Win10 install, missing fonts or translation?

classic Classic list List threaded Threaded
19 messages Options
Reply | Threaded
Open this post in threaded view
|

Win10 install, missing fonts or translation?

Phil Miesle
Hello,

I have installed GrampsAIO64-5.0.2 into Windows 10 using the downloaded installer, and when I launch the program I end up with no letters, only square boxes, no letters apart from the Windows title bar:


I've tried installing a new theme but that hasn't worked, which leads me to believe I have a language problem?  My language is normally "English (Ireland)" but I have switched it to "English (United States)" to no avail...

Am I missing something?


--
Gramps-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gramps-users
https://gramps-project.org
Reply | Threaded
Open this post in threaded view
|

Re: Win10 install, missing fonts or translation?

Dave Scheipers
Hi Phil

In your Theme, what are you using as your font? When exploring the
theme, I noticed that some fonts did not render correctly.

Dave

On Wed, Aug 14, 2019 at 8:22 AM Phil Miesle <[hidden email]> wrote:

>
> Hello,
>
> I have installed GrampsAIO64-5.0.2 into Windows 10 using the downloaded installer, and when I launch the program I end up with no letters, only square boxes, no letters apart from the Windows title bar:
>
> https://drive.google.com/open?id=0BwzY2jGxsoM0NVhuWlViRXhnLXRmVVRHMXlkLWpibTBEUVdB
>
> I've tried installing a new theme but that hasn't worked, which leads me to believe I have a language problem?  My language is normally "English (Ireland)" but I have switched it to "English (United States)" to no avail...
>
> Am I missing something?
> --
> Gramps-users mailing list
> [hidden email]
> https://lists.sourceforge.net/lists/listinfo/gramps-users
> https://gramps-project.org


--
Gramps-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gramps-users
https://gramps-project.org
Reply | Threaded
Open this post in threaded view
|

Re: Win10 install, missing fonts or translation?

prculley
Phil;
Two changes were made regarding the Windows font handling for 5.0.2.  The first is that the Linux style font handler (fontconfig) was made the standard font handler.  The second change was to move the cache for that handler to AppData/Local/fontconfig.

Did you have some version of Gramps previously working properly?  What version?
If so, did you have "Edit/Preferences/Use alternate font handler" turned on?
Does your c:\Users\you\AppData\Local\fontconfig\cache directory have CACHEDIR.TAG and "something-x86_64.cache-7" files present? Any others?
If you run the command "c:\Program Files\GrampsAIO64-5.0.2\fc-cache -f -v" from the command line, do you see something like the following?
C:\ Program Files\GrampsAIO64-5.0.2\fc-cache -f -v
C:/WINDOWS/fonts: caching, new cache contents: 1132 fonts, 0 dirs
C:/Users/prc/.local/share/fonts: skipping, no such directory
C:/Users/prc/.fonts: skipping, no such directory
C:/Users/prc/AppData/Local/fontconfig/cache: cleaning cache directory
C:/Users/prc/.cache/fontconfig: not cleaning non-existent cache directory
C:/Users/prc/.fontconfig: not cleaning non-existent cache directory
fc-cache: succeeded

Note that the above command can take a while to complete, just like it does during the Gramps install.
Let me know if Gramps starts normally after running the command.

If none of this works, you can also try adding the PANGOCAIRO_BACKEND environment variable with the value of 'win32'.  That should allow Gramps to start using the other font handler.

I'd appreciate feedback on this, if there is some unexpected issue with the AIO we will need to deal with it as soon as possible.

Paul C.

On Wed, Aug 14, 2019 at 8:01 AM Dave Scheipers <[hidden email]> wrote:
Hi Phil

In your Theme, what are you using as your font? When exploring the
theme, I noticed that some fonts did not render correctly.

Dave

On Wed, Aug 14, 2019 at 8:22 AM Phil Miesle <[hidden email]> wrote:
>
> Hello,
>
> I have installed GrampsAIO64-5.0.2 into Windows 10 using the downloaded installer, and when I launch the program I end up with no letters, only square boxes, no letters apart from the Windows title bar:
>
> https://drive.google.com/open?id=0BwzY2jGxsoM0NVhuWlViRXhnLXRmVVRHMXlkLWpibTBEUVdB
>
> I've tried installing a new theme but that hasn't worked, which leads me to believe I have a language problem?  My language is normally "English (Ireland)" but I have switched it to "English (United States)" to no avail...
>
> Am I missing something?
> --
> Gramps-users mailing list
> [hidden email]
> https://lists.sourceforge.net/lists/listinfo/gramps-users
> https://gramps-project.org


--
Gramps-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gramps-users
https://gramps-project.org


--
Gramps-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gramps-users
https://gramps-project.org
Reply | Threaded
Open this post in threaded view
|

Re: Win10 install, missing fonts or translation?

Phil Miesle
First time ever installing Gramps (but very excited to try it and possibly extend it - having the ability to support multi-user in a manner other than something Dropbox-like would be great).

Directory of C:\Users\Phil Miesle\AppData\Local\fontconfig\cache

14/08/2019  16:00    <DIR>          .
14/08/2019  16:00    <DIR>          ..
14/08/2019  16:00               204 CACHEDIR.TAG
30/07/2019  21:14         2,035,728 d031bbba323fd9e5b47e0ee5a0353f11-x86_64.cache-7
14/08/2019  16:00         1,985,648 d031bbba323fd9e5b47e0ee5a0353f11-x86_64.cache-7.NEW

C:\Users\Phil Miesle>"c:\Program Files\GrampsAIO64-5.0.2\fc-cache" -f -v
C:/Windows/fonts: caching, new cache contents: 1057 fonts, 0 dirs
C:/Users/Phil Miesle/.local/share/fonts: skipping, no such directory
C:/Users/Phil Miesle/.fonts: skipping, no such directory
C:/Users/Phil Miesle/AppData/Local/fontconfig/cache: cleaning cache directory
C:/Users/Phil Miesle/.cache/fontconfig: cleaning cache directory
C:/Users/Phil Miesle/.fontconfig: not cleaning non-existent cache directory
c:\Program Files\GrampsAIO64-5.0.2\fc-cache: succeeded
 
After running this, the problem remains unresolved. 

Setting PANGOCAIRO_BACKEND=win32 in my User Environment Variables has resolved the problem.

Happy to continue to help troubleshooting this, note that I have Windows Subsystem for Linux installed and I'm beginning to wonder if that may be part of the difficulty?


On Wed, 14 Aug 2019 at 15:57, Paul Culley <[hidden email]> wrote:
Phil;
Two changes were made regarding the Windows font handling for 5.0.2.  The first is that the Linux style font handler (fontconfig) was made the standard font handler.  The second change was to move the cache for that handler to AppData/Local/fontconfig.

