Re: Call for help - understanding how to debug sw module

[Chris Sherlock]

No probs, good pointer I think I can work with that :-)

Chris

On Friday, January 29, 2016, Tomaž Vajngerl  wrote:

> Hi,
>
> On Thu, Jan 28, 2016 at 9:52 AM, Chris Sherlock
> > wrote:
> > I did a calculation from http://www.sven.de/dpi and apparently I have a
> pixel density of 267.02 PPI.
> >
> > Wondering if this is skewing the calculations.
>
> OS usually doesn't tell the correct monitor DPI to the application but
> usually just reports 96 DPI (if no scaling is applied). In windows you
> can change the scale levels - 100%, 125%, 150%, 200%,... which just
> adjusts what DPI is reported to the application (100% - 96DPI, 125% -
> 120DPI, 150% - 144 DPI, 200% - 192DPI...) so the application can scale
> accordingly (at least fonts). I'm not sure how OSX does this.
>
> You should check what DPI is actually reported to LO - OutputDevice
> has mnDPIX and mnDPIY - check where they are set and to what value.
> DPI influences how pixels are converted to and from actual units (mm,
> inch, twips,..) It could be that we transform between units just too
> much and the error accumulates if the DPI is set to a weird value.
> Tracking this down could be tricky.
>
> > Chris
>
> Regards, Tomaž
>
___
LibreOffice mailing list
LibreOffice@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/libreoffice

Re: Call for help - understanding how to debug sw module

[Chris Sherlock]

Well, I just read the AquaSalGraphics code and there is a sanity check that 
limits it to 200DPI!

I’m just testing now what happens if I set the max to 300. 

Chris

> On 29 Jan 2016, at 9:39 PM, Tomaž Vajngerl  wrote:
> 
> Hi,
> 
> On Thu, Jan 28, 2016 at 9:52 AM, Chris Sherlock
>  wrote:
>> I did a calculation from http://www.sven.de/dpi and apparently I have a 
>> pixel density of 267.02 PPI.
>> 
>> Wondering if this is skewing the calculations.
> 
> OS usually doesn't tell the correct monitor DPI to the application but
> usually just reports 96 DPI (if no scaling is applied). In windows you
> can change the scale levels - 100%, 125%, 150%, 200%,... which just
> adjusts what DPI is reported to the application (100% - 96DPI, 125% -
> 120DPI, 150% - 144 DPI, 200% - 192DPI...) so the application can scale
> accordingly (at least fonts). I'm not sure how OSX does this.
> 
> You should check what DPI is actually reported to LO - OutputDevice
> has mnDPIX and mnDPIY - check where they are set and to what value.
> DPI influences how pixels are converted to and from actual units (mm,
> inch, twips,..) It could be that we transform between units just too
> much and the error accumulates if the DPI is set to a weird value.
> Tracking this down could be tricky.
> 
>> Chris
> 
> Regards, Tomaž

___
LibreOffice mailing list
LibreOffice@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/libreoffice

Re: Call for help - understanding how to debug sw module

[Chris Sherlock]

Oops, my calculation is out by quite a lot! 

My resolution is actually 202.57 PPI, still just over that 200PPI hard limit. 

