Re: Broken makeinfo? [was: Re: Fix this]

mb> You must be using some old broken makeinfo. S> C'mon, old != broken. The question is, how many "old makeinfos" are S> out there, and how many of them are going to get tickled by that S> construct? People who download pre-packaged tarballs will not have any problems since their info files will already be up-to-date. Only CVS users will have to upgrade, or get up-to-date info files from the packaged tarballs. S> The lack of reports on xemacs-beta is quite possibly meaningless; as I S> reported recently, "make beta" does not rebuild info by default. ??? Just `make' rebuilds the info files by default. S> Try adding info to the top-level makefile dependency for beta (I don't S> grok Makefile\(.in\)+, sorry, no patch) and see what happens. S> Or maybe we don't care, because we don't care if people can build S> their own info docs or nobody ever does? (Is that why rebuilding info S> doesn't go automatically with make beta?) S> Or maybe we should think about abandoning (!) Texinfo in favor of a S> real DTD not defined by an implementation? mb> there is no obvious workaround for this brokenness, except mb> perhaps to change all the ＠ifnottex to ＠ifinfo. S> What about S> -＠node Top S> +＠node Top, Getting Started, , (dir) I merely imported the info.texi file from the texinfo distribution. Now that I look at it, there is only 2 occurrence of ＠ifnottex in man/, so it is actually reasonable to change them, for now, assuming this causes no problems for texi2html. Do we actually know that makeinfo from 3.12 fails, while 3.12f works fine? S> which reverses the other change to the Texinfo commands in that S> region. (The three ＠nodes that don't have a full complement of S> pointers were all introduced by your patch, probably need to change S> the others too.) Dunno if that will help, but I know makeinfo has S> been sensitive to xref issues in the past, and has never been good at S> giving useful error messages. Martin