Did you have some version of Gramps previously working properly?  What version?
If so, did you have "Edit/Preferences/Use alternate font handler" turned on?
Does your c:\Users\you\AppData\Local\fontconfig\cache directory have CACHEDIR.TAG and "something-x86_64.cache-7" files present? Any others?
If you run the command "c:\Program Files\GrampsAIO64-5.0.2\fc-cache -f -v" from the command line, do you see something like the following?
C:\ Program Files\GrampsAIO64-5.0.2\fc-cache -f -v
C:/WINDOWS/fonts: caching, new cache contents: 1132 fonts, 0 dirs
C:/Users/prc/.local/share/fonts: skipping, no such directory
C:/Users/prc/.fonts: skipping, no such directory
C:/Users/prc/AppData/Local/fontconfig/cache: cleaning cache directory
C:/Users/prc/.cache/fontconfig: not cleaning non-existent cache directory
C:/Users/prc/.fontconfig: not cleaning non-existent cache directory
fc-cache: succeeded

Note that the above command can take a while to complete, just like it does during the Gramps install.
Let me know if Gramps starts normally after running the command.

If none of this works, you can also try adding the PANGOCAIRO_BACKEND environment variable with the value of 'win32'.  That should allow Gramps to start using the other font handler.

I'd appreciate feedback on this, if there is some unexpected issue with the AIO we will need to deal with it as soon as possible.

Paul C.

On Wed, Aug 14, 2019 at 8:01 AM Dave Scheipers <[hidden email]> wrote:
Hi Phil

In your Theme, what are you using as your font? When exploring the
theme, I noticed that some fonts did not render correctly.

Dave

On Wed, Aug 14, 2019 at 8:22 AM Phil Miesle <[hidden email]> wrote:
>
> Hello,
>
> I have installed GrampsAIO64-5.0.2 into Windows 10 using the downloaded installer, and when I launch the program I end up with no letters, only square boxes, no letters apart from the Windows title bar:
>
> https://drive.google.com/open?id=0BwzY2jGxsoM0NVhuWlViRXhnLXRmVVRHMXlkLWpibTBEUVdB
>
> I've tried installing a new theme but that hasn't worked, which leads me to believe I have a language problem?  My language is normally "English (Ireland)" but I have switched it to "English (United States)" to no avail...
>
> Am I missing something?
> --
> Gramps-users mailing list
> [hidden email]
> https://lists.sourceforge.net/lists/listinfo/gramps-users
> https://gramps-project.org


--
Gramps-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gramps-users
https://gramps-project.org


--
Gramps-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gramps-users
https://gramps-project.org
Reply | Threaded
Open this post in threaded view
|

Re: Win10 install, missing fonts or translation?

Phil Miesle
So  fc-cache -V does not work without passing the directory. 

I removed environment variable, renamed the "cache" directory to "cache.old" and then startup took quite a bit longer than usual ("long time" is relative, I have a beast of a machine).  Spent a lot of time on disk I/O and ended up with a working app, as well as a new "cache" directory.

phil@WORK-1:/mnt/c/Users/Phil Miesle/AppData/Local/fontconfig$ ls -l cache
total 1940
-rwxrwxrwx 1 phil phil     204 Aug 16 01:39 CACHEDIR.TAG
-rwxrwxrwx 1 phil phil 1985648 Aug 16 01:39 d031bbba323fd9e5b47e0ee5a0353f11-x86_64.cache-7

phil@WORK-1:/mnt/c/Users/Phil Miesle/AppData/Local/fontconfig$ ls -l cache.old/
total 3932
-rwxrwxrwx 1 phil phil     204 Aug 14 16:00 CACHEDIR.TAG
-rwxrwxrwx 1 phil phil 2035728 Jul 30 21:14 d031bbba323fd9e5b47e0ee5a0353f11-x86_64.cache-7
-rwxrwxrwx 1 phil phil 1985648 Aug 14 16:00 d031bbba323fd9e5b47e0ee5a0353f11-x86_64.cache-7.NEW

phil@WORK-1:/mnt/c/Users/Phil Miesle/AppData/Local/fontconfig$ md5sum cache/d031bbba323fd9e5b47e0ee5a0353f11-x86_64.cache-7
b85a40d91b79ac30a9bbf4b810426b76  cache/d031bbba323fd9e5b47e0ee5a0353f11-x86_64.cache-7

phil@WORK-1:/mnt/c/Users/Phil Miesle/AppData/Local/fontconfig$ md5sum cache.old/d031bbba323fd9e5b47e0ee5a0353f11-x86_64.cache-7.NEW
b85a40d91b79ac30a9bbf4b810426b76  cache.old/d031bbba323fd9e5b47e0ee5a0353f11-x86_64.cache-7.NEW


phil@WORK-1:/mnt/c/Users/Phil Miesle/AppData/Local/fontconfig$ md5sum cache.old/d031bbba323fd9e5b47e0ee5a0353f11-x86_64.cache-7
faf3856210a6201c57a3d705202c0727  cache.old/d031bbba323fd9e5b47e0ee5a0353f11-x86_64.cache-7

So what it looks to me is that there is this ".NEW" file that should really be the cache file, but something happened during installation?

Looks like the workaround might be to remove the ".cache-7" file, and remove the ".NEW" extension from the remaining file; alternately remove the "cache" directory and restart (expecting a long delay in restart time).



On Wed, 14 Aug 2019 at 17:17, Paul Culley <[hidden email]> wrote:
Since the 'win32' setting (which doesn't use fontconfig) made Gramps work, it seems clear that there is a fontconfig issue.  If you have the Windows Subsystem for Linux installed it seems possible that there might be some sort of conflict.  At the moment, I don't have a good way to test this here.

It might be interesting to see if there is more than one fontconfig on your system.  If you try 'fc-cache -V' at a command prompt somewhere away from the Gramps install location, do you get a result?  Is it different than if you do the " c:\Program Files\GrampsAIO64-5.0.2\fc-cache -V"?

If you rename or remove "C:\Users\Phil Miesle\AppData\Local\fontconfig\cache" and remove or rename the PANGOCAIRO_BACKEND so Gramps starts with fontconfig, does it take a long time and rebuild the "C:\Users\Phil Miesle\AppData\Local\fontconfig\cache" again when Gramps starts?  This would give an indication of whether Gramps is using that cache at all.

As you can guess I'm grasping for ideas here...

The Gramps AIO is built on the MSYS2/MINGW64 library system which is not compatible with the Microsoft tools.  The MINGW system tries to use Linux libraries built on that system so that applications think they are still running on Linux.  I don't have any idea if this any be causing confusion with the WSL stuff.  I heard that one user was able to get Gramps running on WSL after installing an X server, in case you want to play with that.