Chris

> On 31 Jan 2016, at 12:00 PM, Chris Sherlock  
> wrote:
> 
> Well, I just read the AquaSalGraphics code and there is a sanity check that 
> limits it to 200DPI!
> 
> I’m just testing now what happens if I set the max to 300. 
> 
> Chris
> 
>> On 29 Jan 2016, at 9:39 PM, Tomaž Vajngerl  wrote:
>> 
>> Hi,
>> 
>> On Thu, Jan 28, 2016 at 9:52 AM, Chris Sherlock
>>  wrote:
>>> I did a calculation from http://www.sven.de/dpi and apparently I have a 
>>> pixel density of 267.02 PPI.
>>> 
>>> Wondering if this is skewing the calculations.
>> 
>> OS usually doesn't tell the correct monitor DPI to the application but
>> usually just reports 96 DPI (if no scaling is applied). In windows you
>> can change the scale levels - 100%, 125%, 150%, 200%,... which just
>> adjusts what DPI is reported to the application (100% - 96DPI, 125% -
>> 120DPI, 150% - 144 DPI, 200% - 192DPI...) so the application can scale
>> accordingly (at least fonts). I'm not sure how OSX does this.
>> 
>> You should check what DPI is actually reported to LO - OutputDevice
>> has mnDPIX and mnDPIY - check where they are set and to what value.
>> DPI influences how pixels are converted to and from actual units (mm,
>> inch, twips,..) It could be that we transform between units just too
>> much and the error accumulates if the DPI is set to a weird value.
>> Tracking this down could be tricky.
>> 
>>> Chris
>> 
>> Regards, Tomaž
> 

___
LibreOffice mailing list
LibreOffice@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/libreoffice

Re: Call for help - understanding how to debug sw module

[Chris Sherlock]

OK, I didn’t read the comment above it correctly - 200 should be more than 
enough our AquaSalGraphics class doesn’t give DPI in terms of actual screen 
resolution but the logical screen resolution. 

I’ve pushed a gerrit patch for review to get the *actual* screen DPI if anyone 
wanted to review it:

https://gerrit.libreoffice.org/#/c/21948/ 

Chris

> On 31 Jan 2016, at 12:18 PM, Chris Sherlock  
> wrote:
> 
> Oops, my calculation is out by quite a lot! 
> 
> My resolution is actually 202.57 PPI, still just over that 200PPI hard limit. 
> 
> Chris
> 
>> On 31 Jan 2016, at 12:00 PM, Chris Sherlock  
>> wrote:
>> 
>> Well, I just read the AquaSalGraphics code and there is a sanity check that 
>> limits it to 200DPI!
>> 
>> I’m just testing now what happens if I set the max to 300. 
>> 
>> Chris
>> 
>>> On 29 Jan 2016, at 9:39 PM, Tomaž Vajngerl  wrote:
>>> 
>>> Hi,
>>> 
>>> On Thu, Jan 28, 2016 at 9:52 AM, Chris Sherlock
>>>  wrote:
 I did a calculation from http://www.sven.de/dpi and apparently I have a 
 pixel density of 267.02 PPI.
 
 Wondering if this is skewing the calculations.
>>> 
>>> OS usually doesn't tell the correct monitor DPI to the application but
>>> usually just reports 96 DPI (if no scaling is applied). In windows you
>>> can change the scale levels - 100%, 125%, 150%, 200%,... which just
>>> adjusts what DPI is reported to the application (100% - 96DPI, 125% -
>>> 120DPI, 150% - 144 DPI, 200% - 192DPI...) so the application can scale
>>> accordingly (at least fonts). I'm not sure how OSX does this.
>>> 
>>> You should check what DPI is actually reported to LO - OutputDevice
>>> has mnDPIX and mnDPIY - check where they are set and to what value.
>>> DPI influences how pixels are converted to and from actual units (mm,
>>> inch, twips,..) It could be that we transform between units just too
>>> much and the error accumulates if the DPI is set to a weird value.
>>> Tracking this down could be tricky.
>>> 
 Chris
>>> 
>>> Regards, Tomaž
>> 
> 

___
LibreOffice mailing list
LibreOffice@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/libreoffice

Re: Call for help - understanding how to debug sw module

[Tomaž Vajngerl]

Hi,

On Thu, Jan 28, 2016 at 9:52 AM, Chris Sherlock
 wrote:
> I did a calculation from http://www.sven.de/dpi and apparently I have a pixel 
> density of 267.02 PPI.
>
> Wondering if this is skewing the calculations.

