[U-Boot] sending updated patch versions

[U-Boot] sending updated patch versions

[Holger Brunck]

Hi Wolfgang,
I have a short question about sending updated patch versions. I have read:

http://www.denx.de/wiki/view/U-Boot/Patches#Sending_updated_patch_versions

and I wonder if it is needed to update the history of *all* patches in a serie
even if for different specific patches nothing was changed. Or if it sufficient
if only the subject indicates the version number for such patches.

E.g patch serie where only 1 patch has changes requested, do I need to update
all other patches with the history:

"Changes for v2: - nothing"

In the past I did this, but I ask myself if this is really needed.

Please let me know what the rules are in such a case. I think the wiki page did
not describe what to do with unchanged but version updated patches.

Best regards
Holger

[U-Boot] sending updated patch versions

[Wolfgang Denk]

Dear Holger,

In message <4E119D36.7040900@keymile.com> you wrote:
>
> and I wonder if it is needed to update the history of *all* patches in a serie
> even if for different specific patches nothing was changed. Or if it sufficient
> if only the subject indicates the version number for such patches.

If you keep the threading of the messages intact (i. e. supply proper
In-reply-to: and References: headers - like you get automatically when
inserting the message ID when "git send-email" asks for the
"Message-ID to be used as In-Reply-To for the first email"), then
there is usually no need to resend unchanged patces at all - what for?
It would be just a waste of effort and bandwidth.

> E.g patch serie where only 1 patch has changes requested, do I need to update
> all other patches with the history:
> 
> "Changes for v2: - nothing"

No. Please don't repost the unchanged patches at all. Repost just the
one that you modified.


There are only few cases where you have biggger or more complicated
series of patches, which undergo a number of changes here and there
where it may make sense to repost a clean version of the whole series.

The basic rule here is once again: "No news is good news."

Best regards,

Wolfgang Denk

-- 
DENX Software Engineering GmbH,     MD: Wolfgang Denk & Detlev Zundel
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: (+49)-8142-66989-10 Fax: (+49)-8142-66989-80 Email: wd at denx.de
"What is wanted is not the will to believe, but the will to find out,
which is the exact opposite." - Bertrand Russell, _Sceptical_Essays_,
1928

[U-Boot] sending updated patch versions

[Holger Brunck]

Hi,

On 07/04/2011 05:05 PM, Wolfgang Denk wrote:

> In message <4E119D36.7040900@keymile.com> you wrote:
>>
>> and I wonder if it is needed to update the history of *all* patches in a serie
>> even if for different specific patches nothing was changed. Or if it sufficient
>> if only the subject indicates the version number for such patches.
> 
> If you keep the threading of the messages intact (i. e. supply proper
> In-reply-to: and References: headers - like you get automatically when
> inserting the message ID when "git send-email" asks for the
> "Message-ID to be used as In-Reply-To for the first email"), then
> there is usually no need to resend unchanged patces at all - what for?
> It would be just a waste of effort and bandwidth.
>

yes I agree and thats why I am asking. Ah ok, yes I have seen the message from
git send-email but wasn't sure how to use this.

So if I get a request to update 4/5 from a already posted patch serie which
message id I have to enter for "In-Reply-To for the first email"? I assume I
have to use the message id from the initial 4/5 Patch Email which has hit the
mailing list or?

>> E.g patch serie where only 1 patch has changes requested, do I need to update
>> all other patches with the history:
>>
>> "Changes for v2: - nothing"
> 
> No. Please don't repost the unchanged patches at all. Repost just the
> one that you modified.
> 

yes this sounds reasonable and is also less work for the sender.

Best regards
Holger Brunck

[U-Boot] sending updated patch versions

[Wolfgang Denk]

Dear Holger,

In message <4E11DB54.9000000@keymile.com> you wrote:
> 
> So if I get a request to update 4/5 from a already posted patch serie which
> message id I have to enter for "In-Reply-To for the first email"? I assume I
> have to use the message id from the initial 4/5 Patch Email which has hit the
> mailing list or?

Right.

If in doubt, you can see the message ID for example in patchwork.

Best regards,

Wolfgang Denk

-- 
DENX Software Engineering GmbH,     MD: Wolfgang Denk & Detlev Zundel
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: (+49)-8142-66989-10 Fax: (+49)-8142-66989-80 Email: wd at denx.de
Every program has at least one bug and can be shortened by  at  least
one  instruction  --  from  which,  by induction, one can deduce that
every program can be reduced to one instruction which doesn't work.

[U-Boot] sending updated patch versions

[Scott Wood]

On Mon, 4 Jul 2011 17:05:28 +0200
Wolfgang Denk <wd@denx.de> wrote:

> Dear Holger,
> 
> In message <4E119D36.7040900@keymile.com> you wrote:
> >
> > and I wonder if it is needed to update the history of *all* patches in a serie
> > even if for different specific patches nothing was changed. Or if it sufficient
> > if only the subject indicates the version number for such patches.
> 
> If you keep the threading of the messages intact (i. e. supply proper
> In-reply-to: and References: headers - like you get automatically when
> inserting the message ID when "git send-email" asks for the
> "Message-ID to be used as In-Reply-To for the first email"), then
> there is usually no need to resend unchanged patces at all - what for?
> It would be just a waste of effort and bandwidth.

IMHO, it makes it easier for the maintainer to collect the full patchset for
applying.  There's also possibility that there is another patch in the
series that got updated, but the message got lost, deleted,
forgotten, filtered to a different folder (especially if the patchset is
crossposted to multiple mailing lists), etc.

It's also more effort for the sender to prune out such patches than to
leave them in, and more effort for people reading the patchset to track
down the missing pieces (even with threading -- the old messages may have
been deleted, especially if it's been a while).

The extra bandwidth is negligible.

> > E.g patch serie where only 1 patch has changes requested, do I need to update
> > all other patches with the history:
> > 
> > "Changes for v2: - nothing"
> 
> No. Please don't repost the unchanged patches at all. Repost just the
> one that you modified.
> 
> 
> There are only few cases where you have biggger or more complicated
> series of patches, which undergo a number of changes here and there
> where it may make sense to repost a clean version of the whole series.

Of course, if it's one or two patches out of 15, updated half an hour after
the previous version went out, or if there's repeated iteration on a subset
of the patchset (especially if the patchset is a loosely-related
collection), then it's reasonable to just post what changed.  I think the
cases where a full repost helps are not that rare, though.

-Scott