Bug#747412: marked as pending
Osamu Aoki
osamu at debian.org
Sat Oct 17 10:12:55 UTC 2015
control: retitle -1 add chk-origtgz script calling uscan/diffoscope/tardiff/cmp.
Hi,
On Fri, Oct 16, 2015 at 10:37:54PM +0800, Paul Wise wrote:
> On Fri, 2015-10-16 at 22:08 +0900, Osamu Aoki wrote:
>
> > You requested
> > uscan: option to verify current upstream tarball
> >
> > Doesn't the following look like what you requested
>
> No, that doesn't look like the new plan I proposed here:
>
> https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=747412#30
I am aware ... but that is the follow up comment/discussion.
I closed this because I addressed the original issue. But if you insist
to change this bug to be wider scope one, let's keep this. I will
remove this bug from the proposed changelog.
> The new plan I proposed doesn't need any changes in uscan itself, just
> a new chk-origtgz script that calls uscan and diffoscope/tardiff/cmp.
If this bug is "doesn't need any changes in uscan itself", the title
must be changed accordingly as above. (This is not wishlist on uscan.)
I am a bit lost hear. There were many ideas discussed.
* A new tool (chkorigsig?) -- in https://bugs.debian.org/747412#25
* "a new chk-origtgz script" -- in https://bugs.debian.org/747412#30
My thought on uscan direction is more like chkorigsig. I also think
uscan is overloaded. So it is good idea to split out functionalities as
much. But for this round of changes, I am fixing structural problem
first. My change enables to turned on and off each step of uscan.
(Avoid conditional short cut "return".) That is done now. I may work
on chkorigsig if there is a clean solution for that. This does not
involve doing diff etc. If successful, uscan will use chkorigsig.
As for the idea of chk-origtgz, the new uscan and uupdate combination
record hush of the original tarball before repacked to orig.tar.gz.
This may be used by the chk-origtgz you are thinking...
Good luck.
Osamu
PS: This discussion gave me a chance to rethink --report option
handling.
More information about the devscripts-devel
mailing list