[U-Boot] [ PATCH] patman: cover letter shows like 00/xx if more than 10 patches

[Simon Glass]

Hi Josh,

On 7 April 2015 at 20:38, Josh Wu <josh.wu at atmel.com> wrote:
>
> On 4/8/2015 4:49 AM, Simon Glass wrote:
>>
>> Hi Josh,
>>
>> On 6 April 2015 at 20:31, Josh Wu <josh.wu at atmel.com> wrote:
>>>
>>> HI, Simon
>>>
>>> Thanks for the feedback.
>>>
>>>
>>> On 4/6/2015 2:31 AM, Simon Glass wrote:
>>>>
>>>> Hi Josh,
>>>>
>>>> On 31 March 2015 at 20:54, Josh Wu <josh.wu at atmel.com> wrote:
>>>>>
>>>>> Hi, Simon
>>>>>
>>>>> On 4/1/2015 10:04 AM, Simon Glass wrote:
>>>>>>
>>>>>> Hi Josh,
>>>>>>
>>>>>> On 30 March 2015 at 19:54, Josh Wu <josh.wu at atmel.com> wrote:
>>>>>>>
>>>>>>> Make cover letter shows like 0/x, 00/xx and 000/xxx etc.
>>>>>>>
>>>>>>> Signed-off-by: Josh Wu <josh.wu at atmel.com>
>>>>>>> ---
>>>>>>
>>>>>> This is a quirk of patman that I've grown comfortable with. Still, we
>>>>>> should fix it. Thanks for the patch.
>>>>>>
>>>>>>>     tools/patman/patchstream.py | 9 +++++++--
>>>>>>>     1 file changed, 7 insertions(+), 2 deletions(-)
>>>>>>>
>>>>>>> diff --git a/tools/patman/patchstream.py b/tools/patman/patchstream.py
>>>>>>> index 8c3a0ec..4bfb1e9 100644
>>>>>>> --- a/tools/patman/patchstream.py
>>>>>>> +++ b/tools/patman/patchstream.py
>>>>>>> @@ -468,8 +468,13 @@ def InsertCoverLetter(fname, series, count):
>>>>>>>         prefix = series.GetPatchPrefix()
>>>>>>>         for line in lines:
>>>>>>>             if line.startswith('Subject:'):
>>>>>>> -            # TODO: if more than 10 patches this should save 00/xx,
>>>>>>> not
>>>>>>> 0/xx
>>>>>>> -            line = 'Subject: [%s 0/%d] %s\n' % (prefix, count,
>>>>>>> text[0])
>>>>>>> +            # if more than 10 patches this should save 00/xx, not 0/xx
>>>>>>
>>>>>> s/save/say/
>>>>>>
>>>>>> (my typo, I think)
>>>>>
>>>>> ;-)  I'll fix this.
>>>>>>
>>>>>>
>>>>>>> +            zero_repeat = 1
>>>>>>> +            while (count / (10 ** zero_repeat) > 0):
>>>>>>> +                zero_repeat = zero_repeat + 1
>>>>>>
>>>>>> How about:
>>>>>>
>>>>>>       zero_repeat = int(math.log10(count)) + 1
>>>>>>
>>>>>> ?
>>>>>
>>>>> yes, it's better. just need to import the match lib.
>>>>> I will change to this and sent v2 patch. Thanks.
>>>>>
>>>>> BTW: speak of patman, I get an issue of using the "Series-prefix".
>>>>>
>>>>> When I use Series-prefix like following in the commit:
>>>>>       Series-prefix: U-Boot][
>>>>> Then I get the patman generated patch like:
>>>>>       [U-Boot][ PATCH]
>>>>>                       ^  a space here.
>>>>>
>>>>> A space is before the 'PATCH', that annoys me. But I don't see you have
>>>>> such
>>>>> space in your patches. Any advice to avoid the extra space?
>>>>> Thanks in advance.
>>>>
>>>> This is intentional, since if you use a prefix of 'RFC' we want to get
>>>> 'RFC PATCH v2' instead of 'RFCPATCH v2'. See GetPatchPrefix().
>>>
>>> yes, understood.
>>>
>>>> Why do
>>>> you want [U-Boot] anyway? That sounds more like the project than a
>>>> patch prefix. Perhaps you could add an option to prepend the project
>>>> in square brackets?
>>>
>>> I tried a the project prefix, and that works for the format-patch command.
>>> But it not work for patman.
>>>
>>> here is my steps:
>>>    git config format.subjectprefix "U-Boot"
>>>
>>> Now, when I run git format-patch, the generated patch will have "[U-Boot]"
>>> prefix. But if I run "patman -c1 -n", the [U-Boot] prefix is gone.
>>> It seems patman overide the format.subjectprefix  option of git. Do you have
>>> an idea about what is the difference between the run "git format-patch -1"
>>> and "patman -c1"?
>>> thanks.
>>
>> Why do you want [U-Boot] anyway?
>
> I always send the u-boot patch with title like "[U-Boot][PATCH]xxxxx".
> Is "[U-Boot]" string not needed?

Nope.

>
>>
>> I suggest you add an option to prepend the patman project name (in
>> square brackets) to patches, from the -p option.
>
> hmm, it's better to get the prefix by 'git config format.subjectprefix', then add this to CreatePatches().

You could certainly do that.

>
>>
>> Patman has to set up its own prefix when generating patches. See
>> CreatePatches().
>
> okay, I'll try this, if it works nice then I will send a patch out.
>
> Thanks a lot.

OK sounds good.

Regards,
Simon