[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Bug#724539: RFS: psocksxx/0.0.1-1 [ITP]



Hey Tobi,

Thanks for the review.

On Wed, Sep 25, 2013 at 6:22 PM, Tobias Frost <tobi@frost.de> wrote:
> Hallo Uditha,
>
> (I'm not a DD/DM... Also quite new to reviewing)
>
> Thanks for your work, however there is already a iostream based socket
> library in Debian: libskstream. So maybe you can elaborate a little why
> you think this library is a must in Debian :) (like reverse
> dependencies...)
>

I have an ITP for bitz-server and the latest version depends on this
library, hence the packaging :).

> The package itself looks quite good, just a few remarks (I did not build
> the package, just a paper-review)
>
> d/control you libpsocksxx0 Provides: libpsocksxx. I think that is wrong,
> as the so-name will be important when installing the library later. (But
> I could be wrong here)
>
> I think it is depreciated to install *.la files.
>
> You should enable multiarch-support

Duly changed and uploaded to mentors.

>
> Maybe you'd like also to provide a -doc package with the generated
> doxygen documentation.

Definitely though about it but the upstream version does not install
doxygen generated docs and I didn't want to include a patch for
debian. I tried searching but couldn't find an example of how this can
be done in a clean way. Any pointers?

>
> Please be aware: Buildd's do not have necessarily have networking, not
> even lo is guaranteed. So you might want to disable (parts of) your
> testsuite which relies on networking

I noticed this when I was building using sbuild and a couple of tests
failed. If these aren't going to have any impact on the builds I'm
tempted to leave them there rather than completely disabling them.
Partly disabling would need a debian patch and I personally don't like
that idea (unless it is really needed of course).

>
> Best regards,
> coldtobi
>

Thanks,
Udi


Reply to: