[pacman-dev] [arch-dev-public] [signoff] filesystem 2007.11-5

[Xavier]

Aaron Griffin wrote:
> Changes in -5:
>     Removal of /tmp workaround - FS#7194 close message

7197 is not fixed unfortunately. Note that you had to be careful when 
testing it, exactly because of the workaround in 7194.
That is, you install fileystem package, then see the /tmp permissions 
are fine, but the permissions were in fact changed by filesystem scriptlet.

See http://bugs.archlinux.org/task/7484#comment17619 where I explain why 
/tmp/ is a special case for pacman.

>     Removal of pacman dirs - FS#6062 comments

That's probably a good thing, pacman needed /var/lib/pacman to install 
filesystem in the first place anyway.
On this topic, shouldn't pacman create this directory itself if it 
doesn't exist?

Note this is exactly the same issue as the recent cachedir change :
http://projects.archlinux.org/git/?p=pacman.git;a=commit;h=1a0aaa20df29234250ecc0f9fc3f56bb8cf22ca1

with the following correspondence :
alpm_option_add_cachedir <-> alpm_option_set_dbpath
_alpm_filecache_setup <-> *_alpm_db_register_local

But I am not sure what to think about the following comment 
(libalpm/db.c, ~560):
       snprintf(path, PATH_MAX, "%slocal", dbpath);
       /* TODO this is rediculous, we try to do this even if we can't */
       if(stat(path, &buf) != 0 || !S_ISDIR(buf.st_mode)) {
               _alpm_log(PM_LOG_DEBUG, "database dir '%s' does not
                         exist, creating it\n", path);
               if(_alpm_makepath(path) != 0) {
                       RET_ERR(PM_ERR_SYSTEM, NULL);
               }
       }