Re: [PATCH v2 0/4] New getprogname module

[T J]

Hello,

I had some trouble posting to the list, so I'm not sure if my previous message got through or not. I apologise if it has.

Anyway, it appears that this has not been tested on Windows/MinGW as it currently does not work. Everything worked fine with the old progname module.

A sample build log is here: https://ci.appveyor.com/project/fontforge/fontforge/build/1.0.32

Thanks,
Jeremy

---

From: bug-gnulib <bug-gnulib-bounces+address@hidden> on behalf of Jim Meyering <address@hidden>
Sent: Tuesday, 6 September 2016 12:47 PM
To: Pino Toscano
Cc: address@hidden List
Subject: Re: [PATCH v2 0/4] New getprogname module

 

On Mon, Sep 5, 2016 at 9:22 AM, Jim Meyering <address@hidden> wrote:
> On Mon, Sep 5, 2016 at 7:28 AM, Pino Toscano <address@hidden> wrote:
>> On Saturday, 3 September 2016 20:47:15 CEST Jim Meyering wrote:
> ...
>> Another thing: should some deprecation warning/note be added regarding
>> the progname module?
>
> I like the idea of adding a deprecation warning.
> If it could be completely replaced, I'd suggest to add the "Status:
> deprecated" attribute to its modules file, but we don't have a
> replacement for set_program_name, so there may still be legitimate
> uses. If a future change were to move set_program_name into its own
> new module, *then*, we could officially deprecate the progname module.
>
>> Is NEWS the proper place for them? Attached there
>> is a small documentation addendum.
>
> Good idea.
> While this is not officially an incompatible change, converting is
> invasive enough that this NEWS blurb belongs in that section.
>
> I've split a long sentence and merged that into your first commit.
> And pushed.
>
>>> I'm prepared to push the attached, but will wait for your ack.
>
> If you're interested, one more thing that may help avoid trouble would
> be to add a syntax-check rule to prohibit new uses of this module,
> including new inclusion of progname.h, new declarations of
> program_name or anything else you can think of that should no longer
> be done here in gnulib.

FYI, while adapting grep to use this module, I encountered a single
new error/warning. The attached patch fixes that: