Zsh Mailing List Archive
Messages sorted by:
RE: zsh-4.0.1 -- some minor build/install problems
- X-seq: zsh-workers 14958
- From: "Andrej Borsenkow" <Andrej.Borsenkow@xxxxxxxxxxxxxx>
- To: "Peter Breitenlohner" <peb@xxxxxxxxxxxx>, <zsh-workers@xxxxxxxxxx>
- Subject: RE: zsh-4.0.1 -- some minor build/install problems
- Date: Mon, 18 Jun 2001 13:24:47 +0400
- Importance: Normal
- In-reply-to: <Pine.LNX.4.21.0106111210040.1321-200000@xxxxxxxxxxxxxxxxxxxxx>
- Mailing-list: contact zsh-workers-help@xxxxxxxxxx; run by ezmlm
> 1. zsh-4.0.1/Test/Makefile.in:
> without the patch "make DESTDIR=some_where check" rather badly fails.
> Why would I want to use "DESTDIR=some_where" together with "make check"?
> Seen in isolation this admittedly doesn't make much sense, nevertheless
> "make check" shouldn't fail. Consider, however,
> "make DESTDIR=some_where all check install"
That makes sense I guess.
> 2. zsh-4.0.1/Makefile.in:
> without the patch I had to say
> "make all; make -C Doc info" as user
> and "make install install.info" as root
> instead of the more natural "make all info".
> You might say that "make install.info" will build the info pages in any
> case. True, but when the build tree is on an nfs-mounted
> filesystem, root may
> not have write permission.
The actual bug is that currently info is not created at all when building
off CVS. install.info depends on texi not on info as it should. Your patch
does not fix it. I'll commit the right one.
There remains the issue of make install.info vs. make info of course.
> 3. zsh-4.0.1/Doc/Makefile.in:
> without the patch "make DESTDIR=some_where install.info" produces
> some weird
> messages (in my case zsh.info from zsh-3.0.5 had no 'INFO-DIR-SECTION', or
> there may not be any /usr/local/info/zsh.info at all).
Now, when I'm looking at it - DESTDIR was instroduced for package creation -
in which case install-info is pretty useless and DESTDIR is plain wrong.
Alternative would be to skip install-info if -n $DESTDIR.
Messages sorted by: