Black screen on Mac OSX

> Hi,
>
> I just noted DrGeo 17.07 (pharo 3 based) starts with a black screen on Mac OS X 10.12.6.
>
> Can anyone with such configuration confirm it?
>
> Thanks
>
> Hilaire
>
>
> Dr. Geo -- http://drgeo.eu
>

> Hi hilaire
>
> I'm on 10.12.6 and I downloaded the latest version of DRGeo I can find
> on the link in your signature.
> First I had problem to launch it because it is blocked by the security
> manager and
> yes I can confirm that the system just open a black screen.
> I had to kill the app.
>
> Stef

> Hi,
>
> Thanks for the test. So I have to update it.
>
> Hilaire
>
>
> Le 26/03/2018 à 19:58, Stephane Ducasse a écrit :
>>
>> Hi hilaire
>>
>> I'm on 10.12.6 and I downloaded the latest version of DRGeo I can find
>> on the link in your signature.
>> First I had problem to launch it because it is blocked by the security
>> manager and
>> yes I can confirm that the system just open a black screen.
>> I had to kill the app.
>>
>> Stef
>
>
> --
> Dr. Geo
> http://drgeo.eu
>
>
>

> Le 26 mars 2018 à 14:19, H. Fernandes <[hidden email]> a écrit :
>
> Hi,
>
> I just noted DrGeo 17.07 (pharo 3 based) starts with a black screen on Mac OS X 10.12.6.
>
> Can anyone with such configuration confirm it?
>
> Thanks
>
> Hilaire
>
>
> Dr. Geo -- http://drgeo.eu
>

> Hi hilaire
>
> I got startup errors
> I debugged a bit and you are calling to create a directory
>
> createDirectory: path
>
>         Path / 'var' / 'folders' / 'gs' /
> '_5w5v50j3m7gl4lcqf9npxhr0000gq' / 'T' / 'AppTranslocation' /
> '863D37C1-90CD-4901-90DA-F4EC3F4B3080' / 'd' / 'locale'
>
> the encode pathString is
> '/var/folders/gs/_5w5v50j3m7gl4lcqf9npxhr0000gq/T/AppTranslocation/863D37C1-90CD-4901-90DA-F4EC3F4B3080/d/locale'
>
>
> And this leads to a primitive failure
>
> createDirectory: path
> "Create a directory for the argument path.
> If the path refers to an existing file, raise FileExists.
> If the path refers to an existing directory, raise DirectoryExists.
> If the parent directory of the path does not exist, raise DirectoryDoesNotExist"
>
> | parent encodedPathString pathString result |
> pathString := self stringFromPath: path.
> encodedPathString := Primitives encode: pathString.
> result := Primitives createDirectory: encodedPathString.
> result
>    ifNil: [
>     parent := path parent.
>    (self exists: path)
>     ifTrue: [
>      (self isFile: path)
>        ifTrue: [ self signalFileExists: path ]
>        ifFalse: [ self signalDirectoryExists: path ] ].
>   (self isDirectory: parent)
>          ifFalse: [ ^ self signalDirectoryDoesNotExist: parent ].
> self primitiveFailed ].
> ^^^^^^^^^^^^^^^^^^
> It fails here
>
> ^ self
>
>
> On my machine I get
>
> var/folders/gs/_5w5v50j3m7gl4lcqf9npxhr0000gq/T/AppTranslocation
>
> and not the folder d
>
>
> This comes from reset of the GetTextTranslator
>
> privateStartUp
>     self reset.
>     GetTextTranslator reset.
>     self localeChanged.
>
>
> Tell me how I can help.
>
>
> On Tue, Mar 27, 2018 at 7:15 PM, Hilaire <[hidden email]> wrote:
>> Thanks for the tip
>>
>>
>> Le 27/03/2018 à 08:59, Christophe Demarey a écrit :
>>>
>>> If the VM is too old, it does not work well on recent distributions.
>>> You could try to use the stable VM of Pharo 4 instead and it should work.
>>
>>
>> --
>> Dr. Geo
>> http://drgeo.eu
>>
>>
>>

> No I am not defining such extra method. I more or less packaged blindly a
> modified universal VMs for P4. I was able to test it on linux and windows
> (throught wine), but not on Mac osx.
>
> I am doing right now the same for P7 based image, but not there yet.
>
>
> Le 03/04/2018 à 22:39, Stephane Ducasse a écrit :
>>
>> Hilaire
>>
>> I looked at the other bug.
>> And I would like to know if you are defining extra
>> methods on SQSurfaceDispatch class.
>> I do not get why I got an error (may be this is related to the source
>> management.
>>
>> Stef
>
>
> --
> Dr. Geo
> http://drgeo.eu
>
>
>

> Le 04/04/2018 à 12:13, Stephane Ducasse a écrit :
>>
>> Let me know for P7 because this is easier to check and help.
>
> Ah, so please test this one on your Mac. It is a P7 image based built:
> https://www.dropbox.com/s/f51fh7wa9x52grv/DrGeo.app-18.06a.zip?dl=0
>
>
> Thanks
>
> Hilaire
>
>
> --
> Dr. Geo
> http://drgeo.eu
>
>
>

> Hi Hilaire
>
> The app goes to the dock but nothing happens. What is strange is that
> it is assigned to none display.
> I tried to open the drgeo.image with my working Pharo6.1VM and it did not work.
> I tried to open the drgeo.image with my working Pharo70VM and it did not work.
> => did not open.
>
> I could open one of my Phaor 70 images with the VM you ship.
>
> I have no log files or others. So I do not know what to do.
>
> Stef

> Hi Hilaire
>
> The app goes to the dock but nothing happens. What is strange is that
> it is assigned to none display.
> I tried to open the drgeo.image with my working Pharo6.1VM and it did not work.
> I tried to open the drgeo.image with my working Pharo70VM and it did not work.
> => did not open.
>
> I could open one of my Phaor 70 images with the VM you ship.
>
> I have no log files or others. So I do not know what to do.

> Hi Hilaire,
>   thanks for spliting the subject, I have missed the discussion in the
> other thread.
>
> The #unload message should be used to clean up the system in case the
> class is unload from the system.
> That's why the logic has been move to the #resetCurrent method that is
> called during the shutdown of the image.
>
> However, the question is not if the method should be included in a
> compatibility package or if the users should send #resetCurrent.
> The problem is why it is needed to clean up the font cache, it should
> not be needed. The only scenario I think is if you want to replace a
> font (in the Pharo world a Face, because there is a font per each
> combination of face and point size) named exactly the same with
> another one. I think this is no normal behavior.
>
> I will try to see why this is needed when loading DrGeo in Pharo7 and
> fix the FT implementation not to need it.
> Maybe it is hiding a bigger issue.
>