Without a working fontconfig, Gramps may have trouble with some unusual utf8 characters, like far eastern ones, depending on your installed Windows fonts.  Also there may be an issue with creating Graphs.  Which is the reason we switched to making this the default.

P.S. Gramps does NOT support multiple users working on the same tree at the same time, trying this will irrecoverably destroy a BSDDB database, and will not be good for the alternate SqLite backend database.

Paul C.

On Wed, Aug 14, 2019 at 10:08 AM Phil Miesle <[hidden email]> wrote:
First time ever installing Gramps (but very excited to try it and possibly extend it - having the ability to support multi-user in a manner other than something Dropbox-like would be great).

Directory of C:\Users\Phil Miesle\AppData\Local\fontconfig\cache

14/08/2019  16:00    <DIR>          .
14/08/2019  16:00    <DIR>          ..
14/08/2019  16:00               204 CACHEDIR.TAG
30/07/2019  21:14         2,035,728 d031bbba323fd9e5b47e0ee5a0353f11-x86_64.cache-7
14/08/2019  16:00         1,985,648 d031bbba323fd9e5b47e0ee5a0353f11-x86_64.cache-7.NEW

C:\Users\Phil Miesle>"c:\Program Files\GrampsAIO64-5.0.2\fc-cache" -f -v
C:/Windows/fonts: caching, new cache contents: 1057 fonts, 0 dirs
C:/Users/Phil Miesle/.local/share/fonts: skipping, no such directory
C:/Users/Phil Miesle/.fonts: skipping, no such directory
C:/Users/Phil Miesle/AppData/Local/fontconfig/cache: cleaning cache directory
C:/Users/Phil Miesle/.cache/fontconfig: cleaning cache directory
C:/Users/Phil Miesle/.fontconfig: not cleaning non-existent cache directory
c:\Program Files\GrampsAIO64-5.0.2\fc-cache: succeeded
 
After running this, the problem remains unresolved. 

Setting PANGOCAIRO_BACKEND=win32 in my User Environment Variables has resolved the problem.

Happy to continue to help troubleshooting this, note that I have Windows Subsystem for Linux installed and I'm beginning to wonder if that may be part of the difficulty?


On Wed, 14 Aug 2019 at 15:57, Paul Culley <[hidden email]> wrote:
Phil;
Two changes were made regarding the Windows font handling for 5.0.2.  The first is that the Linux style font handler (fontconfig) was made the standard font handler.  The second change was to move the cache for that handler to AppData/Local/fontconfig.

Did you have some version of Gramps previously working properly?  What version?
If so, did you have "Edit/Preferences/Use alternate font handler" turned on?
Does your c:\Users\you\AppData\Local\fontconfig\cache directory have CACHEDIR.TAG and "something-x86_64.cache-7" files present? Any others?
If you run the command "c:\Program Files\GrampsAIO64-5.0.2\fc-cache -f -v" from the command line, do you see something like the following?
C:\ Program Files\GrampsAIO64-5.0.2\fc-cache -f -v
C:/WINDOWS/fonts: caching, new cache contents: 1132 fonts, 0 dirs
C:/Users/prc/.local/share/fonts: skipping, no such directory
C:/Users/prc/.fonts: skipping, no such directory
C:/Users/prc/AppData/Local/fontconfig/cache: cleaning cache directory
C:/Users/prc/.cache/fontconfig: not cleaning non-existent cache directory
C:/Users/prc/.fontconfig: not cleaning non-existent cache directory
fc-cache: succeeded

Note that the above command can take a while to complete, just like it does during the Gramps install.
Let me know if Gramps starts normally after running the command.

If none of this works, you can also try adding the PANGOCAIRO_BACKEND environment variable with the value of 'win32'.  That should allow Gramps to start using the other font handler.

I'd appreciate feedback on this, if there is some unexpected issue with the AIO we will need to deal with it as soon as possible.

Paul C.

On Wed, Aug 14, 2019 at 8:01 AM Dave Scheipers <[hidden email]> wrote:
Hi Phil

In your Theme, what are you using as your font? When exploring the
theme, I noticed that some fonts did not render correctly.

Dave

On Wed, Aug 14, 2019 at 8:22 AM Phil Miesle <[hidden email]> wrote:
>
> Hello,
>
> I have installed GrampsAIO64-5.0.2 into Windows 10 using the downloaded installer, and when I launch the program I end up with no letters, only square boxes, no letters apart from the Windows title bar:
>
> https://drive.google.com/open?id=0BwzY2jGxsoM0NVhuWlViRXhnLXRmVVRHMXlkLWpibTBEUVdB
>
> I've tried installing a new theme but that hasn't worked, which leads me to believe I have a language problem?  My language is normally "English (Ireland)" but I have switched it to "English (United States)" to no avail...
>
> Am I missing something?
> --
> Gramps-users mailing list
> [hidden email]
> https://lists.sourceforge.net/lists/listinfo/gramps-users
> https://gramps-project.org


--
Gramps-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gramps-users
https://gramps-project.org


--
Gramps-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gramps-users
https://gramps-project.org
Reply | Threaded
Open this post in threaded view
|

Re: Win10 install, missing fonts or translation?

prculley
Thanks for the assistance on debugging this.

After looking around the internet a bit it looks like there has been a persistent issue with fontconfig on Windows where something is interfering with fontconfig when it updates its cache.  As you noticed the updated cache has a .NEW extension, and it is supposed to end up replacing the .cache-7 file, but it doesn't work sometimes.  Depending on how bad the cache is out of date, users get anything from total failure to long delays every time the application starts as fontconfig tries to scan for and update its cache.

There was a patch applied back in 2017 that was supposed to fix this, but I guess it did not get all the cases.  If I had to guess, I would suspect that this is another case of Windows not deleting a file immediately when requested, so that when the .NEW file tries to get renamed, it fails as the original file is still there.  Or maybe something else, like an issue with virus scanners holding the cache-7 file open at the wrong time.

I suppose I could have the Gramps installer always delete the cache and then rebuild it, that would probably avoid your issue.  Of course if your system had any other applications that used fontconfig, with the same cache location, there is a potential that there could be a conflict.
Will have to think about this...

Paul C.

On Thu, Aug 15, 2019 at 7:52 PM Phil Miesle <[hidden email]> wrote:
So  fc-cache -V does not work without passing the directory. 

