[Buildroot] [PATCH 02/11] manual: minor fix in patch-policy.txt

Samuel Martin s.martin49 at gmail.com
Tue Feb 26 11:15:55 UTC 2013

Hi Thomas,

2013/2/26 Thomas Petazzoni <thomas.petazzoni at free-electrons.com>:
> Dear Samuel Martin,
> On Tue, 26 Feb 2013 10:44:39 +0100, Samuel Martin wrote:
>> >> +- If a patch does some package version fixes, this should be documented in the
>> >> +commit message of the patch itself (or the message prepended to the 'diff'
>> >> +itself).
>> >
>> > I am not sure to understand this part.
>> I don't know how to explain this clearly.
>> My point is, for some packages, BR does not provide the latest version,
>> but some fix may be available upstream, so they have been backported in BR
>> (e.g.: package/libglib2/libglib2-make-codegen-python2-python3-compliant.patch).
>> Any suggestion to explain this point in a better way is welcome.
> I don't understand why you would want to put this here in the
> documentation. What you're saying is just a detail on what the patch
> could possibly contain (i.e a backport from upstream versus a
> Buildroot-specific hack for cross-compilation), I don't see what it
> changes in terms of patch policy.
Fair enough, I'll drop this point.

>> >> +* Otherwise, patch files matching +<packagename>-<description>.patch+
>> >> +  are applied following the +ls+ command order.
>> >
>> > Shouldn't we be enforcing a <packagename>-<number>-<description>.patch
>> > policy, in order to ensure that patches are applied in the right order?
>> Well, here I stick to the current apply-patches.sh implementation.
> The current apply-patches.sh implementation takes the patches in
> alphabetical order, which is why it is important to have the patches
> numbered to make sure that the patches are applied in a well-known
> order, and not an order based on the <description> field alphabetical
> sorting.

Well, explicit numbering is better than implicit, so should be the doc.
I'll fix it and re-spin this patch.



More information about the buildroot mailing list