Bug#781517: [Reproducible-builds] Bug#781517: tracker.d.o: please don't link ftbfs issues on reproducible.d.n

2015-04-29 Thread Holger Levsen
Hi,

On Mittwoch, 29. April 2015, Jérémy Bobbio wrote:
 As dpkg-buildpackage now outputs `-Wdate-time` instead of
 `-Werror=date-time` in the reproducible toolchain, I believe the above
 to be less an issue.

true, but I still think we shouldn't mark known false ftbfs as ftbfs... but:

we know how to exclude these false results (see the ftbfs pages on rb.d.n) so 
we should only flag those real ftbfs in the json output. which is just a 
matter of doing it...

I probably can look into this on Friday but wont if someone else beats me to 
it.


cheers,
Holger




signature.asc
Description: This is a digitally signed message part.


Bug#781517: [Reproducible-builds] Bug#781517: tracker.d.o: please don't link ftbfs issues on reproducible.d.n

2015-04-29 Thread Jérémy Bobbio
Holger Levsen:
 On Montag, 30. März 2015, Paul Wise wrote:
  These seem like FTBFS that should be reported, so the package
  maintainers patch out usage of the macros, especially as the plan was
  to enable warnings for them by default eventually.
 
 yes, they should be reported. thats why they are listed on 
 https://reproducible.debian.net/issues/unstable/timestamps_from_cpp_macros_issue.html
 
 There is *no* need to list them on tracker.d.o, in fact, I consider this 
 *harmful*: once you start showing false negatives, people are less likely to 
 believe the results, thus people start to mentally ignore them.
 
 Please dont make this happen.
  
  How can we show FTBFS that do need to be reported but not show FTBFS
  that don't need to be reported?
 
 Please only show reliable results from reproducible.d.n on tracker.d.o - 
 ftbfs 
 bugs are not reliable. Please exclude them. 
 
 (What's needed to exclude them would be to include logic from 
 reproducible.debian.net which I dont consider sensible.)

As dpkg-buildpackage now outputs `-Wdate-time` instead of
`-Werror=date-time` in the reproducible toolchain, I believe the above
to be less an issue.

-- 
Lunar.''`. 
lu...@debian.org: :Ⓐ  :  # apt-get install anarchism
`. `'` 
  `-   


signature.asc
Description: Digital signature


Bug#781517: [Reproducible-builds] Bug#781517: tracker.d.o: please don't link ftbfs issues on reproducible.d.n

2015-03-30 Thread Holger Levsen
Hi Lunar,

On Montag, 30. März 2015, Jérémy Bobbio wrote:
 Sources embedding timestamps should be fixed. Getting help from
 maintainers would be great! So I think they should be notified when
 their packages are using __TIME__ (et al.) CPP macros.

yes. we should file bugs about those issues. 

But linking all ftbfs issues lists too many false negatives, so I really think 
it's the opposite of helpful to link them.


cheers,
Holger




signature.asc
Description: This is a digitally signed message part.


Bug#781517: [Reproducible-builds] Bug#781517: tracker.d.o: please don't link ftbfs issues on reproducible.d.n

2015-03-30 Thread Jérémy Bobbio
Holger Levsen:
 while you thankfully added links to unreproducible packages on 
 reproducible.debian.net you've also added links to packages which fail to 
 build from source there, as can for example be seen on 
 https://tracker.debian.org/pkg/ocaml-faad
 
 Please don't do that, ftbfs issues on reproducible.d.n can have several 
 causes 
 not steeming from the package (eg half of the 700 ftbfs errors in unstable 
 come from 
 https://reproducible.debian.net/issues/unstable/timestamps_from_cpp_macros_issue.html

Sources embedding timestamps should be fixed. Getting help from
maintainers would be great! So I think they should be notified when
their packages are using __TIME__ (et al.) CPP macros.

-- 
Lunar.''`. 
lu...@debian.org: :Ⓐ  :  # apt-get install anarchism
`. `'` 
  `-   


signature.asc
Description: Digital signature