Re: Is libtool being maintained at all?

[Dalibor Topic]

Salut Nicholas, hello Bob,

Nicholas Wourms wrote:
> Charles Wilson wrote:

> > I'd guess not.  It appears that debian's changes are all on the 1.4.3 
> > branch, which is permanently and irrevocably dead.  We really really 
> > really don't want to provide anybody with more excuses to stay on 
> > 1.4.3.  If John Q. Developer wants new-and-improved libtool 
> > functionality, then we want JQD to use 1.5.x.
>
>
> Try telling that to the gcc people...  They are still using some early 
> prerelease of 1.4a for crying out loud!

As one of kaffe's current build system developers I find the current 
situation very frustrating.

We have changed over to libtool 1.5 shortly after it was released in 
order to get better support for our cross compiling developers and 
users. That works just fine, according to our users, and I'm very 
pleased about the decision to switch over to libtool 1.5. I'm also very 
grateful for the hard work that has gone into improving libtool for 1.5. 
Kaffe would not be as flexible as it is now without using libtool. 
That's the good part. Thank you all for creating libtool!

The somewhat bad part is that we have discovered several problems with 
libtool 1.5, ranging from hour-long configure script checks to real bugs 
leading to crashes. We are getting regular bug reports for libtool 1.5, 
and fortunately, they are often accompanied by patches. ;)

My humble attempts at getting the attention of libtool developers to our 
small and simple patches have so far failed to receive *any* response 
from libtool maintainers for more than a month now, despite repeated 
postings to both address@hidden and address@hidden I find 
that situation quite frustrating.

On one hand, I hear people say: abandon 1.4 and use 1.5, on the other 
hand, when I use 1.5, I'm apparently on my own. If you don't want to 
provide excuses for people to use 1.4, please try to work together with 
people who are using 1.5 to resolve issues that come up. I would like to 
contribute to libtool, but I have the feeling that libtool maintainers 
are not interested in my contributions.  In fact, I don't have the 
feeling that the libtool maintainers are interested in anyone's 
contribtions beside a small circle of initiated. It's not clear to me 
how to get into that circle. ;)

The ugly part is that ignoring potential contributors may not buy you 
time in the long run. You may end up redoing the work already performed 
by contributors you ignore now. I don't believe that's the outcome 
anyone here would want.

I need a working libtool for kaffe, which can (at least in theory ;) run 
on more than 50 platforms. I believe that everyone would be better off, 
if the libtool package maintainers on those platforms would/could work 
more closely with the libtool developers to keep improving the GNU 
libtool source base, instead of further increasing the gap with their 
local forks. Working local forks, frozen at older libtool versions, 
don't make the switch to a broken libtool 1.5 attractive. In order to 
have more people adopt libtool 1.5, I belive that this deadlock has to 
be broken up by everyone cooperating to make the next libtool release 
even better.

I've tried to incite some closer cooperation by posting links to 
distribution specific libtool patches. Some of the package maintainers 
have even offered their support in getting the distribution specific 
patches in. To me, that's good sign, and I hope that these patches make 
it in some day. Just like I hope that kaffe's patches will receive a 
response some day, if I only keep reposting them often enough. ;)

If that's not possible, maybe a 'most general unified' fork of libtool 
under a different umbrella could be a temporary, egcs-like solution to 
what I percieve to be a problem in the making.

cheers,
dalibor topic