bug-ncurses
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: ncurses-5.4 configure bug - ticdir not set properly


From: Thomas Dickey
Subject: Re: ncurses-5.4 configure bug - ticdir not set properly
Date: Fri, 13 Feb 2004 18:50:46 -0500 (EST)

On Fri, 13 Feb 2004, Paul Lew wrote:

> >>>>> "Thomas" == Thomas Dickey <address@hidden> writes:
>
>     Paul>> Problem:
>     Paul>>
>     Paul>> When make install use a prefix different than the prefix
>     Paul>> specified in the configure.  The ticdir is not updated
>     Paul>> accordingly.  The result is the 'make install' wiped out an
>     Paul>> old installation.
>     Paul>>
>     Paul>> Example:
>     Paul>>
>     Paul>> > ./configure --prefix=kit1
>     Paul>> > make
>     Paul>> > make install prefix=kit2
>
>     Thomas> But this would work:
>
>     Thomas> make install DESTDIR=kit2
>
> That's good, how about obsolete the use of prefix=... in make install?
> Why it partially worked?

It only works if the name=value is passed to the subordinate makes
explicitly.  Depending on the particular make program - and unfortunately
in the case of _versions_ of GNU make, the assignments from the command
line may/may not propagate into subordinate makes without an assignment
in the top-level makefile (and note that ncurses/Makefile calls
include/Makefile).  The makefiles are written to assume as little as
possible about the particular make program.

You can look at it in more than one way:  DESTDIR is the conventional way
to shift a whole package to a new location on install.  Arguably any of
the (counting now) 17 other variables _could_ be modified at the
command-line.  Anywhere in between raises the issue of why not _this_ one,
too.  For instance, you might want to (I don't):

        make install \
                ticdir=/usr/local/lib/terminfo \
                bindir=/usr/local/ncurses/bin

>
>     Thomas> It helps - on any package - to use "make -n" to see what
>     Thomas> it will do.
>
> I did that but hard to expect what might be wrong, especially when
> output is long.
>
>     Paul>> ...
>     Paul>>
>     Paul>> datadir=kit2/share \
>     Paul>> ticdir=kit1/share/terminfo \       <== ooch!
>     Paul>> ...
>     Paul>>
>     Paul>> This happens to 5.3 too.
>
>     Thomas> was there an earlier bug report?
>
> I dont know, I did not file this problem last time.

I can recall the general topic some time ago, but it takes time to locate
that sort of thing in email...


-- 
Thomas E. Dickey
http://invisible-island.net
ftp://invisible-island.net




reply via email to

[Prev in Thread] Current Thread [Next in Thread]