Re: [gentoo-dev] Last rites: app-text/cuneiform

[Rich Freeman <rich0@gentoo.org>]

On Sun, Mar 24, 2013 at 6:43 AM, Markos Chandras <hwoarang@gentoo.org> wrote:
> So per https://bugs.gentoo.org/show_bug.cgi?id=462366#c4, the package
> now has a new maintainer so it will not be removed.
> See? This is what I call a good solution instead of going around and
> constantly saying "Ohhh bad bad Gentoo removes awesome packages"

Probably worth noting that the real problem for packages like these is
a barely-existent upstream.  If a package is really important to you
it behooves you to try to support upstream however you can.

I've yet to really dig into the issues with this package, but some of
the Gentoo bugs refer to working implementations in other distros or
upstream filed patches that apparently aren't in the repository yet.
These are really upstream issues.

So, while cuneiform has a lease on life in Gentoo, it is really just a
matter of time before some big dependency change kills it for good if
upstream doesn't pick up momentum.  I don't mind maintaining an odd
patch or two, but there is no way something like this is going to stay
in the tree if it ends up becoming a blocker for some big toolchain
upgrade (unless the fix is trivial).

So, if you find this package really useful consider this whole thread
as a warning.  I don't personally use it, but I think that this
package isn't quite at the point of no return and at least some appear
to be passionate about it so I'm willing to buy them some time.  If it
does reach that point, then I'll put out a call for maintainers (proxy
or otherwise) and put it down myself if there is no response to save
treecleaners the duplicate effort.  If you aren't interested in
developing then offer donations to upstream, or do something to
revitalize the project.  It isn't a lost cause - YET.

On a side note, if you use this instead of tesseract I'd be interested
in hearing about why (off list).  In my very limited tests tesseract
seems to perform better.  The cuneiform community (what little there
is) would do well to understand their niche and exploit it, or
influence healthier projects to address their needs.

Markos - I'm not sure what can be done to try to better flush out user
interest in taking care of packages that are on the verge of death.
I'd suggest announcing pending removals before masking them, but I
suspect that more often than not the only reason we get replies on
-dev is that users notice the masks.  Maybe the package masks could
have a webpage explaining how users can help rescue packages
constructively, and include a link to it in mask notices.  Since I've
tended to be an advocate for not masking as quickly I might go ahead
and toss something together.

Rich