Recent blog entries for ericp

Wrote a few patches and bug fixes... Started cleaning up the multimedia DLLs for a proper separation:

• started by splitting the driver code to be as close as possible of Windows' implementation ; I now got native MCI CDROM run with native WINMM and MMSYSTEM DLLs pair - not very useful, but good news anyway in terms of interoperability
• fixed a couple of remaining bugs in lots of MM areas (MCI, CD ROM...)
• Rewrote some of Marcus patches and fixed some other errors he reported (mainly on MCI)
• uncommitted yet: WINMM/MMSYSTEM are mostly separated, I still ahve to solve the service thread dependency to be finished

Also started working on delay loading for builtin DLLs. This is now functional, but still needs to be ironed out.

Vacations soon... good news, with maybe more time to hack !!

Well, hard to keep up with a regular diary writing.

I've been buzy looking after nice issues.

First one was a bug in GCC (2.95.1 and 2.95.2 are plagued...). Seems to be an optimization bug, but which disappears when -no-strength-reduce (but it's different from the old strength-reduce GCC 2.7.x bug).

Hacked also a new stabs parser for WineDbg, which shall solve some issues with latest GCC versions (see previous diary entries).

I had also a bit of an activity with the new WineHQ look (revamping the WWN, and setting up the right connections). Since, past weeks had quite an heavy traffic on the wine- devel mailing list, the WWN team and myself have been late on publishing... It's maybe time to reorganize the WWN edition mechanisms... We'll see.

Kept on improving the PCM converter in MSACM DLL. Rate conversion is now in place, and working (but still not as fast as it should be ; maybe one day, I'll rewrite it using fixed point math instead of floating point).

Advanced a bit the debugging stuff (fixed anyway a couple of bugs), but not sure if my lazy symbol loading stuff is a reasonable feature... still requires some thinking. Got a bug report from latest gcc which breaks Wine's debugger stabs parsing. Will require a major rewrite.

I've been trying to fix a strange bug in Media Player (probably memory corruption, but doesn't appear when relay debugging is turned on...). Sigh, sounds like hunting ain't over yet.

Shall be shortly on vacation. No Wine in sight... hopefully.

Found some time to do some more Wine hacking. Added some PCM conversion routines to MSACM ; far from being finished, but still better than nothing.

Fixed a regression in exception handling code (debugger was frozen).

Still haven't finished long standing patches in my tree (lazy symbol loading in the debugger to decrease startup time ; final MMIO fixes - François Jacques fixed some of them, but a few remain, sigh). Also discussed a lot with François Jacques who is running into interesting issues in multimedia code (mainly the recording part, which has been poorly tested... more patches to come it seems).

Ove resurfaced after the Linux Kongress, and since the mailing issues seem to be over, the "Keeping Track of Wine section" in Wine Weekly News should be available again.

Less and less time to hack Wine... sigh this means more spent on other subjects... who said girl(s) ?

Found anyway some interesting bugs (loader, DIB drawing...); fixes are underway.

Also installed on Winehq Dimi's tool to track cvs commits as a whole... nice tool, still needs some more polish

Uwe also asked that winedbg (wine debugger) can also compile as a pure Win32 app... this means more hacking on the way.

Back from vacation (yurk)

Back to Wine hacking (lots of multimedia issues to sort out)

Back to work (re-yurk)

Yesterday I finished some Wine loader fixes... Windows Media player is now (at least, sigh) able to stream audio content (asf/asx) files. But (yurk) there are still some winsock issues to iron out... Not so much time to spend on Wine lately, but vacation will be here soon (yeah!!)

Did some more testing on some new wine debugger features. May be I'll submit that tomorrow, when I have time.