Kannan wrote:
> >>> -----BEGIN PGP SIGNED MESSAGE-----
> >>> Hash: SHA1
> >>>
> >>>
> >>> Log:
> >>> Resolve a deprecation warning.
> >>>
> >>> * subversion/svn/info-cmd.c
> >>> (svn_cl__info): Use `svn_client_info3()'.
> >>>
> >>> Patch by: Kannan R <ka...@collab.net>
> >
> >>> Index: subversion/svn/info-cmd.c
> >>> ===================================================================
> >>> --- subversion/svn/info-cmd.c (revision 40224)
> >>> +++ subversion/svn/info-cmd.c (working copy)
> >>> @@ -527,7 +527,7 @@
> >>> && (peg_revision.kind == svn_opt_revision_unspecified))
> >>> peg_revision.kind = svn_opt_revision_head;
> >>>
> >>> - err = svn_client_info2(truepath,
> >>> + err = svn_client_info3(truepath,
> >>> &peg_revision, &(opt_state->start_revision),
> >>> receiver, NULL, opt_state->depth,
> >>> opt_state->changelists, ctx, subpool);
> >> Are you sure you don't need to update the receiver also?
> >
> >> svn_client_info2's docstring says:
> >
> >> /** Similar to svn_client_info3, but uses an svn_info_receiver_t instead of
> >> * a svn_info_receiver2_t.
> >
> > Thank you stsp. After your comments in IRC, changed the receiver type to
> > `svn_info_receiver2_t' and passed the path as absolute to
> > `svn_client_info3()' as it expects an 'abspath_or_url', also to avoid an
> > assertion failed error in `svn_wc__db_read_info()' and just tested `svn
> > info'. It runs fine, before I could change anything in `print_info()'.
> > I'm attaching the patch herewith. Please correct me if I'm wrong.
> >
>
> Any updates on this thread?
Sorry, this got lost.
I've just tested your patch. It changes output -- svn info shows
absolute paths instead of relative paths. This in itself is not
a huge problem, we know that it happens as part of the wc-ng transition,
and we may have to figure out a way to make absolute paths relative
for command line client output if possible.
But the problem is that the changed output breaks tests.
I've only tested info_tests with your patch, and info tests 2 and 3
are failing. Please run all tests and make sure they pass before sending
patches. If you cannot get a test to pass, mention this in your patch
submission and ask for help. It's important to run *all* tests on a
patch, because breakage sometimes shows up in unexpected places. This
is quite time consuming but necessary.
Stefan
------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=2414699
Re: [PATCH] Fix a deprecation warning.
Posted by Kannan <ka...@collab.net>.
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
> Stefan Sperling wrote:
>> On Mon, Oct 26, 2009 at 07:25:35PM +0530, Kannan wrote:
>>> -----BEGIN PGP SIGNED MESSAGE-----
>>> Hash: SHA1
>>>
>>>
>>> Log:
>>> Resolve a deprecation warning.
>>>
>>> * subversion/svn/info-cmd.c
>>> (svn_cl__info): Use `svn_client_info3()'.
>>>
>>> Patch by: Kannan R <ka...@collab.net>
>
>>> Index: subversion/svn/info-cmd.c
>>> ===================================================================
>>> --- subversion/svn/info-cmd.c (revision 40224)
>>> +++ subversion/svn/info-cmd.c (working copy)
>>> @@ -527,7 +527,7 @@
>>> && (peg_revision.kind == svn_opt_revision_unspecified))
>>> peg_revision.kind = svn_opt_revision_head;
>>>
>>> - err = svn_client_info2(truepath,
>>> + err = svn_client_info3(truepath,
>>> &peg_revision, &(opt_state->start_revision),
>>> receiver, NULL, opt_state->depth,
>>> opt_state->changelists, ctx, subpool);
>> Are you sure you don't need to update the receiver also?
>
>> svn_client_info2's docstring says:
>
>> /** Similar to svn_client_info3, but uses an svn_info_receiver_t instead of
>> * a svn_info_receiver2_t.
>
> Thank you stsp. After your comments in IRC, changed the receiver type to
> `svn_info_receiver2_t' and passed the path as absolute to
> `svn_client_info3()' as it expects an 'abspath_or_url', also to avoid an
> assertion failed error in `svn_wc__db_read_info()' and just tested `svn
> info'. It runs fine, before I could change anything in `print_info()'.
> I'm attaching the patch herewith. Please correct me if I'm wrong.
>
Any updates on this thread?
- --
Regards,
Kannan
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.6 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/
iQEVAwUBSvBdEnlTqcY7ytmIAQI5Nwf/TSTNoY6CKR+1J7xSmCHRt26/vv6pp47G
oN12I7/hTxkv3KDeFnv0UZzvU5Za7x++DSYvhA9q/Kuw5jBRk1231xvq0IEe9nS7
CHFtsmhq4E9Od1CfPmeu3ptD/rBUX0aoeIsA9nnFWsff6w8HUxdAxVwhP00RAYLw
+Az89ITzEp3m4K+P/aDS9sjkl5CNnDUycmO6tlSvjO0/LarJ8s61b/bj6ZIG35Jo
MnCtmSAqedobPq/zYE6eQgMi5pQPodCOPi1NDoLeL8CkHqC7Wk5w98yt5LqKuB75
AQsqc3KbNKpxS11hIBkF6tj6ror3qBwkhPdK6QS6p5ALzBxSaJXpMw==
=tomn
-----END PGP SIGNATURE-----
------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=2414153
Re: [PATCH] Fix a deprecation warning.
Posted by Kannan <ka...@collab.net>.
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Stefan Sperling wrote:
> On Mon, Oct 26, 2009 at 07:25:35PM +0530, Kannan wrote:
>> -----BEGIN PGP SIGNED MESSAGE-----
>> Hash: SHA1
>>
>>
>> Log:
>> Resolve a deprecation warning.
>>
>> * subversion/svn/info-cmd.c
>> (svn_cl__info): Use `svn_client_info3()'.
>>
>> Patch by: Kannan R <ka...@collab.net>
>
>> Index: subversion/svn/info-cmd.c
>> ===================================================================
>> --- subversion/svn/info-cmd.c (revision 40224)
>> +++ subversion/svn/info-cmd.c (working copy)
>> @@ -527,7 +527,7 @@
>> && (peg_revision.kind == svn_opt_revision_unspecified))
>> peg_revision.kind = svn_opt_revision_head;
>>
>> - err = svn_client_info2(truepath,
>> + err = svn_client_info3(truepath,
>> &peg_revision, &(opt_state->start_revision),
>> receiver, NULL, opt_state->depth,
>> opt_state->changelists, ctx, subpool);
>
> Are you sure you don't need to update the receiver also?
>
> svn_client_info2's docstring says:
>
> /** Similar to svn_client_info3, but uses an svn_info_receiver_t instead of
> * a svn_info_receiver2_t.
Thank you stsp. After your comments in IRC, changed the receiver type to
`svn_info_receiver2_t' and passed the path as absolute to
`svn_client_info3()' as it expects an 'abspath_or_url', also to avoid an
assertion failed error in `svn_wc__db_read_info()' and just tested `svn
info'. It runs fine, before I could change anything in `print_info()'.
I'm attaching the patch herewith. Please correct me if I'm wrong.
- --
Regards,
Kannan
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.6 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/
iQEVAwUBSuhDBHlTqcY7ytmIAQLk+wgAjezuiOofayrjtaJKv+9JdKdTN2wL+kmz
CAZXbjpJ9LC3DH8EGEcWrj1eikMalzmW2RrYDiMdIrPOZfcsXM/kpP/jkpg560Sj
BZ4vl9+1BqnbfpRESB1fLYJg8UIlkVJrNoUvl2snYy/oUq6pUtcO4To/70Y7josz
DvmowuKrhofD1Gd6v4uN95gL8N1LUx2FX86xY0dj7XcMn2YZRY4E3Lg1EoGx2Pky
+DBELVYBw0HARo5zuawCynhX09sc3ye7q9WqRogd6GOXgX1Jqbjyd4traLRvSltW
CZKbEv58GibwgSbesxr4rcI2NoSMYITcOo60VIcaFt0/S82QejCqaA==
=w/o1
-----END PGP SIGNATURE-----
------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=2412106
Re: [PATCH] Fix a deprecation warning.
Posted by Stefan Sperling <st...@elego.de>.
On Mon, Oct 26, 2009 at 07:25:35PM +0530, Kannan wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
>
> Log:
> Resolve a deprecation warning.
>
> * subversion/svn/info-cmd.c
> (svn_cl__info): Use `svn_client_info3()'.
>
> Patch by: Kannan R <ka...@collab.net>
> Index: subversion/svn/info-cmd.c
> ===================================================================
> --- subversion/svn/info-cmd.c (revision 40224)
> +++ subversion/svn/info-cmd.c (working copy)
> @@ -527,7 +527,7 @@
> && (peg_revision.kind == svn_opt_revision_unspecified))
> peg_revision.kind = svn_opt_revision_head;
>
> - err = svn_client_info2(truepath,
> + err = svn_client_info3(truepath,
> &peg_revision, &(opt_state->start_revision),
> receiver, NULL, opt_state->depth,
> opt_state->changelists, ctx, subpool);
Are you sure you don't need to update the receiver also?
svn_client_info2's docstring says:
/** Similar to svn_client_info3, but uses an svn_info_receiver_t instead of
* a svn_info_receiver2_t.
Stefan
------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=2411436