RE: Pike 8.0 RC1
郭雪松 <[email protected]>
| Newsgroups | gmane.comp.lang.pike.user |
|---|---|
| Message-ID | <[email protected]> |
$ pike80 Pike v8.0 release 10 running Hilfe v3.5 (Incremental Pike Frontend) > Thread.Local()->set(([])); (1) Result: ([ ]) > Thread.Local()->get(); (2) Result: 0 From: [email protected] To: [email protected] Subject: RE: Pike 8.0 RC1 Date: Sat, 25 Oct 2014 11:02:31 +0800 Bug report: Pike v7.9 release 11 running Hilfe v3.5 (Incremental Pike Frontend) > Thread.Local()->set(([])); (1) Result: ([ ]) > Thread.Local()->get(); (2) Result: 0 > Date: Fri, 24 Oct 2014 09:54:52 +0200 > From: [email protected] > To: [email protected] > Subject: Re: Pike 8.0 RC1 > > > On 10/22/14 20:21, Arne Goedeke wrote: > > > > > > On Wed, 22 Oct 2014, Henrik Grubbstr�m wrote: > > > >> On Wed, 22 Oct 2014, Arne Goedeke wrote: > >> > >>> > Also, if you have anything you really want to fix in pike 8.0, please > >>> > say something about it, or commit a fix soon. > >>> > >>> We are currently not completely compatible with nettle 2.7, which is > >>> what gentoo is still shipping. The problem, which has already been > >>> discussed here, is that at some point nettle changed the signatures for > >>> random functions. Should that be fixed? > >> > >> I've never noticed. In what way does the problem manifest? > >> > >> BTW: I've just added an ebuild for Pike 8.0.10 to the pike gentoo > >> overlay, > >> and it seems to install fine on at least on of my machines. > > > > The second argument of the random function was changed from unsigned int > > to size_t at some point (I guess in nettle 3 or so). > > > > dsa_generate_keypair(struct dsa_public_key *pub, > > /usr/include/nettle/dsa.h:200:1: note: expected 'void (*)(void *, > > unsigned int, uint8_t *)' but argument is of type 'void (*)(void *, > > size_t, uint8_t *)' > > dsa_generate_keypair(struct dsa_public_key *pub, > > > > It builds and installs fine for me, too. The code also works; I guess it > > doesnt matter with x86-64 calling conventions. > > On reading the nettle git history it seems like this was part of the > same changes that pike_nettle_size_t was introduced for. >