On Fri, Oct 4, 2019 at 8:26 PM Przemek Klosowski via devel
wrote:
>
> On 10/3/19 12:19 PM, Matthew Miller wrote:
> > On Thu, Oct 03, 2019 at 11:13:32AM -0500, Michael Cronenworth wrote:
> >>> Remote changelog URLs might become inaccessible over time, making
> >>> tracking down
> >>> behavior chan
On 10/3/19 12:19 PM, Matthew Miller wrote:
On Thu, Oct 03, 2019 at 11:13:32AM -0500, Michael Cronenworth wrote:
Remote changelog URLs might become inaccessible over time, making tracking down
behavior changes & tricky bugs problematic.
Yes, there are systems that do not have Internet access.
Ex
On Fri, Oct 04, 2019 at 08:52:45AM -0400, Neal Gompa wrote:
> I would personally make the same request as well for anything I've got
> a changelog entry for.
This is reasonable and I think anything which moves changelog entries out of
spec files (and ideally also RPM metadata) must preserve this.
On Fri, Oct 4, 2019 at 4:39 AM Daniel P. Berrangé wrote:
>
> On Thu, Oct 03, 2019 at 01:12:10PM -0500, Jason L Tibbitts III wrote:
> > > "MM" == Matthew Miller writes:
> >
> > MM> Whether or not it's documented policy (and I can't remember or find
> > MM> anything either), many packages have
On Thu, Oct 03, 2019 at 01:12:10PM -0500, Jason L Tibbitts III wrote:
> > "MM" == Matthew Miller writes:
>
> MM> Whether or not it's documented policy (and I can't remember or find
> MM> anything either), many packages have the practice of trimming very
> MM> old entries.
>
> You can't alway
* Daniel P. Berrangé [03/10/2019 14:47] :
>
> FWIW, my approach is to purge all changelog entries older than 2 years
> the first time I touch a package in January each year. Is there any value
> in having guidelines to encourage some policy in this area, so maintainer
> approach it in a consistent
> "MM" == Matthew Miller writes:
MM> Whether or not it's documented policy (and I can't remember or find
MM> anything either), many packages have the practice of trimming very
MM> old entries.
You can't always do this. I tried to purge changelog entries from a
package older than 2010 and wa
On Thu, Oct 03, 2019 at 11:13:32AM -0500, Michael Cronenworth wrote:
> >Remote changelog URLs might become inaccessible over time, making tracking
> >down
> >behavior changes & tricky bugs problematic.
> Yes, there are systems that do not have Internet access.
> Examples:
> - Classified systems wi
On 10/3/19 9:45 AM, Martin Kolman wrote:
Also, the current changelogs are self contained & do not require internet
access.
Remote changelog URLs might become inaccessible over time, making tracking down
behavior changes & tricky bugs problematic.
Yes, there are systems that do not have Intern
On Thu, Oct 03, 2019 at 10:19:40AM -0500, Chris Adams wrote:
> The other thing that makes source code changelogs less useful in some
> cases is that they are often very verbose, with info that isn't clear to
> end users. They show changes that are often not relevant (like maybe
> between release 1
Once upon a time, Daniel P. Berrangé said:
> I think having a record of upstream SCM would be useful regardless. Many
> times when submitting patches to Fedora packages, I've been told to send
> my patch to upstream insteadwhich means trying to figure out where
> that upstream is for this give
On Thu, Oct 03, 2019 at 09:37:31AM -0500, Chris Adams wrote:
> Once upon a time, Daniel P. Berrangé said:
> > Or just add some RPM metadata tags to record the upstream SCM type + URL +
> > branch / release tag, etc. The user can thus easily find the upstream
> > full commit logs corresponding to t
On Thu, 2019-10-03 at 09:37 -0500, Chris Adams wrote:
> Once upon a time, Daniel P. Berrangé said:
> > Or just add some RPM metadata tags to record the upstream SCM type + URL +
> > branch / release tag, etc. The user can thus easily find the upstream
> > full commit logs corresponding to the paca
On Thu, Oct 03, 2019 at 09:16:15AM -0500, Chris Adams wrote:
> > I think rather than this, we should bite the bullet and remove changelogs
> > entirely from spec files.
> I find "rpm -q --changelog" useful (at least when maintainers put useful
> info there, which isn't always), so please don't.
I
On 10/3/19 10:29 AM, Daniel P. Berrangé wrote:
On Thu, Oct 03, 2019 at 10:21:37AM -0400, David Cantrell wrote:
On 10/3/19 10:16 AM, Chris Adams wrote:
Once upon a time, Matthew Miller said:
I think rather than this, we should bite the bullet and remove changelogs
entirely from spec files.
