1
0
mirror of https://github.com/moparisthebest/pacman synced 2024-11-10 03:25:01 -05:00
pacman/TODO.aaron
Aaron Griffin 170d63190a * Shuffled some of the alpm_list free funtions - still not perfect, but better
* Added alpm_list_remove_node for single list node removal
* Proper error checking/output for failed db_read/db_write (missing files)
* Invalid packages (missing files) are now removed from the package cache
* -Qs and -Ss output now look the same
* config.rpath causes errors on one machine I had, so I added it to CVS
* Fixed a "clobbered memory" issue when installing groups - only the outer list
  should be free'd, not the contained data
2007-01-24 08:51:50 +00:00

71 lines
2.9 KiB
Plaintext

== This is my custom TODO file ==
* Make sure all the alpm_list_t changes are valid (especially alpm_list_free)
* src/pacman:
- There's quite a few single function headers which contain the pacman_*
functions. We should move these to a single header (pacman.h) to clean up
the source a bit.
- config_t duplicates much of what can be accessed from the alpm_option_*
interface. Determine what can / can't be removed and do so.
- initialize alpm BEFORE parsing args, so we can get rid of a handful of
local lists and things in pacman.c (see below)
- fix the "--debug" parameter. requiring a bitmask from the user is silly.
we can parse this can send a real bitmask to alpm.
* lib/libalpm:
- Remove `root` param from alpm_initialize. We can simply initialize to /
and call set_root at a later time. This allows us to initiliaze earlier.
- move logging (alpm_logaction) out of the library. log files should be
written by the app, not the library. Adding a PM_LOG_* constant that
frontends can recognize and send to a file is probably a good idea.
- maybe it's my fault, but I see little difference between PM_LOG_FLOW1,
PM_LOG_FLOW2, and PM_LOG_DEBUG. We might want to just do away with flow1
and flow2 and use debug across the board.
- clear up list allocation/deallocation - some lists need to be free'd, some
do not and there is no clear indication WHEN this should happen.
- remove DB entries (directories) on a read error?
* pacman: A LOT of functions are way too long. There should be an upper limit of
100-200 lines. _alpm_add_commit is around 600 lines, and is far too complex.
Quite a lot of this needs to be refactored as we go.
Function list (from Dan McGee):
add.c:
_alpm_add_commit
be_files.c:
_alpm_db_read
_alpm_db_write
conflict.c:
_alpm_checkconflicts
_alpm_db_find_conflicts
package.c:
_alpm_pkg_load (getting a bit long)
remove.c:
_alpm_remove_commit
server.c:
_alpm_downloadfiles_forreal
sync.c:
_alpm_sync_sysupgrade (getting there)
_alpm_sync_prepare
_alpm_sync_commit
* pacman: fixup doxygen documentation for public interface
* pacman: use gcc visibility (static) for non-public functions (all _alpm
functions should be static)
**** BUGS ****
* Removal of conflicts on -A and -U (INCOMPLETE)
* ordering of installed packages based on dependencies
i.e. pkga depends=(pkgb pkgc), pkgb depends=(pkgc)
pacman -S pkga pkgb pkgc
> a depends on b and c, move them: pkgb pkgc pkga
> b depends on c, move it: pkgc pkgb pkga
> pacman -S pkgc pkgb pkga
* gensync: get all info from packages instead of PKGBUILDs
for now, we only need to check the PKGBUILD in case of a
'force' option
* pacman: add pactest cases/functionality for all -Q operations and make sure
the output is sane