On Sun, Feb 24, 2008 at 09:24:38PM +0200, Martin-Éric Racine wrote: > Package: dictionaries-common-dev > Version: 0.96.0 > Severity: important > > -----BEGIN PGP SIGNED MESSAGE----- > Hash: SHA1 > > ***During package building*** > > dh_installdirs -paspell-lv > installdeb-aspell --package=aspell-lv > Name "main::no_pre_post" used only once: possible typo at > /usr/bin/installdeb-aspell line 58.
This is harmless, as a matter of fact a fix was committed last friday and I planned an upload for tomorrow, but just to avoid noise. > # Override: usr-lib-in-arch-all (all Aspell dictionaries are packaged this > way by design) > cp debian/aspell-lv.linda > debian/aspell-lv/usr/share/linda/overrides/aspell-lv > cp debian/aspell-lv.lintian > debian/aspell-lv/usr/share/lintian/overrides/aspell-lv I this a new lintian test?. I do not remember it. > > ***During install*** > > <[EMAIL PROTECTED]:/home/q-funk/Projects/myspell-lv-0.7.3>$ LC_ALL=C sudo > aspell-autobuildhash --debug Did it complain during normal install? > lv => aspell_compat: > [0.60_c1]; lang_compat: [0] > aspell-autobuildhash: processing: lv [lv] ----------------- > " does not guarantee that "s" can always be stripped.ion "s ----------------- The above is the error, and is an aspell error message regarding your dict, where the bug seems to be located. I do not have a buggy dict at hand to try this, but IIRC, that error is triggered by the absence of the --dont-validate-affixes option. The use of --debug enables this test. > dictionaries-common: Running aspell-autobuildhash > - ------------------------------------------------- ... > ***This is all nice and dandy, except that the --debug option is basically > useless, because it doesn't show which > word or which line number makes the hash building fail. What the --debug option intends is to help debugging of aspell-autobuildhash, not aspell, that provides no debug option, and the only --debug extra addition is that if set, will disable --dont-validate-affixes and so, a check of affixes will be carried out, so checking of dicts is easier. You have all the info provided by aspell, that may be enough. > Then again, even that could be cause by a bug in the -dev package?*** or even in libc6 ;-). However this is usually a problem in the dict, so that is the place to start. Not sure about the package (rus-ispell?), but I remember something like this being the reason for adding the --dont-validate-affixes option, so packages do not get stalled in case of non responsive upstream. Will try looking into my ancient mailbox tomorrow. Unless I find a real problem I will downgrade and retitle this bug report as $no_pre_post noise related and close it with the upload. Will also try looking at your package tomorrow in case I can suggest something. If is a new upstream version, let me know. -- Agustin
