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

Re: Bug#331114: debian-el: leaves temporary files in /tmp after installation



>>>>>  PSG == Peter S Galbraith [2005-10-2]

PSG> Romain Francoise <rfrancoise@debian.org> wrote:
>> Peter S Galbraith <p.galbraith@globetrotter.net> writes:
>> 
>> > Any comments from the Debian-Emacsen crowd?
>> 
>> Some packages (like muse-el) use a different approach and store the
>> compilation log in the flavor-dependent site-lisp directory, which
>> solves this problem.

PSG> I think it makes it worse.  Instead of having log files stored in /tmp
PSG> where they can esaily be deleted, they end up within the core
PSG> filysystem.

I use both approaches with auctex:

  - site-lisp/auctex/CompilationLog.gz contains the output of byte
    compiling the package; it is usually very useful for me when I
    receive a bug report (I have a bug script in the package which
    attach it to bug reports compiled with reportbug for this purpose,
    otherwise I ask the bug submitters to send it to me), because it
    also contains the output of configure, make, and so on, so that I
    can easily understand why (if) the installation has failed;

  - a temporary file in /tmp from update-auctex-elisp (according to the
    user debconf preferences), containing the (usually) very useless
    output of byte compilation of AUCTeX style files (a package specific
    collection of very simple elisp forms); this file gets created on
    package installation by the background running update-auctex-elisp
    and once a week by a cron script (see the auctex package for more);
    if I were to delete it afterwards, it would have no sense to create
    it first of all: that would be more or less equivalent to redirect
    the output of update-auctex-elisp to /dev/null, which is not an
    option I would like to embrace.

I use the first approach with mailcrypt also.

PSG> I used to do this, but decided to move away from it some time ago
PSG> because of the clutter.  I'm not going back.

Don't you think that it can be useful to have a log of the installation
specific to the host of bug submitters?

-- 
Ciao, Davide



Reply to: