digitalmars.D - Holy Win32 Voids, Batman!
- kris (17/17) Jan 23 2007 W00t! Poll Time!
- Pragma (7/30) Jan 23 2007 I must be late to the party, or mislaid my invitation.
- kris (6/40) Jan 23 2007 It's all legal D, and kosher for C-based code. The downside is that when...
- John Reimer (7/14) Jan 23 2007 Aye... I forgot about that D feature... and it's even documented. It's...
W00t! Poll Time! All you Win32 developers out there -- how do you feel about marking all instances of Win32 structs like so: POINT p = void; ... ... ... TV_ITEM tv = void; ... TEXTMETRIC tm = void; ensuring that there is an "= void;" affixed to each decl? If you're all happy with that, we can eliminate those issues with Win32 bloat without a fix from Walter. Woohoo! All those in favour please say 'aye'. All those not in favour say 'nae'. Those in favour of a true fix say nothing at all -- that way, we can truthfully assert most ppl want a real fix instead ;) (see related posts also)
Jan 23 2007
kris wrote:W00t! Poll Time! All you Win32 developers out there -- how do you feel about marking all instances of Win32 structs like so: POINT p = void; ... ... ... TV_ITEM tv = void; ... TEXTMETRIC tm = void; ensuring that there is an "= void;" affixed to each decl? If you're all happy with that, we can eliminate those issues with Win32 bloat without a fix from Walter. Woohoo! All those in favour please say 'aye'. All those not in favour say 'nae'. Those in favour of a true fix say nothing at all -- that way, we can truthfully assert most ppl want a real fix instead ;) (see related posts also)I must be late to the party, or mislaid my invitation. Is this attempting to solve the link size of executables? If so, and this changes nothing but the level of bloat we've all come to expect, then "AYE". As long as it can't be misinterpreted by other (potential) compilers, and is side-effect free code then there's no reason not to do this. -- - EricAnderton at yahoo
Jan 23 2007
Pragma wrote:kris wrote:It's all legal D, and kosher for C-based code. The downside is that when you miss out certain " = void;" (for specific Win32 structs) you'll get a ungrokable linker error. See Sean's recent post on "Big problem with Small programs" for the true culprit.W00t! Poll Time! All you Win32 developers out there -- how do you feel about marking all instances of Win32 structs like so: POINT p = void; ... ... ... TV_ITEM tv = void; ... TEXTMETRIC tm = void; ensuring that there is an "= void;" affixed to each decl? If you're all happy with that, we can eliminate those issues with Win32 bloat without a fix from Walter. Woohoo! All those in favour please say 'aye'. All those not in favour say 'nae'. Those in favour of a true fix say nothing at all -- that way, we can truthfully assert most ppl want a real fix instead ;) (see related posts also)I must be late to the party, or mislaid my invitation. Is this attempting to solve the link size of executables? If so, and this changes nothing but the level of bloat we've all come to expect, then "AYE". As long as it can't be misinterpreted by other (potential) compilers, and is side-effect free code then there's no reason not to do this.
Jan 23 2007
On Tue, 23 Jan 2007 13:47:28 -0800, kris wrote:It's all legal D, and kosher for C-based code. The downside is that when you miss out certain " = void;" (for specific Win32 structs) you'll get a ungrokable linker error. See Sean's recent post on "Big problem with Small programs" for the true culprit.Aye... I forgot about that D feature... and it's even documented. It's a perfect workaround for this situation (and you've verified it works, I assume). You must have distracted me by titling you posts with all those annoyingly catchy quotes and references. :D -JJR
Jan 23 2007