I
Once upon a time, Daniel P. Berrangé said:
> Or just add some RPM metadata tags to record the upstream SCM type + URL +
> branch / release tag, etc. The user can thus easily find the upstream
> full commit logs corresponding to the pacakge.
IMHO that is only good when the Fedora package is nothin
On Thu, Oct 03, 2019 at 10:21:37AM -0400, David Cantrell wrote:
> On 10/3/19 10:16 AM, Chris Adams wrote:
> > Once upon a time, Matthew Miller said:
> > > I think rather than this, we should bite the bullet and remove changelogs
> > > entirely from spec files.
> >
> > I find "rpm -q --changelog"
On 03. 10. 19 16:16, Chris Adams wrote:
Once upon a time, Matthew Miller said:
I think rather than this, we should bite the bullet and remove changelogs
entirely from spec files.
I find "rpm -q --changelog" useful (at least when maintainers put useful
info there, which isn't always), so pleas
On 10/3/19 10:16 AM, Chris Adams wrote:
Once upon a time, Matthew Miller said:
I think rather than this, we should bite the bullet and remove changelogs
entirely from spec files.
I find "rpm -q --changelog" useful (at least when maintainers put useful
info there, which isn't always), so pleas
Once upon a time, Matthew Miller said:
> I think rather than this, we should bite the bullet and remove changelogs
> entirely from spec files.
I find "rpm -q --changelog" useful (at least when maintainers put useful
info there, which isn't always), so please don't.
--
Chris Adams
On Thu, Oct 03, 2019 at 09:51:01AM -0400, Matthew Miller wrote:
> On Thu, Oct 03, 2019 at 02:47:27PM +0100, Daniel P. Berrangé wrote:
> > FWIW, my approach is to purge all changelog entries older than 2 years
> > the first time I touch a package in January each year. Is there any value
> > in havin
On Thu, Oct 03, 2019 at 02:47:27PM +0100, Daniel P. Berrangé wrote:
> FWIW, my approach is to purge all changelog entries older than 2 years
> the first time I touch a package in January each year. Is there any value
> in having guidelines to encourage some policy in this area, so maintainer
> appr
On Thu, Oct 03, 2019 at 09:23:16AM -0400, Josh Boyer wrote:
> On Thu, Oct 3, 2019 at 9:21 AM Vitaly Zaitsev via devel
> wrote:
> >
> > Hello all.
> >
> > Is it possible to remove old %changelog entries from SPECs? I can't find
> > information about this in Fedora packaging guidelines.
>
> Yes.
F
On Thu, Oct 03, 2019 at 03:40:43PM +0200, Miroslav Suchý wrote:
> Rpmbuild actually removes all entries older than %_changelog_trimtime
> In Fedora this macro is defined as
> -13: _changelog_trimtime%{lua:print(os.time() - 2 * 365 * 86400)}
> I.e. everything older than 2 years is discarded
On Thu, Oct 03, 2019 at 02:45:20PM +0200, Vitaly Zaitsev via devel wrote:
> Hello all.
>
> Is it possible to remove old %changelog entries from SPECs? I can't find
> information about this in Fedora packaging guidelines.
>
> All history still can be found in git log.
We already set an rpm variab
Dne 03. 10. 19 v 14:45 Vitaly Zaitsev via devel napsal(a):
Hello all.
Is it possible to remove old %changelog entries from SPECs? I can't find
information about this in Fedora packaging guidelines.
All history still can be found in git log.
Rpmbuild actually removes all entries older than %_
On Thu, Oct 03, 2019 at 02:45:20PM +0200, Vitaly Zaitsev via devel wrote:
> Is it possible to remove old %changelog entries from SPECs? I can't find
> information about this in Fedora packaging guidelines.
> All history still can be found in git log.
Whether or not it's documented policy (and I ca
On Thu, Oct 3, 2019 at 9:21 AM Vitaly Zaitsev via devel
wrote:
>
> Hello all.
>
> Is it possible to remove old %changelog entries from SPECs? I can't find
> information about this in Fedora packaging guidelines.
Yes.
josh
>
> All history still can be found in git log.
>
> --
> Sincerely,
> Vi
Hello all.
Is it possible to remove old %changelog entries from SPECs? I can't find
information about this in Fedora packaging guidelines.
All history still can be found in git log.
--
Sincerely,
Vitaly Zaitsev (vit...@easycoding.org)
___
devel mail
29 matches
Mail list logo
