Rafael Sadowski <raf...@sizeofvoid.org> writes:

> On Sun Oct 09, 2016 at 09:42:22AM +0200, Jeremie Courreges-Anglas wrote:
>> Rafael Sadowski <raf...@sizeofvoid.org> writes:
>> 
>> > On Sat Oct 08, 2016 at 09:07:47PM +0200, Dmitrij D. Czarkoff wrote:
>> >> Rafael Sadowski <raf...@sizeofvoid.org> wrote:
>> >> 
>> >> >>> -HOMEPAGE=  http://tcpreplay.appneta.com/   
>> >> >>> +HOMEPAGE=  http://tcpreplay.appneta.com/
>> >> >> 
>> >> >> Does removing whitespace warrant REVISION bump?
>> >> >
>> >> >It is a tab directly after the HOMEPAGE.
>> >> 
>> >> I think it does not really call for REVISION bump.  Yes, it is a change,
>> >> but it is not really necessary to force update.  Or am I getting
>> >> REVISION wrong again?
>> >
>> > I'm not so familiar with the internal package process and how make/perl 
>> > handle
>> > $HOMEPAGE but sh(1) catch the tab:
>> >
>> > url=`make show=HOMEPAGE`
>> 
>> Here it's not a matter of "is this change really worth it to force
>> update?".  The issue at hand is that if you repackage after fixing the
>> homepage, the checksum of +DESC changes, and register-plist will
>> rightfully complain.  So a bump is needed.
>> 
>> (If you don't see this behavior because you set an empty
>> PLIST_REPOSITORY, well, don't do that.)
>> 
>
> Thanks jca@ for your illustration. Last night I checked the both .tgz
> packages. With and without tab after the URL and surprise, surprise the
> difference between the two packages are 7 Bytes. So a REVISION bump is
> necessary, isn't it?

You can't assume that if the resulting package size changes in
successive builds, then the package needs a revision bump.  Many other
reasons may change the resulting package size.

As mentioned in the manpage, register-plist(1) exits by default if the
HOMEPAGE changes.  That's on purpose.  And since any register-plist
error should be avoided, yes, you need a REVISION bump.

-- 
jca | PGP : 0x1524E7EE / 5135 92C1 AD36 5293 2BDF  DDCC 0DFA 74AE 1524 E7EE

Reply via email to