Large patch/diff refuses to apply

[Sean Bruno]

I'm doing a large amount of change to ixgbe(4) in support of IFLIB
implementations and running into failures when trying to apply large
diffs. This is causing phabricator reviews to be unuseable as well.

I've setup two trees to test this. The first tree is used to generate
the diff and the second (vanilla) is used to apply the diff. The entire
patch fails to apply, so I'm assuming that the size of the diff is
failing because of a sanity check or something.

https://people.freebsd.org/~sbruno/test.diff

% svn stat sys/dev/ixgbe/if_ix.c
M sys/dev/ixgbe/if_ix.c

% svn diff sys/dev/ixgbe/if_ix.c > ~/test.diff


% cd ../vanilla/

% patch -p0 < ~/test.diff
Hmm... Looks like a unified diff to me...
The text leading up to this was:
--------------------------
|Index: sys/dev/ixgbe/if_ix.c
|===================================================================
|--- sys/dev/ixgbe/if_ix.c (revision 306673)
|+++ sys/dev/ixgbe/if_ix.c (working copy)
--------------------------
Patching file sys/dev/ixgbe/if_ix.c using Plan A...
Hunk #1 failed at 1.
1 out of 1 hunks failed--saving rejects to sys/dev/ixgbe/if_ix.c.rej
done

% ls -l sys/dev/ixgbe/if_ix.c*
-rw-r--r-- 1 sbruno wheel 166692 Oct 6 06:41 sys/dev/ixgbe/if_ix.c
-rw-r--r-- 1 sbruno wheel 166692 Oct 6 06:37 sys/dev/ixgbe/if_ix.c.orig
-rw-r--r-- 1 sbruno wheel 172779 Oct 6 06:41 sys/dev/ixgbe/if_ix.c.rej

% svn stat sys/dev/ixgbe/
? sys/dev/ixgbe/if_ix.c.orig

[Slawa Olhovchenkov]

On Thu, Oct 06, 2016 at 07:57:57AM -0600, Sean Bruno wrote:

> I'm doing a large amount of change to ixgbe(4) in support of IFLIB
> implementations and running into failures when trying to apply large
> diffs. This is causing phabricator reviews to be unuseable as well.
>
> I've setup two trees to test this. The first tree is used to generate
> the diff and the second (vanilla) is used to apply the diff. The entire
> patch fails to apply, so I'm assuming that the size of the diff is
> failing because of a sanity check or something.

No. This is expanded/collapsed keywords issuse:

===
-******************************************************************************/
-/*$FreeBSD$*/
===

svn diff over repo generate patch w/ collapsed keywords.
At working copy all keywords expanded.
_______________________________________________
freebsd...@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/freebsd-current
To unsubscribe, send any mail to "freebsd-curre...@freebsd.org"

[Sean Bruno]

On 10/06/16 08:23, Slawa Olhovchenkov wrote:
> On Thu, Oct 06, 2016 at 07:57:57AM -0600, Sean Bruno wrote:
>
>> I'm doing a large amount of change to ixgbe(4) in support of IFLIB
>> implementations and running into failures when trying to apply large
>> diffs. This is causing phabricator reviews to be unuseable as well.
>>
>> I've setup two trees to test this. The first tree is used to generate
>> the diff and the second (vanilla) is used to apply the diff. The entire
>> patch fails to apply, so I'm assuming that the size of the diff is
>> failing because of a sanity check or something.
>
> No. This is expanded/collapsed keywords issuse:
>
> ===
> -******************************************************************************/
> -/*$FreeBSD$*/
> ===
>
> svn diff over repo generate patch w/ collapsed keywords.
> At working copy all keywords expanded.
>

Ah, I see. Thank you.

I am regenerating the failed files now. That seems to work (if I leave
the keywords alone).

sean

[Alan Somers]

Also, I think the "svn patch" command (as opposed to plain "patch")
can deal with the RCS keywords.

