> > As already said, I have never used libcloud and maybe should hence not > care that much, but I still want you to notice that it will probably be > removed from Debian if the bug is not fixed: > http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=598463#30
If we're going to get kicked from repos for this, I think it's time to drop the "dependencies are generically bad" thing and fix this issue. It doesn't seem like we would even need to explicitly *require* M2Crypto (though, IMO, we might as well since it's on pypi). On Fri, Nov 12, 2010 at 5:05 PM, Pietro Battiston <[email protected]>wrote: > Il giorno lun, 08/11/2010 alle 19.40 +0100, Pietro Battiston ha scritto: > > Il giorno lun, 08/11/2010 alle 11.18 +0100, Tomaž Muraus ha scritto: > > > Hello, > > > > > > A user has already created an issue on Jira about this[1] some time ago > and > > > because the root issue is in the Python module only a warning has been > added > > > to the README. > > > > > > I still personally think that the better solution would to fix the > problem > > > and subclass the HTTPSConnection class and manually check the hostname > or > > > switch to the M2Crypto library like you have suggested. > > > > > > Only problem with switching to the M2Crypto library is that it adds an > extra > > > dependency. > > > > > > Sure. Hence, using M2Crypto if available and printing a warning > > otherwise is to my eyes the optimum. > > > > That's what I'm doing in the attached patch (from "svn diff" on svn > > trunk). > > > > Would you mind reviewing/testing it? Would you/some other developer > > suggest if there are other places (a quick grep found none) in libcloud > > where https connections are made? > > > > thanks a lot > > > > Pietro > > > As already said, I have never used libcloud and maybe should hence not > care that much, but I still want you to notice that it will probably be > removed from Debian if the bug is not fixed: > http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=598463#30 > > bye > > Pietro > > > > > > > > > > [1]: https://issues.apache.org/jira/browse/LIBCLOUD-55 > > > > > > On Mon, Nov 8, 2010 at 11:00 AM, Pietro Battiston < > [email protected]>wrote: > > > > > > > Hello, > > > > > > > > I' coping with bug > > > > http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=598463 > > > > and it seems the only upstream reference to it is > > > > > > > > > http://mail-archives.apache.org/mod_mbox/incubator-libcloud/201009.mbox/%3c5860913.463891285776633273.javamail.j...@thor%3e > > > > > > > > Now, there is no doubt that it's indeed an annoying thing, and that > many > > > > other > > > > projects just fixed it (waiting for python devs doing it). But I'm > not > > > > in search of a flame: I just would like to fix it (as a Debian patch, > if > > > > you are not interested). > > > > > > > > In a project of mine, the analogous fix took very few lines of code: > > > > > > > > > > > > > http://code.google.com/p/galleryremote/source/diff?spec=svn6&r=6&format=side&path=/trunk/galleryremote/gallery.py > > > > > > > > and I would be happy to try to do the same on libcloud, though I > > > > perfectly know it will be slightly harder. > > > > > > > > But the main point is: I never used this library, neither have an > > > > account on any cloud provider, so I would totally appreciate if some > dev > > > > or at least user could cooperate with me. Feel free to answer in > mailing > > > > list of contact me privately. > > > > > > > > Thanks > > > > > > > > Pietro Battiston > > > > > > > > > > > > >