I removed environment variable, renamed the "cache" directory to "cache.old" and then startup took quite a bit longer than usual ("long time" is relative, I have a beast of a machine).  Spent a lot of time on disk I/O and ended up with a working app, as well as a new "cache" directory.

phil@WORK-1:/mnt/c/Users/Phil Miesle/AppData/Local/fontconfig$ ls -l cache
total 1940
-rwxrwxrwx 1 phil phil     204 Aug 16 01:39 CACHEDIR.TAG
-rwxrwxrwx 1 phil phil 1985648 Aug 16 01:39 d031bbba323fd9e5b47e0ee5a0353f11-x86_64.cache-7

phil@WORK-1:/mnt/c/Users/Phil Miesle/AppData/Local/fontconfig$ ls -l cache.old/
total 3932
-rwxrwxrwx 1 phil phil     204 Aug 14 16:00 CACHEDIR.TAG
-rwxrwxrwx 1 phil phil 2035728 Jul 30 21:14 d031bbba323fd9e5b47e0ee5a0353f11-x86_64.cache-7
-rwxrwxrwx 1 phil phil 1985648 Aug 14 16:00 d031bbba323fd9e5b47e0ee5a0353f11-x86_64.cache-7.NEW

phil@WORK-1:/mnt/c/Users/Phil Miesle/AppData/Local/fontconfig$ md5sum cache/d031bbba323fd9e5b47e0ee5a0353f11-x86_64.cache-7
b85a40d91b79ac30a9bbf4b810426b76  cache/d031bbba323fd9e5b47e0ee5a0353f11-x86_64.cache-7

phil@WORK-1:/mnt/c/Users/Phil Miesle/AppData/Local/fontconfig$ md5sum cache.old/d031bbba323fd9e5b47e0ee5a0353f11-x86_64.cache-7.NEW
b85a40d91b79ac30a9bbf4b810426b76  cache.old/d031bbba323fd9e5b47e0ee5a0353f11-x86_64.cache-7.NEW


phil@WORK-1:/mnt/c/Users/Phil Miesle/AppData/Local/fontconfig$ md5sum cache.old/d031bbba323fd9e5b47e0ee5a0353f11-x86_64.cache-7
faf3856210a6201c57a3d705202c0727  cache.old/d031bbba323fd9e5b47e0ee5a0353f11-x86_64.cache-7

So what it looks to me is that there is this ".NEW" file that should really be the cache file, but something happened during installation?

Looks like the workaround might be to remove the ".cache-7" file, and remove the ".NEW" extension from the remaining file; alternately remove the "cache" directory and restart (expecting a long delay in restart time).



On Wed, 14 Aug 2019 at 17:17, Paul Culley <[hidden email]> wrote:
Since the 'win32' setting (which doesn't use fontconfig) made Gramps work, it seems clear that there is a fontconfig issue.  If you have the Windows Subsystem for Linux installed it seems possible that there might be some sort of conflict.  At the moment, I don't have a good way to test this here.

It might be interesting to see if there is more than one fontconfig on your system.  If you try 'fc-cache -V' at a command prompt somewhere away from the Gramps install location, do you get a result?  Is it different than if you do the " c:\Program Files\GrampsAIO64-5.0.2\fc-cache -V"?

If you rename or remove "C:\Users\Phil Miesle\AppData\Local\fontconfig\cache" and remove or rename the PANGOCAIRO_BACKEND so Gramps starts with fontconfig, does it take a long time and rebuild the "C:\Users\Phil Miesle\AppData\Local\fontconfig\cache" again when Gramps starts?  This would give an indication of whether Gramps is using that cache at all.

As you can guess I'm grasping for ideas here...

The Gramps AIO is built on the MSYS2/MINGW64 library system which is not compatible with the Microsoft tools.  The MINGW system tries to use Linux libraries built on that system so that applications think they are still running on Linux.  I don't have any idea if this any be causing confusion with the WSL stuff.  I heard that one user was able to get Gramps running on WSL after installing an X server, in case you want to play with that.

Without a working fontconfig, Gramps may have trouble with some unusual utf8 characters, like far eastern ones, depending on your installed Windows fonts.  Also there may be an issue with creating Graphs.  Which is the reason we switched to making this the default.

P.S. Gramps does NOT support multiple users working on the same tree at the same time, trying this will irrecoverably destroy a BSDDB database, and will not be good for the alternate SqLite backend database.

Paul C.

On Wed, Aug 14, 2019 at 10:08 AM Phil Miesle <[hidden email]> wrote:
First time ever installing Gramps (but very excited to try it and possibly extend it - having the ability to support multi-user in a manner other than something Dropbox-like would be great).

Directory of C:\Users\Phil Miesle\AppData\Local\fontconfig\cache

14/08/2019  16:00    <DIR>          .
14/08/2019  16:00    <DIR>          ..
14/08/2019  16:00               204 CACHEDIR.TAG
30/07/2019  21:14         2,035,728 d031bbba323fd9e5b47e0ee5a0353f11-x86_64.cache-7
14/08/2019  16:00         1,985,648 d031bbba323fd9e5b47e0ee5a0353f11-x86_64.cache-7.NEW

C:\Users\Phil Miesle>"c:\Program Files\GrampsAIO64-5.0.2\fc-cache" -f -v
C:/Windows/fonts: caching, new cache contents: 1057 fonts, 0 dirs
C:/Users/Phil Miesle/.local/share/fonts: skipping, no such directory
C:/Users/Phil Miesle/.fonts: skipping, no such directory
C:/Users/Phil Miesle/AppData/Local/fontconfig/cache: cleaning cache directory
C:/Users/Phil Miesle/.cache/fontconfig: cleaning cache directory
C:/Users/Phil Miesle/.fontconfig: not cleaning non-existent cache directory
c:\Program Files\GrampsAIO64-5.0.2\fc-cache: succeeded
 
After running this, the problem remains unresolved. 

Setting PANGOCAIRO_BACKEND=win32 in my User Environment Variables has resolved the problem.

Happy to continue to help troubleshooting this, note that I have Windows Subsystem for Linux installed and I'm beginning to wonder if that may be part of the difficulty?


On Wed, 14 Aug 2019 at 15:57, Paul Culley <[hidden email]> wrote:
Phil;
Two changes were made regarding the Windows font handling for 5.0.2.  The first is that the Linux style font handler (fontconfig) was made the standard font handler.  The second change was to move the cache for that handler to AppData/Local/fontconfig.

Did you have some version of Gramps previously working properly?  What version?
If so, did you have "Edit/Preferences/Use alternate font handler" turned on?
Does your c:\Users\you\AppData\Local\fontconfig\cache directory have CACHEDIR.TAG and "something-x86_64.cache-7" files present? Any others?
If you run the command "c:\Program Files\GrampsAIO64-5.0.2\fc-cache -f -v" from the command line, do you see something like the following?
C:\ Program Files\GrampsAIO64-5.0.2\fc-cache -f -v
C:/WINDOWS/fonts: caching, new cache contents: 1132 fonts, 0 dirs
C:/Users/prc/.local/share/fonts: skipping, no such directory
C:/Users/prc/.fonts: skipping, no such directory
C:/Users/prc/AppData/Local/fontconfig/cache: cleaning cache directory
C:/Users/prc/.cache/fontconfig: not cleaning non-existent cache directory
C:/Users/prc/.fontconfig: not cleaning non-existent cache directory
fc-cache: succeeded

Note that the above command can take a while to complete, just like it does during the Gramps install.
Let me know if Gramps starts normally after running the command.

If none of this works, you can also try adding the PANGOCAIRO_BACKEND environment variable with the value of 'win32'.  That should allow Gramps to start using the other font handler.

I'd appreciate feedback on this, if there is some unexpected issue with the AIO we will need to deal with it as soon as possible.

Paul C.

On Wed, Aug 14, 2019 at 8:01 AM Dave Scheipers <[hidden email]> wrote:
Hi Phil

In your Theme, what are you using as your font? When exploring the
theme, I noticed that some fonts did not render correctly.

Dave

On Wed, Aug 14, 2019 at 8:22 AM Phil Miesle <[hidden email]> wrote:
>
> Hello,
>
> I have installed GrampsAIO64-5.0.2 into Windows 10 using the downloaded installer, and when I launch the program I end up with no letters, only square boxes, no letters apart from the Windows title bar:
>
> https://drive.google.com/open?id=0BwzY2jGxsoM0NVhuWlViRXhnLXRmVVRHMXlkLWpibTBEUVdB
>
> I've tried installing a new theme but that hasn't worked, which leads me to believe I have a language problem?  My language is normally "English (Ireland)" but I have switched it to "English (United States)" to no avail...
>
> Am I missing something?
> --
> Gramps-users mailing list
> [hidden email]
> https://lists.sourceforge.net/lists/listinfo/gramps-users
> https://gramps-project.org


--
Gramps-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gramps-users
https://gramps-project.org


--
Gramps-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gramps-users
https://gramps-project.org
Reply | Threaded
Open this post in threaded view
|

Re: Win10 install, missing fonts or translati

GRAMPS - User mailing list
Thanks Paul,

Your reply about potential slow Gramps startup due to corrupted font cache sets me to wondering...
how long should Gramps take to Start on Windows? 

Perhaps I have a corruption that needs to be addressed but I had just chalked the delay up to the fragmented nature of the huge number of Python modules.

Starting Gramps takes so long that I either play a game of Solitaire. My normal method is to simply leave it running as a background app. (In that case, Gramps becomes sluggish after a few days and requires Restarting. And the original timer-triggered backup was a real liability in terms of unnecessary disk space use.) 

When I was developing a highly modular application in a different environment, I put in a diagnostic switch that enabled logging timestamps at the start of each module. This allowed gauging the timesavings potential for optimizing each module.  It forced me to identify a number of inefficiencies inherent to functions of that environment... then write more efficient workarounds when the built-in functions were fatally flawed.

Does Gramps have a similar diagnostic mode? Perhaps in the console app?


-Brian

On Fri, Aug 16, 2019 at 9:18, Paul Culley
...
After looking around the internet a bit it looks like there has been a persistent issue with fontconfig on Windows where something is interfering with fontconfig when it updates its cache.  As you noticed the updated cache has a .NEW extension, and it is supposed to end up replacing the .cache-7 file, but it doesn't work sometimes.  Depending on how bad the cache is out of date, users get anything from total failure to long delays every time the application starts as fontconfig tries to scan for and update its cache.
...


--
Gramps-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gramps-users
https://gramps-project.org

Untitled Download Attachment
Untitled (206 bytes) Download Attachment
Reply | Threaded
Open this post in threaded view
|

Re: Win10 install, missing fonts or translati

Nick Hall
On 16/08/2019 16:02, Emyoulation--- via Gramps-users wrote:
When I was developing a highly modular application in a different environment, I put in a diagnostic switch that enabled logging timestamps at the start of each module. This allowed gauging the timesavings potential for optimizing each module.  It forced me to identify a number of inefficiencies inherent to functions of that environment... then write more efficient workarounds when the built-in functions were fatally flawed.

Does Gramps have a similar diagnostic mode? Perhaps in the console app?

Yes.  Some modules contain debug statements.  See:

https://gramps-project.org/wiki/index.php/Debugging_Gramps

You can also do some profiling described here:

https://gramps-project.org/wiki/index.php/GEPS_016:_Enhancing_Gramps_Processing_Speed#Profiling

Let us know if you need any help.  Ideally Gramps should start within a few seconds.


Nick.




--
Gramps-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gramps-users
https://gramps-project.org
Reply | Threaded
Open this post in threaded view
|

Re: Win10 install, missing fonts or translati

Phil Miesle
As a reference point, with my laptop (6 core 8th gen Intel, 32GB memory, and M.2 NVMe), after removing the cache directory startup took about 90 seconds; that is a *long* time on a machine of this spec. Rebuilding cache didn’t change the “normal” start time which is about 12 seconds.

Clearly some room for optimisation, I should imagine on an older computer with a mechanical hard drive this could be long enough to cook a roast dinner :/

Not being sure how “normal” this all is, I shall have a look at profiling and may spin up another thread on that topic.

For now, this thread relates to cache corruption which looks to have a viable work-around; if this doesn’t work it would be great to learn that! 

And hopefully a cleaner solution can be devised :) 


On Fri, 16 Aug 2019 at 17:22, Nick Hall <[hidden email]> wrote:
On 16/08/2019 16:02, Emyoulation--- via Gramps-users wrote:
When I was developing a highly modular application in a different environment, I put in a diagnostic switch that enabled logging timestamps at the start of each module. This allowed gauging the timesavings potential for optimizing each module.  It forced me to identify a number of inefficiencies inherent to functions of that environment... then write more efficient workarounds when the built-in functions were fatally flawed.

Does Gramps have a similar diagnostic mode? Perhaps in the console app?

Yes.  Some modules contain debug statements.  See:

https://gramps-project.org/wiki/index.php/Debugging_Gramps

You can also do some profiling described here:

https://gramps-project.org/wiki/index.php/GEPS_016:_Enhancing_Gramps_Processing_Speed#Profiling

Let us know if you need any help.  Ideally Gramps should start within a few seconds.


Nick.


--
Gramps-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gramps-users
https://gramps-project.org


--
Gramps-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gramps-users
https://gramps-project.org
Reply | Threaded
Open this post in threaded view
|

Re: Win10 install, missing fonts or translati

prculley
If you have ALL the addons installed, it can take longer to start. Since I normally am running in my debugger (about 1/4 normal speed), this can get tedious, so I normally only have the addons installed that I need (one of the reasons I developed the Enhanced Addon Manager).

Running at full speed it usually takes less than 5 seconds to get to the Family Tree Manager dialog.

Loading the db can be anything from <1 sec to very long depending on the db size.

Paul C.

On Fri, Aug 16, 2019 at 1:39 PM Phil Miesle <[hidden email]> wrote:
As a reference point, with my laptop (6 core 8th gen Intel, 32GB memory, and M.2 NVMe), after removing the cache directory startup took about 90 seconds; that is a *long* time on a machine of this spec. Rebuilding cache didn’t change the “normal” start time which is about 12 seconds.

Clearly some room for optimisation, I should imagine on an older computer with a mechanical hard drive this could be long enough to cook a roast dinner :/

Not being sure how “normal” this all is, I shall have a look at profiling and may spin up another thread on that topic.

For now, this thread relates to cache corruption which looks to have a viable work-around; if this doesn’t work it would be great to learn that! 

And hopefully a cleaner solution can be devised :) 


On Fri, 16 Aug 2019 at 17:22, Nick Hall <[hidden email]> wrote:
On 16/08/2019 16:02, Emyoulation--- via Gramps-users wrote:
When I was developing a highly modular application in a different environment, I put in a diagnostic switch that enabled logging timestamps at the start of each module. This allowed gauging the timesavings potential for optimizing each module.  It forced me to identify a number of inefficiencies inherent to functions of that environment... then write more efficient workarounds when the built-in functions were fatally flawed.

Does Gramps have a similar diagnostic mode? Perhaps in the console app?

Yes.  Some modules contain debug statements.  See:

https://gramps-project.org/wiki/index.php/Debugging_Gramps

You can also do some profiling described here:

https://gramps-project.org/wiki/index.php/GEPS_016:_Enhancing_Gramps_Processing_Speed#Profiling

Let us know if you need any help.  Ideally Gramps should start within a few seconds.


Nick.


--
Gramps-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gramps-users
https://gramps-project.org
--
Gramps-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gramps-users
https://gramps-project.org


--
Gramps-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gramps-users
https://gramps-project.org
Reply | Threaded
Open this post in threaded view
|

Re: Win10 install, missing fonts or translati

Ron Johnson

256GB SSD
Xubuntu 18.04
Gramps 5.0.2
SQLite
9,729 people
3,173 families
18,392 events
3,519 places
2,128 citations
794 Media objects
2,297 Notes

3 seconds to load from the Family Trees window.

On 8/17/19 9:48 AM, Paul Culley wrote:
If you have ALL the addons installed, it can take longer to start. Since I normally am running in my debugger (about 1/4 normal speed), this can get tedious, so I normally only have the addons installed that I need (one of the reasons I developed the Enhanced Addon Manager).

Running at full speed it usually takes less than 5 seconds to get to the Family Tree Manager dialog.

Loading the db can be anything from <1 sec to very long depending on the db size.

Paul C.

On Fri, Aug 16, 2019 at 1:39 PM Phil Miesle <[hidden email]> wrote:
As a reference point, with my laptop (6 core 8th gen Intel, 32GB memory, and M.2 NVMe), after removing the cache directory startup took about 90 seconds; that is a *long* time on a machine of this spec. Rebuilding cache didn’t change the “normal” start time which is about 12 seconds.

Clearly some room for optimisation, I should imagine on an older computer with a mechanical hard drive this could be long enough to cook a roast dinner :/

Not being sure how “normal” this all is, I shall have a look at profiling and may spin up another thread on that topic.

For now, this thread relates to cache corruption which looks to have a viable work-around; if this doesn’t work it would be great to learn that! 

And hopefully a cleaner solution can be devised :) 


On Fri, 16 Aug 2019 at 17:22, Nick Hall <[hidden email]> wrote:
On 16/08/2019 16:02, Emyoulation--- via Gramps-users wrote:
When I was developing a highly modular application in a different environment, I put in a diagnostic switch that enabled logging timestamps at the start of each module. This allowed gauging the timesavings potential for optimizing each module.  It forced me to identify a number of inefficiencies inherent to functions of that environment... then write more efficient workarounds when the built-in functions were fatally flawed.

Does Gramps have a similar diagnostic mode? Perhaps in the console app?

Yes.  Some modules contain debug statements.  See:

https://gramps-project.org/wiki/index.php/Debugging_Gramps

You can also do some profiling described here:

https://gramps-project.org/wiki/index.php/GEPS_016:_Enhancing_Gramps_Processing_Speed#Profiling

Let us know if you need any help.  Ideally Gramps should start within a few seconds.


Nick.


--
Angular momentum makes the world go 'round.


--
Gramps-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gramps-users
https://gramps-project.org
Reply | Threaded
Open this post in threaded view
|

Re: Win10 install, missing fonts or translati

Rich Lakey
3.4 seconds

 python    : 3.6.8
 gramps    : 5.0.1
 gtk++     : 3.22.30
 SQLite
 All addons
 LinuxMint 19 Cinnamon
 AMD Ryzen 5 2600X
 M.2 NVMe 500 GB SSD
16GB memory

13,787 individuals
6,076 notes
5,60 4 places
25,453 Events

I am happy w/ Performance
Rich


On 8/17/19 12:18 PM, Ron Johnson wrote:

256GB SSD
Xubuntu 18.04
Gramps 5.0.2
SQLite
9,729 people
3,173 families
18,392 events
3,519 places
2,128 citations
794 Media objects
2,297 Notes

3 seconds to load from the Family Trees window.

On 8/17/19 9:48 AM, Paul Culley wrote:
If you have ALL the addons installed, it can take longer to start. Since I normally am running in my debugger (about 1/4 normal speed), this can get tedious, so I normally only have the addons installed that I need (one of the reasons I developed the Enhanced Addon Manager).

Running at full speed it usually takes less than 5 seconds to get to the Family Tree Manager dialog.

Loading the db can be anything from <1 sec to very long depending on the db size.

Paul C.

On Fri, Aug 16, 2019 at 1:39 PM Phil Miesle <[hidden email]> wrote:
As a reference point, with my laptop (6 core 8th gen Intel, 32GB memory, and M.2 NVMe), after removing the cache directory startup took about 90 seconds; that is a *long* time on a machine of this spec. Rebuilding cache didn’t change the “normal” start time which is about 12 seconds.

Clearly some room for optimisation, I should imagine on an older computer with a mechanical hard drive this could be long enough to cook a roast dinner :/

Not being sure how “normal” this all is, I shall have a look at profiling and may spin up another thread on that topic.

For now, this thread relates to cache corruption which looks to have a viable work-around; if this doesn’t work it would be great to learn that! 

And hopefully a cleaner solution can be devised :) 


On Fri, 16 Aug 2019 at 17:22, Nick Hall <[hidden email]> wrote:
On 16/08/2019 16:02, Emyoulation--- via Gramps-users wrote:
When I was developing a highly modular application in a different environment, I put in a diagnostic switch that enabled logging timestamps at the start of each module. This allowed gauging the timesavings potential for optimizing each module.  It forced me to identify a number of inefficiencies inherent to functions of that environment... then write more efficient workarounds when the built-in functions were fatally flawed.

Does Gramps have a similar diagnostic mode? Perhaps in the console app?

Yes.  Some modules contain debug statements.  See:

https://gramps-project.org/wiki/index.php/Debugging_Gramps

You can also do some profiling described here:

https://gramps-project.org/wiki/index.php/GEPS_016:_Enhancing_Gramps_Processing_Speed#Profiling

Let us know if you need any help.  Ideally Gramps should start within a few seconds.


Nick.


--
Angular momentum makes the world go 'round.





--
Gramps-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gramps-users
https://gramps-project.org
Reply | Threaded
Open this post in threaded view
|

Re: Win10 install, missing fonts or translati

GRAMPS - User mailing list
In reply to this post by prculley
10 seconds on MacBook -  Mojave on 1.3 Ghz Core I5.  Gramps 5.01.  


On Aug 17, 2019, at 7:48 AM, Paul Culley <[hidden email]> wrote:

If you have ALL the addons installed, it can take longer to start. Since I normally am running in my debugger (about 1/4 normal speed), this can get tedious, so I normally only have the addons installed that I need (one of the reasons I developed the Enhanced Addon Manager).

Running at full speed it usually takes less than 5 seconds to get to the Family Tree Manager dialog.

Loading the db can be anything from <1 sec to very long depending on the db size.

Paul C.

On Fri, Aug 16, 2019 at 1:39 PM Phil Miesle <[hidden email]> wrote:
As a reference point, with my laptop (6 core 8th gen Intel, 32GB memory, and M.2 NVMe), after removing the cache directory startup took about 90 seconds; that is a *long* time on a machine of this spec. Rebuilding cache didn’t change the “normal” start time which is about 12 seconds.

Clearly some room for optimisation, I should imagine on an older computer with a mechanical hard drive this could be long enough to cook a roast dinner :/

Not being sure how “normal” this all is, I shall have a look at profiling and may spin up another thread on that topic.

For now, this thread relates to cache corruption which looks to have a viable work-around; if this doesn’t work it would be great to learn that! 

And hopefully a cleaner solution can be devised :) 


On Fri, 16 Aug 2019 at 17:22, Nick Hall <[hidden email]> wrote:
On 16/08/2019 16:02, Emyoulation--- via Gramps-users wrote:
When I was developing a highly modular application in a different environment, I put in a diagnostic switch that enabled logging timestamps at the start of each module. This allowed gauging the timesavings potential for optimizing each module.  It forced me to identify a number of inefficiencies inherent to functions of that environment... then write more efficient workarounds when the built-in functions were fatally flawed.

Does Gramps have a similar diagnostic mode? Perhaps in the console app?

Yes.  Some modules contain debug statements.  See:

https://gramps-project.org/wiki/index.php/Debugging_Gramps

You can also do some profiling described here:

https://gramps-project.org/wiki/index.php/GEPS_016:_Enhancing_Gramps_Processing_Speed#Profiling

Let us know if you need any help.  Ideally Gramps should start within a few seconds.


Nick.


--
Gramps-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gramps-users
https://gramps-project.org
--
Gramps-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gramps-users
https://gramps-project.org
--
Gramps-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gramps-users
https://gramps-project.org



--
Gramps-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gramps-users
https://gramps-project.org
Reply | Threaded
Open this post in threaded view
|

Re: Win10 install, missing fonts or translati

Ron Johnson

What kind of disk (rotational or SSD)?
How big of a tree?
What database back end?

On 8/17/19 9:17 PM, mikeh via Gramps-users wrote:
10 seconds on MacBook -  Mojave on 1.3 Ghz Core I5.  Gramps 5.01.  


On Aug 17, 2019, at 7:48 AM, Paul Culley <[hidden email]> wrote:

If you have ALL the addons installed, it can take longer to start. Since I normally am running in my debugger (about 1/4 normal speed), this can get tedious, so I normally only have the addons installed that I need (one of the reasons I developed the Enhanced Addon Manager).

Running at full speed it usually takes less than 5 seconds to get to the Family Tree Manager dialog.

Loading the db can be anything from <1 sec to very long depending on the db size.

Paul C.

On Fri, Aug 16, 2019 at 1:39 PM Phil Miesle <[hidden email]> wrote:
As a reference point, with my laptop (6 core 8th gen Intel, 32GB memory, and M.2 NVMe), after removing the cache directory startup took about 90 seconds; that is a *long* time on a machine of this spec. Rebuilding cache didn’t change the “normal” start time which is about 12 seconds.

Clearly some room for optimisation, I should imagine on an older computer with a mechanical hard drive this could be long enough to cook a roast dinner :/

Not being sure how “normal” this all is, I shall have a look at profiling and may spin up another thread on that topic.

For now, this thread relates to cache corruption which looks to have a viable work-around; if this doesn’t work it would be great to learn that! 

And hopefully a cleaner solution can be devised :) 


On Fri, 16 Aug 2019 at 17:22, Nick Hall <[hidden email]> wrote:
On 16/08/2019 16:02, Emyoulation--- via Gramps-users wrote:
When I was developing a highly modular application in a different environment, I put in a diagnostic switch that enabled logging timestamps at the start of each module. This allowed gauging the timesavings potential for optimizing each module.  It forced me to identify a number of inefficiencies inherent to functions of that environment... then write more efficient workarounds when the built-in functions were fatally flawed.

Does Gramps have a similar diagnostic mode? Perhaps in the console app?

Yes.  Some modules contain debug statements.  See:

https://gramps-project.org/wiki/index.php/Debugging_Gramps

You can also do some profiling described here:

https://gramps-project.org/wiki/index.php/GEPS_016:_Enhancing_Gramps_Processing_Speed#Profiling

Let us know if you need any help.  Ideally Gramps should start within a few seconds.



--
Angular momentum makes the world go 'round.


--
Gramps-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gramps-users
https://gramps-project.org
Reply | Threaded
Open this post in threaded view
|

Re: Win10 install, missing fonts or translati

GRAMPS - User mailing list
SSD
3,200 people
BSDDB

On Aug 17, 2019, at 7:51 PM, Ron Johnson <[hidden email]> wrote:


What kind of disk (rotational or SSD)?
How big of a tree?
What database back end?

On 8/17/19 9:17 PM, mikeh via Gramps-users wrote:
10 seconds on MacBook -  Mojave on 1.3 Ghz Core I5.  Gramps 5.01.  


