[Top][All Lists]

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

Re: Automake bug #516 -- and more

From: Peter Breitenlohner
Subject: Re: Automake bug #516 -- and more
Date: Wed, 23 Jan 2008 10:28:28 +0100 (CET)

On Tue, 22 Jan 2008, Ralf Wildenhues wrote:

Hello Ralf,

I did take a look at history now, and think that your patch is ok, and
the order reversing just an oversight.  The patch is missing a ChangeLog
entry and a test.  Providing them would have saved me time processing
the bug.

Thanks a lot.

I am terribly disappointed that I never got any reaction and that this bug
is still present in the new release.

And you've made that quite clear by spamming it to five lists instead of
to bug-automake as noted.  I've limited that to automake lists now.

Sorry, that was caused by simply replying to the original mail.

My frustration was mostly due to the fact that I never got any reaction. I
think I could have provided a ChangeLog entry and a test, if asked for it.


I recently stumbled (once again) at two other manpage related automake
problems, where I'd like to try for solutions provided there is a realistic
chance that such solutions would be included in a future automake release.

The problems are these:

(1) In some cases it is completely inadequate to apply the program name
transform to manpage names. E.g., in the man-db package there is manpath.1
describing the manpath command (should be transformed), and manpath.5
describing the format of the manpath configuration file (should certainly
not be transformed).

Maybe that could be taken care of by something like notransform_MAN_MANS
(needs a better name).

Moreover, renaming "manpath"=>"xmanpath" and  "manpath.1"=>"xmanpath.1"
without changing the manpage contents is sort of strange.

(2) There are packages with translated manpages -- good. But automake
provides no way to supress their installation or to install them for a
subset of the available languages -- not good.

At the moment all available translated manpages are installed
unconditionally (under the assumption that the packaging system of a
distribution takes care of this?). I think something analogous to
--disable-nls and/or LINGUAS=xxx for translated manpages is very desirable.


What is your opinion?

Peter Breitenlohner <address@hidden>

reply via email to

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