Re: Releasing ncurses
On 2010-02-18 16:59 +0100, Sven Joachim wrote:
> On 2010-02-18 11:27 +0100, Craig Small wrote:
>> I'd like #556378 fixed beforehand, if possible. This is the -static
>> problem, which Sven tried to fix with the --without-dlsym and had
>> build troubles. If I can get some idea on how tricky it is to fix we
>> can then work out wether or not we fix it this time.
This has been fixed in yesterday's patch, but it may be tricky to
backport because we would need to regenerate configure, and the Debian
build system does not currently have rules for that. Besides, Thomas
uses a rather strange autoconf version (2.52.20081225), and I have no
idea what might break if we use the Debian autoconf version.
> It seems to me that we need a shlibs bump as there are two new
> functions, is_pad() and is_subwin(), see the 20090906 NEWS entry.
> Which means that we could disrupt whatever transition is going on right
> now. Sorry for not noticing this earlier.
In the light of that and given that additional symbols have been
introduced since then, I would like to switch to the current upstream
patchlevel now, as there might be not another chance to do this before
the squeeze freeze. It would also make it easier to fix some other
issues like #562134.
This would of course require a few days testing of the resulting
packages. I have already verified that building with my proposed patch
for #556378 works fine.
> Reading the 20100123 NEWS entry I'm also not sure if #542031 is really
> fixed in master yet.
The following Fedora bug suggests that it's not, so this would be
another argument for switching to a newer patchlevel: