issues

开放源码的编译器open watcom 1.6.0版的源代码

Project plan for the browser
----------------------------

Well, the browser is moving along quite well, but there are still a few
major issues left to solve.

**** Big issues that _have_ to be done

- have to resolve global symbols and types
- references to symbols is unimplemented
- change the .mbr file format to ELF
- port to other OS's (waiting on wclass port to LUI)
- use "infinite" list boxes (waiting on wclass port to LUI/DUI)
- graphic tree options (waiting on wclass port to LUI)

**** Would be nice (but not absolutely necessary)

- some kind of hardcopy reports			(possibly higher priority?)
- integrated with message translation stuff	(possibly higher priority?)
- support for templates, exceptions and labels. (possibly higher priority?)
- fortran support. 				(possibly higher priority?)
- dwarf macro stuff is unimplemented (waiting on compiler support)
- pinned windows (& resultant communication??).	
- references to a macro were handled

there are also a fair number of little "cleanup" items, but these are the
biggest issues.
Changing the file format to ELF gets rid of the rather primitive and
proprietary "format" we had before, and will let us use the linker to "link" the
browse information.  This "linking" would involve reading in the browse
information and interpreting it, scanning for and removing redundant
information.  This will greatly reduce browser memory requirements and increase
browser speed, and when we move to DWARF as a debugging information format
will give the same benefits to the debugger as an added bonus.

After these features are added and the porting is done, then there probably
should be some time for the browser to be used internally so I can get some
user feedback on what features need to be modified or added.

So... my development plan for the browser is this:
1) finish off the big issues
2) do cleanup then "would be nice" things until wclass is ported to LUI
3) do the LUI dependant things
4) do an "internal release" to get it tested & get feedback.

Note that there probably won't be much time in step 2 to do "would be nice"
things, as the cleanup and big issues are still a fair amount of work, and
since I will probably spend a fair amount of my time working on the linker
a lot of it will have to be done by my coop.  (This assumes that the port
of wclass to LUI does not encounter a major delay from the current plan)

---- jim.