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

Wow.. it almost works



Okay folks, the new version is pretty much working now. The only thing it
doesn't do is actually install things (ie -d is permanently fixed to
'on')

So.. cvs update, compile it up and make sure it can do the proper things
for you and let me know if it works.

Do this:

cvs co apt
cd apt
make startup
make
cd build/bin
export LD_LIBRARY_PATH=`pwd`
./apt-get <blah>

Now we need a bunch o things before we can get .debs of this new
experimental version out.
  - Packaging. How many packages should we make? I'm queasy about a
    libapt-pkg package because the soname will change frequently which
    means the package name will change frequently and I really don't
    expect terribly many people to use it. Interest in a -doc package
    has been brought up and interest in a libapt-pkg-dev package too
  - Docs. The man pages are out of date and someone needs to write a big
    huge man page for the configuration file. (Branden??) 
  - The ftp method is gone again

This is version 0.3.x incidently.

It also manages to fix these bugs, we maybe competing with X here for
sear number of closures :>

#25026: apt: Why do you list the packages you're _not_ doing anything to
instead of the ones you are?
#22507: apt: apt-get - listing packages to be upgraded (wishlist)
 Summary: Feature request
 Status: Fixed in 0.3.0 via the -u option
#21113: [Apt] In dselect, update dies and exits if the packages file is
not found.
 Status: Fixed in 0.3.0
#22675: APT does not honor the --admin-dir option
#22836: Disk/Zip-Method for apt?
 Summary: APT does not provide a way to download packages onto a
          removable media for another computer
 Status: 0.3.0 has substantially better support for this to the point
         that it is doable by using a seperate configuration file and
         the -c option
#27601: srange errors from dselect
 Summary: Couldn't locate an archive source
 Status: Require status file
         Believed to be fixed in 0.1.9, was not reproducable w/ given
         status file
#27841: apt: apt depends on a missing library
 Status: New versions of APT in slink have been compiled with libstdc++2.9
#23984: apt: support for "no_proxy" would be nice
 Status: Planed to be integrated into the new methods via the configuration
         file
         Done - Use Acquire::http::proxy::host.com="DIRECT"
#25021: apt: Need some control over multiple connections
 Status: Probable that 0.3.x will have support for configuing some
         parameters
         Done - Use Acquire::Queue-Mode="access"
#25019: apt: Confusing progress report
 Summary: Gripes about the progress meter
 Status: I do not intend to implement very many of these ideas in apt-get.
         The GUI will naturally be better
         I think the new progress meter address basically everyone's
         concerns.
#25022: apt: Lack of feedback on date checking
 Summary: Wants to know what package files were not updated
 Status: There is no place for this in the current apt-get design,
         probably won't make the GUI either.
         Wee, the new acquire code allows this, it now prints out 'Hit'
#26019: apt may report wrong line speed
#26433: apt: claims to fetch things it doesn't actually fetch (wishlist)
#28778: apt: apt's fetched message is wrong for update of packages files
 Summary: APT includes the fetch time from the local cache in its
          calculations
 Status: Probably will be fixed with new acquire code
         And it actually was fixed with the new acquire code
#26670: apt: apt-get dumps core after checking integrity
 Summary: Some terminal environments cause APT to crash
          Win95 telnet and emacs term-in-a-buffer are two at least
 Status: I have no idea why.
         Seems to be gone in .3, whatever it was...
#25001: apt: cleaned out archive even though not all files were installed
 Status: Apparently in some cases APT can return a success code even
         though it failed. I'm paying very close attention to this in
         0.3.x. Not to mention that the clean behavior will be
         configurable..   


Reply to: