Development mailing list

On 2013-03-16 11:00, X512 wrote: I have x86 based Acer W500 tablet PC and try to use Haiku on it. One Nice dos it run ok on that thing? of most important problem is impossibility of using Bluetooth keyboard. Without keyboard it's hard to do something useful. Onscreen ...

I'm not particularly responding to any part of this conversation in general, but wanted to thank everyone, like Pawel, for exploring this. I also just noticed that the GCC 4.9 branch is in effect as of today, so maybe I'll start looking at targeting GCC 4.8.0 for Haiku instead of GCC 4.7.2. There are some dependency bugs that were afflicting us (that I think are fixed in the 4.8 branch), so maybe that can be leveraged to get our GCC up to date compared to 4.6.3. I'm not sure if issues affecting ftree-vrp are addressed in a satisfactory way in the 4.8 branch without looking further, but it's possible, I suppose. ...

Haiku Developers (esp. Oliver), If you could spend a few minutes of your time to help me figure out a problem I'm having I'd really appreciate it. Take a look at this screenshot from the User Guide: Under version info look at the short description and long description text boxes. The text in the short description, which is represented by ...

I have x86 based Acer W500 tablet PC and try to use Haiku on it. One of most important problem is impossibility of using Bluetooth keyboard. Without keyboard it's hard to do something useful. Onscreen keyboard is also missing. Keymap preflet can be used as onscreen keyboard, but it can input only to specific field. To input in another window it's required to drag buttons to that window (that's very funny but not so useful) or copy-paste text from sample input field. So it will be nice to have Bluetooth support. Any progress had been made ...

2013/3/15 Axel Dörfler axeld@xxxxxxxxxxxxxxxx: Am 15/03/2013 15:41, schrieb Pawel Dziepak: 2013/3/14 Axel Dörfler axeld@xxxxxxxxxxxxxxxx: Isn't that just a porting issue? Somewhere GCC stores that information, and it's usually right. Just not on Haiku. IMO one should simply update GCC to know that strdup(), free(), etc. accept NULL pointers. ...

Am 15/03/2013 15:41, schrieb Pawel Dziepak: 2013/3/14 Axel Dörfler axeld@xxxxxxxxxxxxxxxx: Isn't that just a porting issue? Somewhere GCC stores that information, and it's usually right. Just not on Haiku. IMO one should simply update GCC to know that strdup(), free(), etc. accept NULL pointers. ...

2013/3/14 Axel Dörfler axeld@xxxxxxxxxxxxxxxx: Am 14/03/2013 18:01, schrieb Pawel Dziepak: 2013/3/14 Ingo Weinhold ingo_weinhold@xxxxxx: Or IOW, the feature is just buggy. It would be interesting whether the bug has been fixed in newer gcc versions. I did some tests on a simplified example on GCC 4.7.2. What I found ...

A little status update, I've been trying to get binutils to add PE target so objcopy can convert ELF EFI apps to PE apps, but it is not going well. In buildtools/build-gcc4-optional-package-Haiku.sh I'd expect to add '--enable-targets=[possible same as --target],i586-haiku-pe,x86_64-haiku-pe and maybe --enable-64-bit-bfd as I'm hoping to do both 32- and 64-bit. But atm it seems I have to do a lot of brute force testing maybe restricting to just 32 bit or 64 bit depending on what we are building. ...

Am 14/03/2013 18:01, schrieb Pawel Dziepak: 2013/3/14 Ingo Weinhold ingo_weinhold@xxxxxx: Or IOW, the feature is just buggy. It would be interesting whether the bug has been fixed in newer gcc versions. I did some tests on a simplified example on GCC 4.7.2. What I found out is that although GCC does not replace strdup() with its builtin ...

Ingo Weinhold писал 14.03.2013 16:47: Or IOW, the feature is just buggy. It would be interesting whether the bug has been fixed in newer gcc versions. By the way, after workarounding the issue with Desktop crash following applications cannot start with the same strlen-strdup backtrace: Web+, ...

2013/3/14 Ingo Weinhold ingo_weinhold@xxxxxx: On 03/14/2013 04:20 PM, Pawel Dziepak wrote: 2013/3/14 Siarzhuk Zharski zharik@xxxxxx: -fno-tree-vrp is not set: 0009aa60 strdup: snip ...

On 03/14/2013 04:20 PM, Pawel Dziepak wrote: 2013/3/14 Siarzhuk Zharski zharik@xxxxxx: -fno-tree-vrp is not set: 0009aa60 strdup: snip ...

Hi Pawel, Von: Pawel Dziepak pdziepak@xxxxxxxxxxx [...] Looks like the NULL pointer check indeed is removed. GCC manual [1] may give an answer to why it is happening. Description of -ftree-vrp: Perform Value Range Propagation on trees. (...) This allows the optimizers to remove unnecessary range checks like array bound checks and null pointer checks. (...) Null pointer check elimination is only done if -fdelete-null-pointer-checks is enabled. ...

On Thu, Mar 14, 2013 at 3:27 PM, Pawel Dziepak pdziepak@xxxxxxxxxxx wrote: 2013/3/14 Edward Robbins edd.robbins@xxxxxxxxxxxxxx: Interesting discussion. Are you sure those aren't the wrong way round? Because the NULL test looks to be in the 2nd but not the first: 9aa78: 31 f6 xor %esi,%esi 9aa7a: 85 ff test %edi,%edi 9aa7c: 74 2f je 9aaad strdup+0x4d 9aa7e: 83 ec 0c sub $0xc,%esp ...

2013/3/14 Pawel Dziepak pdziepak@xxxxxxxxxxx: Both of these flags are enabled by default by -O2. Actually it looks like delete-null-pointer-checks is doing a bit more (probably recognizing strdup) than it is said in its description. Siarzhuk, could you check whether setting -fno-delete-null-pointer-checks and leaving vrp enabled solves the problem? It is also worth to mention flag -fdump-tree-vrp which dumps each function after VRP. Enabling this (probably only for one file since it ...

2013/3/14 Edward Robbins edd.robbins@xxxxxxxxxxxxxx: Interesting discussion. Are you sure those aren't the wrong way round? Because the NULL test looks to be in the 2nd but not the first: 9aa78: 31 f6 xor %esi,%esi 9aa7a: 85 ff test %edi,%edi 9aa7c: 74 2f je 9aaad strdup+0x4d 9aa7e: 83 ec 0c sub $0xc,%esp The second one is with VRP disabled (a flag which disables it: ...