Date: Tue, 19 Mar 2002 21:58:48 +0900 (JST) From: Hiroki Sato <hrs@eos.ocn.ne.jp> To: n@nectar.cc Cc: murray@FreeBSD.ORG, freebsd-doc@FreeBSD.ORG Subject: Re: Security advisories and XML Message-ID: <20020319.215848.45176388.hrs@eos.ocn.ne.jp> In-Reply-To: <20020317225416.GA36506@hellblazer.nectar.cc> References: <20020316131007.GX17499@freebsdmall.com> <20020317.043511.115904137.hrs@eos.ocn.ne.jp> <20020317225416.GA36506@hellblazer.nectar.cc>
index | next in thread | previous in thread | raw e-mail
[-- Attachment #1 --] "Jacques A. Vidrine" <n@nectar.cc> wrote in <20020317225416.GA36506@hellblazer.nectar.cc>: n> This looks good, but it misses the mark a bit in that it is very n> document-oriented. Something that is more structure-oriented (but n> still document friendly!) is needed. e.g., component names and n> versions must be clearly identified (for searching and subscription n> processing), advisory revisions must be dealt with, and the revisions n> must be logically grouped by branch and issue, so that they may be n> used to generate a patchfile (this is possibly the hardest part, and n> perhaps too much for a first cut). Another one is attached. In this version, fixed files and revisions of each component can be grouped in <module>, and this includes two of XSLT stylesheet which output HTML and sh-script to retrieve patches from CVS repo or via http. For searching, we can use <issuenum> (by number), <catname> (by category), <modname>/<submodname> (by module), branch attributes in <affects> (by branch), and <corrected> (by date). I think it is not enough and still needs more refinement. What is insufficient do you think? -- | Hiroki Sato <hrs@FreeBSD.org> | <hrs@eos.ocn.ne.jp> [-- Attachment #2 --] .3<
