Re: Android port

[Po Lu]

Po Lu <luangruo@yahoo.com> writes:

> Eli Zaretskii <eliz@gnu.org> writes:
>
>>> Date: Sun, 6 Aug 2023 10:44:09 -0700
>>> Cc: Eli Zaretskii <eliz@gnu.org>, angelo.g0@libero.it, emacs-devel@gnu.org,
>>>  Bruno Haible <bruno@clisp.org>
>>> From: Paul Eggert <eggert@cs.ucla.edu>
>>> 
>>> I understand the reluctance from the Android point of view. However, 
>>> printf-posix imports 69 new source files to Emacs, and these files have 
>>> not been tested extensively with Emacs on non-Android platforms. From 
>>> the viewpoint of non-Android Emacs platforms, it's significantly less 
>>> disruptive if merging the Android branch does not add 69 new source 
>>> files that will require testing on these platforms.
>>> 
>>> And even from the Android viewpoint, no matter what we do to fix the 
>>> problem some testing needs to be done anyway, as the fix is likely to 
>>> affect Emacs in test-relevant ways.
>>
>> I think Paul makes a good point here about those modules not being
>> tested in Emacs on other platforms.  Avoiding addition of 69 files to
>> Emacs is also a non-trivial gain.
>>
>> Since Emacs 30.1 will not be released any time soon, I think we will
>> have ample time to test it without the *printf modules, and find out
>> and fix any issues this could create.
>>
>> So I suggest to give this solution a chance.
>>
>>> The idea is to get feature/android merged quickly. We can revisit the 
>>> use of Gnulib's printf-posix module later, as needed. With luck 
>>> printf-posix won't be needed, as Emacs historically has avoided use of 
>>> unusual printf features (for obvious portability reasons).
>>
>> Right.
>
> I plan to quickly test this on some old versions of Android and ack.

Cursory inspection indicates that Emacs functions correctly without the
*printf modules on Android 2.3, 4.0.1 and 4.4, so I've ommitted the
modules and merged the branch to master, after removing the w32-specific
code that disables them.

Thanks for your help.