> are you talking about dmidecode.c only or kernel things too? dmidecode.c only. Two more things make me think we need to merge all changes with care. * If I read the list carefully, the lm_sensors version is based on the files found at http://www.domsch.com/linux/patches/dmi/. I suppose that the base is dmidecode-1.1.c with one of the three patches applied. Unfortunately, none of these patches leads to the "official" version available from Alan's files (but -take3 is the closest, of course). * Arjan van de Ven (Red Hat) fixed a bug in December 2001, which, it seems, never made it back to Alan. Arjan's version is available at http://people.redhat.com/arjanv/. I suggest we apply my patch first (and call the result version 1.3) since I already sent it to Alan and he seems to agree. Then, I'll submit a new patch with both Arjan's fix and lm_sensors improvements and we have version 1.4, we everything so far included. We'd consider Alan's files repository the official place for the file, and I'd ask Matt Domsch and Arjan to either delete or update their files. After some testing is done (comparing old and new results on a number of systems should be sufficient), we correct the last bugs[1], bump version to 1.5 and commit all changes to lm_sensors. This is the plan. (Then and only then, I'll add a call to dmidecode in sensors-detect to make it ibm-bogus-hardware-resistant.) If nobody objects, I'll start working ASAP. Else, feel free to express yourself. [1] Yeah, I know, there is no such thing as "the last bug". Think of it as "the last bug we could find after a reasonable number of tests". -- Jean "Khali" Delvare http://www.ensicaen.ismra.fr/~delvare/
