Custom Query (242 matches)

In audfiles.scr of UNIAUD package there is a line to copy MMPM2.INI. This should not be done since it just walks over the file that is present there. So what if you have Open MPEG from IBM installed or multiple CD players. You break the MMOMS2 system that way.

The uniaud16 repository contains build products that should not be version controlled. For example

j:\sla_dev2\uniaud16.dev\trunk\install
>svn ls -v
     33 vladest          1249 Dec 07  2005 AUDHELP.HLP
     33 vladest         12301 Dec 07  2005 CARDINFO.DLL
     33 vladest           958 Dec 07  2005 CARDINFO.RES
     33 vladest          7328 Dec 07  2005 CARDINFO.zip
     33 vladest           326 Dec 07  2005 audfiles.scr
     33 vladest           876 Dec 07  2005 audhelp.itl
     33 vladest          4395 Dec 07  2005 audplay.ico
     33 vladest           130 Dec 07  2005 cardinfo.def
     33 vladest         10050 Dec 07  2005 cardinfo.rc
     33 vladest          1965 Dec 07  2005 control.scr
     33 vladest          3625 Dec 07  2005 makefile.os2
     33 vladest          4395 Dec 07  2005 midiplay.ico
     33 vladest         88041 Dec 07  2005 rc.exe
     33 vladest         48128 Dec 07  2005 rcpp.exe
     33 vladest          2273 Dec 07  2005 rcstub.c
     33 vladest          4395 Dec 07  2005 vidplay.ico

In the above cardinfo.dll, cardlinfo.zip and others are generated by the build process. These should be deleted from the respository with svn rm.

In the makefile.os2's of the drivers, isa, and pci directories under alsa-kernel, there are driver source files listed that have been commented out. Edits have been made at some point to those source files, so a port was at least begun. AFAICT there does not appear to be support for these in the other places necessary to make them work even if they were not commented out of the makefiles.

An attempt to compile pci/ice1712 after the merge to 0.9.0rc4 failed with errors.

The problems with this scenario are:

1. Hardware that could be supported by Uniaud is not.

2. Uniaud is now tailored to work only with a few driver files with significant modifications. Tailoring it instead to work with a greater number of ALSA drivers with fewer modifications would reduce the work required to add support for new hardware, and reduce the chances for bugs peculiar to each hardware configuration.

3. The present AlsaResync is merging these source files without any testing, which leaves room for error and reduces or nullifies the benefit of the work.

My proposed action at this point is to test whether these files compile successfully before the resync, and whether that has been broken by any stage of the resync, fixing as needed. Even though this does not in itself enable support for the hardware, it preserves that potential, reduces merge error, and also provides more clues about what in Uniaud outside of alsa-kernel might need done to improve general compatibility with ALSA.