Re: fmt2.myr: Landed new string formatting code.

[Ori Bernstein <ori@xxxxxxxxxxxxxx>]

And I moved to '{' and '}'.

There are still a few questions in my mind about escaping -- right now I use
'{{' and '}}', but it's a bit ugly/confusing. Still, {} works better for a
semi-common usecase:

    std.fmt("%%%", prefix, sep, suffix)

will fail with '%' because the first '%%' will be taken as a literal '%',
but it's fine with {}:

    std.fmt("{}{}{}", prefix, sep, suffix)

On Sun, 7 Jun 2015 14:36:38 -0700, Ori Bernstein <ori@xxxxxxxxxxxxxx> wrote:

> Ah, I see what you mean; I don't feel strongly about it -- if you want to
> submit a patch to change it, then it shouldn't be hard.
> 
> On Sun, 07 Jun 2015 16:13:51 -0500
> Ryan Gonzalez <rymg19@xxxxxxxxx> wrote:
> 
> > Yeah, I just found it odd. In Python, the braces are the placeholder *and* the options list.
> > 
> > 
> > On June 7, 2015 3:30:46 PM CDT, Ori Bernstein <ori@xxxxxxxxxxxxxx> wrote:
> > >Braces are only for specifying options. % is the placeholder.
> > >
> > >
> > >    std.fmt("% % %", 1, 2, 255) => "1, 2, 255"
> > >    std.fmt("% % %{x}", 1, 2, 255) => "1, 2, ff"
> > >
> > >On Sun, 07 Jun 2015 12:56:35 -0500
> > >Ryan Gonzalez <rymg19@xxxxxxxxx> wrote:
> > >
> > >> Looks neat! One thing: it seems weird to specify a percent *and* the
> > >braces. Python uses no percent with the .format function; it's just the
> > >braces.
> > >> 
> > >> I can test it on Plan 9 once I can figure out how to "git pull" with
> > >an rc script I found...
> > >> 
> > >> Curiosity: how to you commit changes to the repo from Plan 9?
> > >> 
> > >> 
> > >> On June 7, 2015 11:57:10 AM CDT, Ori Bernstein <ori@xxxxxxxxxxxxxx>
> > >wrote:
> > >> >I've got a new version of std.fmt() which uses the generated type
> > >info
> > >> >to know the types of its arguments, and which can support custom
> > >format
> > >> >functions.
> > >> >
> > >> >Format strings are now very different. '%' is the value placeholder.
> > >> >There are no other specifiers. Formatting options are passed via the
> > >> >'{opts}' after the placeholder. '%%' escapes the '%', so that a
> > >single
> > >> >'%' is printed. {} is the empty parameter list, and needs to be
> > >passed
> > >> >if you are printing a { immediately after a placeholder. Examples:
> > >> >
> > >> >    std.f2put("x=%\n", 123)
> > >> >        prints "x=123"
> > >> >
> > >> >    std.f2put("x=%{x}\n", 255)
> > >> >        prints "x=ff"
> > >> >
> > >> >    std.f2put("x=%{}{asdf}\n", 123)
> > >> >        prints "x=123{asdf}"
> > >> >
> > >> >    std.f2put("x=%, %\n", 255)
> > >> >        Kills your program because your arg list doesn't match the
> > >> >        format string (there are too few args)
> > >> >
> > >> >    std.f2put("x=%\n", 123, 456)
> > >> >        Kills your program because your arg list doesn't match the
> > >> >        format string (there are too many args)
> > >> >
> > >> >
> > >> >It hasn't replaced std.fmt() yet, and is accessible via the f2fmt()
> > >> >functions, but I intend to kill off the current std.fmt as soon as
> > >> >I verify that this is working well on plan9.
> > >> >
> > >> >Code is here:
> > >> >    http://git.eigenstate.org/ori/mc.git/tree/libstd/fmt2.myr
> > >> >
> > >> >It depends on the new type introspection api:
> > >> >    http://git.eigenstate.org/ori/mc.git/tree/libstd/introspect.myr
> > >> >
> > >> >API design reviews would be welcome. I'm not sure how happy I am
> > >> >with the introspection API, although I do like that it lets me
> > >> >look at stuff without descending into it, and without needing any
> > >> >heap allocations.
> > >> >
> > >> >-- 
> > >> >Ori Bernstein <ori@xxxxxxxxxxxxxx>
> > >> 
> > >> -- 
> > >> Sent from my Android device with K-9 Mail. Please excuse my brevity.
> > >
> > >
> > >-- 
> > >Ori Bernstein <ori@xxxxxxxxxxxxxx>
> > 
> > -- 
> > Sent from my Android device with K-9 Mail. Please excuse my brevity.
> 
> 
> -- 
> Ori Bernstein <ori@xxxxxxxxxxxxxx>
> 


-- 
    Ori Bernstein