Re: glib 1.1.12, gtk1.1.12 for debian/ppc
On Sat, 16 Jan 1999, Sven LUTHER wrote:
> I don't understand you, ... first you complain that the gtk package is too old,
> then you say it is because of a known egcs buug, and there is no solution apart
> from no compiling the test program that cause problem, and then you said it is
> not the place of the debian package to deal with it. So why do you complain
> then that the gtk package is too old ? Best would be to inform the list of the
> problem, and complain to the egcs folk like you said.
I feel that I have been misunderstood. I stated -rather than complained-
that because the existing gtk was a little old -I had no problem with
that, but some programs did- I took the liberty of creating my own, which
I placed for public access, and offered it to be include in the official
tree. I also said that the solution was to manually compile the offending
files -that is different from not compiling them at all, after all they
are test files- and then proceed with the building process. As you
correctly state, the best thing was to inform the list, which I believe I
have already done, otherwise I could keep the final package for myself.
And yes, an egcs bug should not be reported to the debian bug list, and I
don't think I am being paranoid about this. Maybe someone else feels
different about this?
And in any case, complaints -which mine was not- are good when they lead
in constructive criticism. But I have the impression that you accuse me
of having great expectations, whilst underestimating the work that is
needed for all this. I really apologize if I have lead you to such
conclusions. My only intention was to help.
> I have tried (and succeeded) in building the glib1.1.12 packages, but i had to
> manually remove the noninst_PROGRAMM line in the MAkefile.[am,in¸..] not very
> easy.
although I presume this works ok, I am being a little sceptical about
manually altering the Makefiles, esp. when they actually work for everyone
including ppc, albeit minding the compiler problem.
> Also perhaps you know why since 1.1.9 or earlier, gtk complains about undefined
> target deNONE when trying to compile the po files ? I guess i have a
> misconfigurated locale stuff or something like that. But then my 2.2.0-pre4
> kernel died on me yesterday night in middle compile, without even any kind of
> debug message and messed all my partitions, i will have to reinstall stuff
> again.
no can't say I do, I never noticed anything like that.
> I don't agree, a little patch that only compiles testglib and friends if we are
> not on powerpc would be nice, Anyway, they don't get into the debian package,
> so there is no need to compile them ...
I strongly disagree. A wrongly built library should show that it is so,
by way of the test files. If they don't build ok, or don't run ok, because
of a library error -NOT a compiler error- then we should know that, and
save ourselves from making an incorrect library from distributing. There
is no need for them to be included, but there is certainly a need to
compile them. That's why they are included in the first place.
> no offense taken, it is just that you have to be aware of the volunteer status
> of every debian developper, and that there are only 24hours a day ...
As I said about this, I am completely in accord with that, and that's why
I offered my help in the first place. I imagine that you agree with me
about ending this seemingly never ending conversation, and proceeding with
the important stuff, such as actually producing.
Konstantinos Margaritis
kmargar@cc.uoa.gr
Reply to: