Windows build broken on trunk

Windows build broken on trunk

[Andy Moreton]

Hi,

Bootstrapping trunk on Windows (mingw32 GCC 4.6.1) failed this morning:

w32font.c:78:40: error: static declaration of 'Qsymbol' follows non-static declaration
lisp.h:2335:41: note: previous declaration of 'Qsymbol' was here
mingw32-make[3]: *** [oo/i386/w32font.o] Error 1
mingw32-make[3]: Leaving directory `C:/emacs/src/emacs/trunk/src'

This appers to be from r109131 committed by Dmitry Antipov.

I'm not familiar with the code: is the static Qsymbol used in w32font.c
referring to some aspect of font handling, or is it the same thing as the
global that was added to src/lisp.h ?

   AndyM

Re: Windows build broken on trunk

[Dmitry Antipov]

On 07/18/2012 12:42 PM, Andy Moreton wrote:

> Bootstrapping trunk on Windows (mingw32 GCC 4.6.1) failed this morning:
>
> w32font.c:78:40: error: static declaration of 'Qsymbol' follows non-static declaration
> lisp.h:2335:41: note: previous declaration of 'Qsymbol' was here
> mingw32-make[3]: *** [oo/i386/w32font.o] Error 1
> mingw32-make[3]: Leaving directory `C:/emacs/src/emacs/trunk/src'
>
> This appers to be from r109131 committed by Dmitry Antipov.
>
> I'm not familiar with the code: is the static Qsymbol used in w32font.c
> referring to some aspect of font handling, or is it the same thing as the
> global that was added to src/lisp.h ?

IIUC it's font handling stuff, and (IMHO) it should be renamed from too
generic name to something which is closely related to symbol's real use.

Dmitry

Re: Windows build broken on trunk

[Andreas Schwab]

Dmitry Antipov <dmantipov@yandex.ru> writes:

> IIUC it's font handling stuff, and (IMHO) it should be renamed from too
> generic name to something which is closely related to symbol's real use.

The symbol's real use is the Lisp symbol symbol.

Andreas.

-- 
Andreas Schwab, schwab@linux-m68k.org
GPG Key fingerprint = 58CA 54C7 6D53 942B 1756  01D3 44D5 214B 8276 4ED5
"And now for something completely different."

Re: Windows build broken on trunk

[Andy Moreton]

On Wed 18 Jul 2012, Andreas Schwab wrote:

> Dmitry Antipov <dmantipov@yandex.ru> writes:
>
>> IIUC it's font handling stuff, and (IMHO) it should be renamed from too
>> generic name to something which is closely related to symbol's real use.
>
> The symbol's real use is the Lisp symbol symbol.
>
> Andreas.

I think you are incorrect on this point. The manual has this in node
"Windows Fonts":

`script'
     Specifies a Unicode subrange the font should support.

     The following scripts are recognized on Windows: `latin', `greek',
     `coptic', `cyrillic', `armenian', `hebrew', `arabic', `syriac',
     `nko', `thaana', `devanagari', `bengali', `gurmukhi', `gujarati',
     `oriya', `tamil', `telugu', `kannada', `malayam', `sinhala',
     `thai', `lao', `tibetan', `myanmar', `georgian', `hangul',
     `ethiopic', `cherokee', `canadian-aboriginal', `ogham', `runic',
     `khmer', `mongolian', `symbol', `braille', `han',
     `ideographic-description', `cjk-misc', `kana', `bopomofo',
     `kanbun', `yi', `byzantine-musical-symbol', `musical-symbol', and
     `mathematical'.

From src/w32font.c, line 79:

static Lisp_Object Qkhmer, Qmongolian, Qsymbol, Qbraille, Qhan;

This Qsymbol is used for the script property, and is distinct from the
Lisp symbol symbol used in the rest of emacs.

The name clash is unfortunate.

    AndyM

Re: Windows build broken on trunk

[Andreas Schwab]

Andy Moreton <andrewjmoreton@gmail.com> writes:

> This Qsymbol is used for the script property, and is distinct from the
> Lisp symbol symbol used in the rest of emacs.

Interned Lisp symbols are globally unique.

Andreas.

-- 
Andreas Schwab, schwab@linux-m68k.org
GPG Key fingerprint = 58CA 54C7 6D53 942B 1756  01D3 44D5 214B 8276 4ED5
"And now for something completely different."

Re: Windows build broken on trunk

[Jason Rumney]

Andy Moreton <andrewjmoreton@gmail.com> writes:

> This Qsymbol is used for the script property, and is distinct from the
> Lisp symbol symbol used in the rest of emacs.
>
> The name clash is unfortunate.

The name clash is inconsequential.  The static definition in w32font.c
can be removed if symbol is now globally defined.

Re: Windows build broken on trunk

[Stefan Monnier]

> The name clash is inconsequential.  The static definition in w32font.c
> can be removed if symbol is now globally defined.

The "Lisp_Object Q<foo>" declarations should be auto-generated in
globals.h from the DEFSYM calls, IMNSHO.


        Stefan

Re: Windows build broken on trunk

[Samuel Bronson]

On Jul 19, 2012, at 3:05 AM, Stefan Monnier wrote:

>> The name clash is inconsequential.  The static definition in  
>> w32font.c
>> can be removed if symbol is now globally defined.
>
> The "Lisp_Object Q<foo>" declarations should be auto-generated in
> globals.h from the DEFSYM calls, IMNSHO.

Isn't that a bit excessive?

Re: Windows build broken on trunk

[Stefan Monnier]

>>> The name clash is inconsequential.  The static definition in w32font.c
>>> can be removed if symbol is now globally defined.
>> The "Lisp_Object Q<foo>" declarations should be auto-generated in
>> globals.h from the DEFSYM calls, IMNSHO.
> Isn't that a bit excessive?

Why?


        Stefan

Re: Windows build broken on trunk

[Andreas Schwab]

Stefan Monnier <monnier@IRO.UMontreal.CA> writes:

>> The name clash is inconsequential.  The static definition in w32font.c
>> can be removed if symbol is now globally defined.
>
> The "Lisp_Object Q<foo>" declarations should be auto-generated in
> globals.h from the DEFSYM calls, IMNSHO.

Those are definitions, not declarations, so they cannot go to globals.h.

Andreas.

-- 
Andreas Schwab, schwab@linux-m68k.org
GPG Key fingerprint = 58CA 54C7 6D53 942B 1756  01D3 44D5 214B 8276 4ED5
"And now for something completely different."

Re: Windows build broken on trunk

[Stephen J. Turnbull]

Andreas Schwab writes:
 > Stefan Monnier <monnier@IRO.UMontreal.CA> writes:
 > 
 > >> The name clash is inconsequential.  The static definition in w32font.c
 > >> can be removed if symbol is now globally defined.
 > >
 > > The "Lisp_Object Q<foo>" declarations should be auto-generated in
 > > globals.h from the DEFSYM calls, IMNSHO.
 > 
 > Those are definitions, not declarations, so they cannot go to globals.h.

Doesn't matter, they are data, *not* code, for Stefan's application.

In-best-Andreas-Schwab-style-ly y'rs,

Re: Windows build broken on trunk

[Stefan Monnier]

>>> The name clash is inconsequential.  The static definition in w32font.c
>>> can be removed if symbol is now globally defined.
>> The "Lisp_Object Q<foo>" declarations should be auto-generated in
>> globals.h from the DEFSYM calls, IMNSHO.
> Those are definitions, not declarations, so they cannot go to globals.h.

Duh, you're right.


        Stefan