OS usually doesn't tell the correct monitor DPI to the application but
usually just reports 96 DPI (if no scaling is applied). In windows you
can change the scale levels - 100%, 125%, 150%, 200%,... which just
adjusts what DPI is reported to the application (100% - 96DPI, 125% -
120DPI, 150% - 144 DPI, 200% - 192DPI...) so the application can scale
accordingly (at least fonts). I'm not sure how OSX does this.

You should check what DPI is actually reported to LO - OutputDevice
has mnDPIX and mnDPIY - check where they are set and to what value.
DPI influences how pixels are converted to and from actual units (mm,
inch, twips,..) It could be that we transform between units just too
much and the error accumulates if the DPI is set to a weird value.
Tracking this down could be tricky.

> Chris

Regards, Tomaž
___
LibreOffice mailing list
LibreOffice@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/libreoffice

Re: Call for help - understanding how to debug sw module

[Chris Sherlock]

I did a calculation from http://www.sven.de/dpi and apparently I have a pixel 
density of 267.02 PPI.

Wondering if this is skewing the calculations. 

Chris

> On 28 Jan 2016, at 12:13 PM, Chris Sherlock  
> wrote:
> 
> Hey guys, 
> 
> I know a few people have built on a Retina based Mac and haven’t had this 
> issue…
> 
> Any ideas why this might be occurring?
> 
> Chris
> 
>> On 22 Jan 2016, at 11:07 AM, Chris Sherlock  
>> wrote:
>> 
>> Hi all, 
>> 
>> I’m getting an error that never appears to any of the OS X buildbots or 
>> tinderboxes. I suspect that it has something to do with the fact that I have 
>> a Retina iMac. 
>> 
>> Basically, the unit test that is failing is in CppunitTest_sw_ooxmlimport:
>> 
>> Test name: testGroupshapeChildRotation::Import
>> equality assertion failed
>> - Expected: 686
>> - Actual  : 688
>> 
>> Failures !!!
>> Run: 208   Failure total: 1   Failures: 1   Errors: 0
>> 
>> I’ve got an Ubuntu pastebin of the error here:
>> 
>> http://paste.ubuntu.com/14593673/
>> 
>> Can anyone give me any pointer on how I would even *start* to troubleshoot 
>> this? A good start is a pointer to the code that this actually runs, because 
>> whilst I can read the test code, I have no knowledge of sw whatsoever and 
>> I’d like to know where in the code the size is failing on my machine. 
>> 
>> Thanks in advance for any help!
>> 
>> Chris
> 

___
LibreOffice mailing list
LibreOffice@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/libreoffice

Re: Call for help - understanding how to debug sw module

[Chris Sherlock]

If it is at all helpful, these are the hardware specs (this machine is 
seriously overkill, I got it with redundancy money a year ago):

Hardware Overview:

  Model Name:   iMac
  Model Identifier: iMac15,1
  Processor Name:   Intel Core i5
  Processor Speed:  3.5 GHz
  Number of Processors: 1
  Total Number of Cores:4
  L2 Cache (per Core):  256 KB
  L3 Cache: 6 MB
  Memory:   8 GB

AMD Radeon R9 M290X:

  Chipset Model:AMD Radeon R9 M290X
  Type: GPU
  Bus:  PCIe
  PCIe Lane Width:  x16
  VRAM (Total): 2048 MB
  Vendor:   ATI (0x1002)
  Device ID:0x6810
  Revision ID:  0x
  ROM Revision: 113-C408A0-729
  EFI Driver Version:   01.00.729
  Displays:
iMac:
  Display Type: Retina LCD
  Resolution:   5120 x 2880 Retina
  Retina:   Yes
  Pixel Depth:  30-Bit Color (ARGB2101010)
  Main Display: Yes
  Mirror:   Off
  Online:   Yes
  Built-In: Yes


At this point, I’m sort of wondering if the resolution of my screen might be 
part of the issue… another user sent me their specs and it seems that the other 
Retina displays are on laptops…

