Hurd task/todo list

[Marco Gerards]

Some of the tasks and todo items should be removed or at least changed
because they have been done. Can someone please correct me if I'm
wrong and otherwise fix it (or tell me to send in a patch to fix it)?

I hope this mail makes sense (it is a little chaotic), please have a
look at the task and todo list while reading the rest of my mail.

Thanks,
Marco

>From the tasks list:

* Make thread death in C threads really free resources.

This item should be removed because it is not relevant
anymore. Perhaps It should be replaced by:

!!!
* Replace cthreads by pthreads.

I've added "!!!" here because I think it is high priority, am I
mistaken about this?



!!!
  * An RPC trace program to aid debugging.  Look at utils/rpctrace.c
  for starters and ask roland for more info.

RPC trace works fine for me. Also the RPC ID -> names translation
works perfectly for me. Perhaps this task should be replaced by:

* Create and install the RPC list. Use "mig -list" to do this.



* A fancy terminal driver that uses readline and supports
  detach/attach, color, and virtual consoles.

Marcus wrote this and it works. There are more specific todo items on
savannah. This should be removed, perhaps a short description of every
todo item on savannah should be added to the todo list.




* Handy shell programs to send msgport msgs, and change default init
  ports and ints.

This (/bin/msgport) works perfectly for me. This task should be
removed. Perhaps a todo item should be added if some functionality is
still missing.



>From the todo list:

* Write coding standards suggestions for Hurd

Isn't this a pre-GCS todo item? If it is it should be removed, if not
can someone please explain what this is about?



*** Implement file_notice_changes.

Marcus did this AFAIK. If it is not completely done this should be
updated to match the libraries/servers that don't support this
interface.



For libdiskfs:

*** Use off_t correctly. !!

Does this have something to do with the 64-bits changes? Has this been
done?




For ext2fs:
*** If the target of a symlink is the empty string, stat seems to spin forever 
!!!

IIRC Moritz reported this (correct me if I'm wrong). And I think this
has been fixed, does someone know more about this?