On Sep 16, 2016 07:06, "Jason Zaman" <jason@perfinion.com> wrote:
>
> On Fri, Sep 16, 2016 at 06:51:25AM -0700, William Roberts wrote:
> > On Fri, Sep 16, 2016 at 6:43 AM, William Roberts
> > <bill.c.roberts@gmail.com> wrote:
> > > On Fri, Sep 16, 2016 at 6:31 AM, Jason Zaman <jason@perfinion.com>
wrote:
> > >> On Fri, Sep 16, 2016 at 06:15:01AM -0700, William Roberts wrote:
> > >>> On Fri, Sep 16, 2016 at 6:09 AM, Janis Danisevskis <
jdanis@google.com> wrote:
> > >>> > I don't mind. Then before sefcontext_compile -r gets widely
adapted we
> > >>> > should change the semantic quickly. I'll prepare a patch.
> > >>>
> > >>> Did I miss something and this was merged? Iv'e been out recovering
> > >>> from a surgery so I haven't been
> > >>> following this as well as I normally would have,
> > >>>
> > >>> If its merged, just leave it.
> > >>
> > >> Its the very latest thing in master yeah, but I do also agree with
changing it.
> > >
> > > I'd prefer it changed myself. Another argument pro-change is that one
doesn't
> > > know if its a PCRE2 or PCRE1 compatable version of the libraries, we
should
> > > probably have an ability to intergoate that via API so we can print it
> > > out in help
> > > dialogue, that this tool so we at least know what it can support.
> > >
> > > The biggest thing that needs to get fixed with this, is that no matter
> > > if it contains the
> > > pre-compiled regexs or not, it should always load and work on Android.
> > > In distros,
> > > it will fall back to file_contexts, but we don't have this in Android.
> > > This ties into the arch
> > > version information below. But if the arch differs, always recompile.
> > > The alternative to
> > > this, is just go back to textual fc file son Android, since we won't
> > > be using any of the
> > > features of binary fc's. Better yet, in the Android build, I would
> > > check to see if host arch
> > > is the same as target arch, or let an OEM set a flag, to do the
> > > compilation. But I would
> > > consider those stretch goals, and just revert android back to textual
files.
> >
> > My ramblings might not be clear here. If we just version it with arch
> > and require
> > an exact match, we don't need -r, so that option goes away.
>
> This is what I was aiming for too, currently the pcre_version() is just
> strcmp'd so changing it to cat(pcre2_version(), arch_string()) should
> avoid any problems. It may turn out to be overzealous but it will always
> be safe. I dont think splitting pcre_version and arch into separate
> fields makes a huge difference if you prefer that. Currently the version

It can make a difference if we ever need to parse any information for some
reason in the future. We'd then have to split them back apart. This is just
a bit more future proof IMO.

> numbers are not parsed tho, so checking arch only for pcre2 seems like a
> lot more work.
>
> -- Jason