Chris



> On 28 Jan 2016, at 12:13 PM, Chris Sherlock  
> wrote:
> 
> Hey guys, 
> 
> I know a few people have built on a Retina based Mac and haven’t had this 
> issue…
> 
> Any ideas why this might be occurring?
> 
> Chris
> 
>> On 22 Jan 2016, at 11:07 AM, Chris Sherlock  
>> wrote:
>> 
>> Hi all, 
>> 
>> I’m getting an error that never appears to any of the OS X buildbots or 
>> tinderboxes. I suspect that it has something to do with the fact that I have 
>> a Retina iMac. 
>> 
>> Basically, the unit test that is failing is in CppunitTest_sw_ooxmlimport:
>> 
>> Test name: testGroupshapeChildRotation::Import
>> equality assertion failed
>> - Expected: 686
>> - Actual  : 688
>> 
>> Failures !!!
>> Run: 208   Failure total: 1   Failures: 1   Errors: 0
>> 
>> I’ve got an Ubuntu pastebin of the error here:
>> 
>> http://paste.ubuntu.com/14593673/
>> 
>> Can anyone give me any pointer on how I would even *start* to troubleshoot 
>> this? A good start is a pointer to the code that this actually runs, because 
>> whilst I can read the test code, I have no knowledge of sw whatsoever and 
>> I’d like to know where in the code the size is failing on my machine. 
>> 
>> Thanks in advance for any help!
>> 
>> Chris
> 

___
LibreOffice mailing list
LibreOffice@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/libreoffice

Re: Call for help - understanding how to debug sw module

[Chris Sherlock]

Hey guys, 

I know a few people have built on a Retina based Mac and haven’t had this issue…

Any ideas why this might be occurring?

Chris

> On 22 Jan 2016, at 11:07 AM, Chris Sherlock  
> wrote:
> 
> Hi all, 
> 
> I’m getting an error that never appears to any of the OS X buildbots or 
> tinderboxes. I suspect that it has something to do with the fact that I have 
> a Retina iMac. 
> 
> Basically, the unit test that is failing is in CppunitTest_sw_ooxmlimport:
> 
> Test name: testGroupshapeChildRotation::Import
> equality assertion failed
> - Expected: 686
> - Actual  : 688
> 
> Failures !!!
> Run: 208   Failure total: 1   Failures: 1   Errors: 0
> 
> I’ve got an Ubuntu pastebin of the error here:
> 
> http://paste.ubuntu.com/14593673/
> 
> Can anyone give me any pointer on how I would even *start* to troubleshoot 
> this? A good start is a pointer to the code that this actually runs, because 
> whilst I can read the test code, I have no knowledge of sw whatsoever and I’d 
> like to know where in the code the size is failing on my machine. 
> 
> Thanks in advance for any help!
> 
> Chris

___
LibreOffice mailing list
LibreOffice@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/libreoffice

Call for help - understanding how to debug sw module

[Chris Sherlock]

Hi all, 

I’m getting an error that never appears to any of the OS X buildbots or 
tinderboxes. I suspect that it has something to do with the fact that I have a 
Retina iMac. 

Basically, the unit test that is failing is in CppunitTest_sw_ooxmlimport:

Test name: testGroupshapeChildRotation::Import
equality assertion failed
- Expected: 686
- Actual  : 688

Failures !!!
Run: 208   Failure total: 1   Failures: 1   Errors: 0

I’ve got an Ubuntu pastebin of the error here:

http://paste.ubuntu.com/14593673/

Can anyone give me any pointer on how I would even *start* to troubleshoot 
this? A good start is a pointer to the code that this actually runs, because 
whilst I can read the test code, I have no knowledge of sw whatsoever and I’d 
like to know where in the code the size is failing on my machine. 

Thanks in advance for any help!

Chris
___
LibreOffice mailing list
LibreOffice@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/libreoffice