Nicholas Nethercote fe3cc3bdf9 Change hp2ps to output .ps files in the same dir that the .hp input files
are in.  This fixes bug #117686.



git-svn-id: svn://svn.valgrind.org/valgrind/trunk@5287
2005-12-05 20:45:59 +00:00

103 lines
3.6 KiB
Plaintext

=============================================================================
Valgrind Roadmap
=============================================================================
This file serves as a rough roadmap for Valgrind development. It shows a
minimal set of features we hope to implement for each version. It's in
reverse chronological order.
-----------------------------------------------------------------------------
3.2.0
-----------------------------------------------------------------------------
Scheduled for end-Mar 06 (3.1.0 + 4 months) ?
In order of increasing speculativeness
--------------------------------------
* Add ppc64-linux support.
* Fold in the V bit compression stuff if it works well (early signs
are promising) and get rid of Addrcheck.
* Get function wrapping working again. Reinstate basic thread checks.
Reinstate Helgrind.
* Performance tuning:
- faster register allocation in vex
- improve stack-update pass
- assess effect of branch misprediction in dispatchers
* Try to accelerate development for Darwin ?
Smaller things
--------------
* Consider using the following defaults:
--leak-check=yes
--num-callers=20
* Expose some of m_redir's functionality to tools so that Memcheck
can replace strlen/strcmp on PPC32 (remove the 3.1.0 hack for this
which checked in m_redir.c if the current tool was Memcheck).
-----------------------------------------------------------------------------
3.1.1
-----------------------------------------------------------------------------
Tentatively end-Dec 05 (3.1.0 + 1 month): fix any critical bugs in
3.1.0.
-----------------------------------------------------------------------------
3.1.0
-----------------------------------------------------------------------------
Scheduled for around November 2005. (Released 27 Nov 05).
Definite
--------
* Get 32-bit and 64-bit programs working smoothly on AMD64 (Tom?). Several
levels of smoothness here, we should aim for at least level 3.
1. Be able to build a 32-bit valgrind on a 64-bit machine, so you can
build and install both, and manually choose between bin/valgrind and
bin64/valgrind.
2. Build both automatically when installing.
3. Choose the appropriate executable automatically at startup just from
"valgrind".
4. With --trace-children=yes, allow 32-bit programs to exec 64-bit
programs and vice versa, and invoke the appropriate Valgrind
automatically.
[All four levels done by Tom]
* Get PPC32 working usably with Memcheck (Julian). Has already improved a
lot since. Get Cachegrind working with it (Nick).
[Both done by Julian]
* Rewrite address space manager; statically link the core with
each tool; remove all glibc dependencies (Julian).
[Done by Julian]
* What about --time-stamp=yes?
[Fixed by Julian to give relative time since startup]
* Make it work with GCC 2.95 (bug #111781) -- don't put declarations after
statements in blocks. Do it after merging ASPACEM with the trunk.
-Wdeclaration-after-statement is the GCC warning that detects this, but
it is only present in GCC after 3.4.0 (ie. not in 3.0.X--3.3.X)...
[Done by Tom and others]
* We need to reintroduce some kind of core/tool interface versioning,
so that if external tools link with libcoregrind.a incompatibilities
are detected.
Maybe
-----
* Get pthread modelling and Helgrind working again. Requires function
wrapping (Nick).
[Won't happen for 3.1.0. Function wrapping is difficult.]
* Reinstate Addrcheck and/or implement V-bit compression in Memcheck (?).
[Won't happen for 3.1.0.]
* Allow suppressions by filename + line number? (Joseph Link's patch)