LSB conf call notes for 2008-06-25

George, Alexey, Kay Jeff: there's a new task status table at the top of ProjectPlan40. If anyone has information to add or sees anything missing, please add it or let me know. Robert: maybe talk a little more about __ symbols. Could we ignore the __ things in appchk, instead of reporting error, but still not add to LSB? Jeff: what happens if someone drops the interface? Compatibility guarantee. Robert: as a developer, what do I do with __ errors? Jeff: fix some in LSB 4, better docs for those we don't include. program_invocation_name. Jeff: objections? Mats: something to do, not sure what. Jeff: should we lean towards inclusion? Mats: yes. getgrent_r and friends. Mats: the ones without "f" are done. "f" ones mean "work on FILE *". Not really controversial, just not in POSIX. Jeff: probably not, except the usual "we have to write the man pages". Robert: found some man pages online for the f* functions. Mats: will think about it. May be an issue with specifying some things that we don't already. argz_*. Jeff: functions to work with Pascal-style strings. Seems useful. No apps seem to use them, though. Alexey: argz_append has apps: ImageMagick and gphoto2. Mats: thinks he's build an ImageMagick in LSB mode. Probably caught by configure scripts for open-source apps. Robert: maybe leave out until we hear a request. Jeff: not necessarily reject, just downgrade the priority. [Correction, found after the call: argz_* don't do Pascal strings, but rather C string arrays, with null bytes separating each string, along with a length.] sysinfo, get_nprocs*. Mats: two issues. The get_nprocs stuff can be gotten another way; perhaps just put a pointer to the other way in the database. Jeff: little usage in navigator, note already present. Robert: sysinfo looks like stuff Ted is nervous about exposing. Jeff: good number of apps that use it, sounds like useful and fairly well-understood info, maybe we should add it. Robert: Ted would be the best person to ask. Agrees that the info is useful. Will it change again? Jeff: no idea, but the change made before looks scalable. mempcpy, wmempcpy. Jeff: __ version is in. Robert: did we miss something? Jeff: 8 apps in navigator, looks like a convenience function, can do pointer arithmetic to do the same thing. Kay: back to "how many code changes do we request of people?" Jeff: maybe we want to figure out why the __ version is in. Ron: there was a bug for __mempcpy, will look for it. Jeff: on vacation for next two calls. Mats will lead unless someone (Ted?) has a burning desire to.