Chris Gerhard <Chris.Gerhard at Sun.COM> writes: [redirecting somewhat]
> While preparing an RTI I ran hg pbchk which in turn calls hg comchk > and it always reports an ARC case as missing. > > : eacces.eu FSS 219 $; hg comchk > remote: Not trusting file /share/bld/ONclones/onnv-clone-eu/.hg/hgrc > from untrusted user blduser, group keepers > Comments check: > These bugs/ARC cases were not found in the databases: > PSARC 2009/171 > : eacces.eu FSS 220 $; Something smashed the CSV data that these commands use. The script on os.org is doing the correct thing, the data is absent. (assuming I'm looking at the right data, and I assume I am, because this change is obviously recent) This even breaks the SCM test suite, which queries for 2006/051 in one of its test, this case is no longer in the CSV, at all. (clearly, it was for the longest time). The CSV should contain *every* case, open or otherwise. It seems it no longer does. I have no information as to why this is the case, but clearly it will break comchk until it was rectified in whatever fashion. It does -appear- that the CSV is now eliding closed cases. This was never previously the case, and when we discussed this interface with (at the time) John Plocher, were told this was by intent, and it was reasonable to rely on. Alan, did you change anything here? SCM people, did you? If this was changed intentionally to elide "unpublished" cases (I haven't investigated, beyond looking 2006/051), clearly that is internal SMI business, and I thus leave it up to people employed there to rectify, though obviously they must do so with alacrity. I will note, however, that all information contained in the CSV *does* display on the (new) open caselog, so it seems doubtful it would be hiding things for that reason? -- Rich