On Aug 17, 2019, at 7:48 AM, Paul Culley <[hidden email]> wrote:

If you have ALL the addons installed, it can take longer to start. Since I normally am running in my debugger (about 1/4 normal speed), this can get tedious, so I normally only have the addons installed that I need (one of the reasons I developed the Enhanced Addon Manager).

Running at full speed it usually takes less than 5 seconds to get to the Family Tree Manager dialog.

Loading the db can be anything from <1 sec to very long depending on the db size.

Paul C.

On Fri, Aug 16, 2019 at 1:39 PM Phil Miesle <[hidden email]> wrote:
As a reference point, with my laptop (6 core 8th gen Intel, 32GB memory, and M.2 NVMe), after removing the cache directory startup took about 90 seconds; that is a *long* time on a machine of this spec. Rebuilding cache didn’t change the “normal” start time which is about 12 seconds.

Clearly some room for optimisation, I should imagine on an older computer with a mechanical hard drive this could be long enough to cook a roast dinner :/

Not being sure how “normal” this all is, I shall have a look at profiling and may spin up another thread on that topic.

For now, this thread relates to cache corruption which looks to have a viable work-around; if this doesn’t work it would be great to learn that! 

And hopefully a cleaner solution can be devised :) 


On Fri, 16 Aug 2019 at 17:22, Nick Hall <[hidden email]> wrote:
On 16/08/2019 16:02, Emyoulation--- via Gramps-users wrote:
When I was developing a highly modular application in a different environment, I put in a diagnostic switch that enabled logging timestamps at the start of each module. This allowed gauging the timesavings potential for optimizing each module.  It forced me to identify a number of inefficiencies inherent to functions of that environment... then write more efficient workarounds when the built-in functions were fatally flawed.

Does Gramps have a similar diagnostic mode? Perhaps in the console app?

Yes.  Some modules contain debug statements.  See:

https://gramps-project.org/wiki/index.php/Debugging_Gramps

You can also do some profiling described here:

https://gramps-project.org/wiki/index.php/GEPS_016:_Enhancing_Gramps_Processing_Speed#Profiling

Let us know if you need any help.  Ideally Gramps should start within a few seconds.



--
Angular momentum makes the world go 'round.
--
Gramps-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gramps-users
https://gramps-project.org



--
Gramps-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gramps-users
https://gramps-project.org
Reply | Threaded
Open this post in threaded view
|

Re: Win10 install, missing fonts or translati

Ron Johnson
It would be interesting to make copies of your tree using #1 the SQLite backend and #2 the BSDDB (freshly built without any fragmentation).

Then you'd know how much of it is due to the 1.3GHz CPU.

On 8/17/19 11:42 PM, mikeh via Gramps-users wrote:
SSD
3,200 people
BSDDB

On Aug 17, 2019, at 7:51 PM, Ron Johnson <[hidden email]> wrote:


What kind of disk (rotational or SSD)?
How big of a tree?
What database back end?

On 8/17/19 9:17 PM, mikeh via Gramps-users wrote:
10 seconds on MacBook -  Mojave on 1.3 Ghz Core I5.  Gramps 5.01.  



--
Angular momentum makes the world go 'round.


--
Gramps-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gramps-users
https://gramps-project.org
Reply | Threaded
Open this post in threaded view
|

Re: Win10 install, missing fonts or translati

GRAMPS - User mailing list
Startup times (both with fresh import from a GEDOM file):

Sqlite - 4 seconds
BSDDB - 10 seconds



On Aug 17, 2019, at 10:12 PM, Ron Johnson <[hidden email]> wrote:

It would be interesting to make copies of your tree using #1 the SQLite backend and #2 the BSDDB (freshly built without any fragmentation).

Then you'd know how much of it is due to the 1.3GHz CPU.

On 8/17/19 11:42 PM, mikeh via Gramps-users wrote:
SSD
3,200 people
BSDDB

On Aug 17, 2019, at 7:51 PM, Ron Johnson <[hidden email]> wrote:


What kind of disk (rotational or SSD)?
How big of a tree?
What database back end?

On 8/17/19 9:17 PM, mikeh via Gramps-users wrote:
10 seconds on MacBook -  Mojave on 1.3 Ghz Core I5.  Gramps 5.01.  



--
Angular momentum makes the world go 'round.
--
Gramps-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gramps-users
https://gramps-project.org



--
Gramps-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gramps-users
https://gramps-project.org
Reply | Threaded
Open this post in threaded view
|

Re: Win10 install, missing fonts or translati

Ron Johnson

Interesting!

(If you decide to stay with the SQLite back end, restore from a .gramps file, because -- I think -- GEDCOM destroys all shared Events, Places, etc.)

On 8/18/19 12:40 AM, mikeh via Gramps-users wrote:
Startup times (both with fresh import from a GEDOM file):

Sqlite - 4 seconds
BSDDB - 10 seconds



On Aug 17, 2019, at 10:12 PM, Ron Johnson <[hidden email]> wrote:

It would be interesting to make copies of your tree using #1 the SQLite backend and #2 the BSDDB (freshly built without any fragmentation).

Then you'd know how much of it is due to the 1.3GHz CPU.

On 8/17/19 11:42 PM, mikeh via Gramps-users wrote:
SSD
3,200 people
BSDDB

On Aug 17, 2019, at 7:51 PM, Ron Johnson <[hidden email]> wrote:


What kind of disk (rotational or SSD)?
How big of a tree?
What database back end?

On 8/17/19 9:17 PM, mikeh via Gramps-users wrote:
10 seconds on MacBook -  Mojave on 1.3 Ghz Core I5.  Gramps 5.01.  


--
Angular momentum makes the world go 'round.


--
Gramps-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gramps-users
https://gramps-project.org
Reply | Threaded
Open this post in threaded view
|

Re: Win10 install, missing fonts or translati

enno
In reply to this post by Phil Miesle
Hello Phil,
> As a reference point, with my laptop (6 core 8th gen Intel, 32GB
> memory, and M.2 NVMe), after removing the cache directory startup took
> about 90 seconds; that is a *long* time on a machine of this spec.
> Rebuilding cache didn’t change the “normal” start time which is about
> 12 seconds.

I get less than 6 seconds on a laptop with a 2 core (4 with HT) i5, 8 GB
RAM, mechanical HDD, almost 12 000 people, BSDDB, Gramps 5.0.2, in
Windows 10.

And since I get similar results on Windows 10 running in VirtualBox,
inside Linux Mint, with 4 GB RAM, and 1 core of the host i7 allocated to
the VM, I think that your 12 seconds are pretty slow. This host runs on
a mechanical HDD too.

Sometimes, a virus scanner can slow down Gramps pretty much, because it
tries to scan all Python files that Gramps runs from. Which one do you use?

Regards,

Enno




--
Gramps-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gramps-users
https://gramps-project.org