[Sean Bruno]

On 10/06/16 09:00, Alan Somers wrote:
> On Thu, Oct 6, 2016 at 8:41 AM, Sean Bruno <sbr...@freebsd.org> wrote:
>>
>>
>> On 10/06/16 08:23, Slawa Olhovchenkov wrote:
>>> On Thu, Oct 06, 2016 at 07:57:57AM -0600, Sean Bruno wrote:
>>>
>>>> I'm doing a large amount of change to ixgbe(4) in support of IFLIB
>>>> implementations and running into failures when trying to apply large
>>>> diffs. This is causing phabricator reviews to be unuseable as well.
>>>>
>>>> I've setup two trees to test this. The first tree is used to generate
>>>> the diff and the second (vanilla) is used to apply the diff. The entire
>>>> patch fails to apply, so I'm assuming that the size of the diff is
>>>> failing because of a sanity check or something.
>>>
>>> No. This is expanded/collapsed keywords issuse:
>>>
>>> ===
>>> -******************************************************************************/
>>> -/*$FreeBSD$*/
>>> ===
>>>
>>> svn diff over repo generate patch w/ collapsed keywords.
>>> At working copy all keywords expanded.
>>>
>>
>> Ah, I see. Thank you.
>>
>> I am regenerating the failed files now. That seems to work (if I leave
>> the keywords alone).
>>
>> sean
>>
>
> Also, I think the "svn patch" command (as opposed to plain "patch")
> can deal with the RCS keywords.
>

The use of "svn cp" also seems to generate a lot of failures as well.

I do want to maintain history between certain files when I go to commit.
e.g. svn cp fileA new_fileA, but this seems to confuse the crap out of
diff/patch no matter what I do. This leads to suffering with
phabricator as well.

For now, I'm not going to do the "svn cp" part of my commit, but I will
try and do this when I am ready to shove my work into current.

sean

[Sean Bruno]

A good example of what I was trying to do:
svn mkdir sys/modules/ix/ix_iflib
svn cp sys/modules/ix/Makefile sys/modules/ix/ix/ix_iflib/Makefile
<edit sys/modules/ix/ix_iflib/Makefile>

That seems to really confuse things.

sean

[Dimitry Andric]

On 06 Oct 2016, at 17:23, Sean Bruno <sbr...@freebsd.org> wrote:
>
> On 10/06/16 09:19, Sean Bruno wrote:
>>
>> On 10/06/16 09:00, Alan Somers wrote:
>>> On Thu, Oct 6, 2016 at 8:41 AM, Sean Bruno <sbr...@freebsd.org> wrote:
>>>>
>>>> On 10/06/16 08:23, Slawa Olhovchenkov wrote:
>>>>> On Thu, Oct 06, 2016 at 07:57:57AM -0600, Sean Bruno wrote:

..

> A good example of what I was trying to do:
> svn mkdir sys/modules/ix/ix_iflib
> svn cp sys/modules/ix/Makefile sys/modules/ix/ix/ix_iflib/Makefile
> <edit sys/modules/ix/ix_iflib/Makefile>
>
> That seems to really confuse things.

Try using: svn diff --show-copies-as-adds

-Dimitry

[Sean Bruno]

On 10/06/16 10:58, Dimitry Andric wrote:
> On 06 Oct 2016, at 17:23, Sean Bruno <sbr...@freebsd.org> wrote:
>>
>> On 10/06/16 09:19, Sean Bruno wrote:
>>>
>>> On 10/06/16 09:00, Alan Somers wrote:
>>>> On Thu, Oct 6, 2016 at 8:41 AM, Sean Bruno <sbr...@freebsd.org> wrote:
>>>>>
>>>>> On 10/06/16 08:23, Slawa Olhovchenkov wrote:
>>>>>> On Thu, Oct 06, 2016 at 07:57:57AM -0600, Sean Bruno wrote:

> ...

>> A good example of what I was trying to do:
>> svn mkdir sys/modules/ix/ix_iflib
>> svn cp sys/modules/ix/Makefile sys/modules/ix/ix/ix_iflib/Makefile
>> <edit sys/modules/ix/ix_iflib/Makefile>
>>
>> That seems to really confuse things.
>
> Try using: svn diff --show-copies-as-adds
>
> -Dimitry
>

Most of what I was trying to do was with arcanist. The review
generation seemed to be fine, but applying the diff from arcanist seemed
to blow up.

sean