Re: [fpc-devel] darwin indirect symbols

2009-11-10 Thread Jonas Maebe
On 10 Nov 2009, at 08:35, dmitry boyarintsev wrote: There're 2 ways to define indirect symbol by the compiler: tai_const with type aitconst_indirect_symbol or tai_directive with asd_indirect_symbol either way produces identical assembler output: .indirect_symbol %symbol_name So i'm

Re: [fpc-devel] darwin indirect symbols

2009-11-10 Thread dmitry boyarintsev
On Tue, Nov 10, 2009 at 3:50 PM, Jonas Maebe jonas.ma...@elis.ugent.be wrote: I think simply that I forgot that one way was already available when I added the other way :) It's fine to remove either one. hmm... I'll try to make the patch. thanks, dmitry

Re: [fpc-devel] darwin indirect symbols

2009-11-10 Thread dmitry boyarintsev
On Tue, Nov 10, 2009 at 3:53 PM, dmitry boyarintsev skalogryz.li...@gmail.com wrote: hmm... I'll try to make the patch. here's the one: http://bugs.freepascal.org/view.php?id=15050 Removing tai_const with indirect_symbol doesn't seem to harm the generated asm code at all, at least for i386

[fpc-devel] darwin indirect symbols

2009-11-09 Thread dmitry boyarintsev
Hello Jonas, There're 2 ways to define indirect symbol by the compiler: tai_const with type aitconst_indirect_symbol or tai_directive with asd_indirect_symbol either way produces identical assembler output: .indirect_symbol %symbol_name So i'm wondering, what was the reason to separate them