Threaded Mode | Linear Mode

fhub · 07-28-2015, 08:13 PM

(07-27-2015 04:10 PM)Marcus von Cube Wrote: It will take some time until I can build everything. When I had the time SF was down... Now I'm a bit busy. Maybe another sleepless night will do.

Hi Marcus,

I see that you have already made a new build (3800), but the Windows emulator has been updated only for the 'complex branch' - the 'normal' WP34s (and also the WP31s) emulators are still the old versions.
Have you overlooked this or didn't you still have the time to build all versions?

Franz

Marcus von Cube · 07-29-2015, 09:31 PM

(07-28-2015 08:13 PM)fhub Wrote: The 'normal' WP34s (and also the WP31s) emulators are still the old versions.
Have you overlooked this or didn't you still have the time to build all versions?

I had some technical problems with my development infrastructure. This made me overlook the Windows builds. The next build will hopefully include everything.

fhub · 07-29-2015, 10:48 PM

(07-29-2015 09:31 PM)Marcus von Cube Wrote: I had some technical problems with my development infrastructure. This made me overlook the Windows builds. The next build will hopefully include everything.

Thanks Marcus!

Now almost everything is up-to-date again, only the Qt-emulator is still missing.
Maybe Sourceforge has 'forgotten' (during their last problems) that Pascal is a WP34s group member, and thus doesn't allow him to upload his new package!?

Franz

pascal_meheut · 07-30-2015, 06:09 AM

I do not know. A few days ago, I had the "upload" button but it did not work. Now it is gone and I found I cannot even upload files to another project I own.
I'll contact them.

BarryMead · (This post was last modified: 08-01-2015 03:25 AM by BarryMead.)

(07-30-2015 06:09 AM)pascal_meheut Wrote: I do not know. A few days ago, I had the "upload" button but it did not work. Now it is gone and I found I cannot even upload files to another project I own.
I'll contact them.

Is there anyone with write access to the wp-34s project that can re-enable Pascal's permissions so that he can update the emulator? As I understand it, Pascal is still LOCKED OUT of the project (still unable to upload files) since Source Forge went down.

pascal_meheut · 08-01-2015, 04:28 AM

No need. Sourceforge gave me access back and I've just uploaded the new versions.

**rprosperi** · 08-01-2015, 01:58 PM

(08-01-2015 04:28 AM)pascal_meheut Wrote: No need. Sourceforge gave me access back and I've just uploaded the new versions.

I just pulled and copy of the new build. Thanks for building and uploading Pascal.

If I use the VERS command, it shows "34S 3.3 [Ptr icon] 380"

I assume this is build 3800, so perhaps something is up with the display formatting?

fhub · 08-01-2015, 02:16 PM

(08-01-2015 01:58 PM)rprosperi Wrote: If I use the VERS command, it shows "34S 3.3 [Ptr icon] 380"

No problem here, it shows the correct build 3802 in both fonts ('Use real fonts' checked or unchecked).

Franz

pascal_meheut · 08-01-2015, 05:08 PM

(08-01-2015 01:58 PM)rprosperi Wrote: If I use the VERS command, it shows "34S 3.3 [Ptr icon] 380"

Same here: it displays 3802.
Which version are you running? Linux 32, 64, Windows, OSX?

**rprosperi** · 08-01-2015, 05:39 PM

(08-01-2015 05:08 PM)pascal_meheut Wrote: Same here: it displays 3802.
Which version are you running? Linux 32, 64, Windows, OSX?

I'm running the windows version on Win-7, x64.

With "Use real fonts" checked, I see the 380 as reported, however with that option unchecked, I see the full 3802. (thanks for the hint Franz)

I'm actually running this on a remote terminal session which could affect this I suppose. I will run locally and report results later today.

Thanks!

**rprosperi** · 08-01-2015, 06:34 PM

(08-01-2015 05:39 PM)rprosperi Wrote: I will run locally and report results later today.

Thanks!

Slight difference running locally on the PC (vs. on RDP terminal session on ThinkPad in the other room).

On the local PC, with "Use real fonts" checked and with the Medium V3 skin, the display shows "380_" vs. the expected "3802" With other skins, the full 3802 appears (except with the Modern V3 skin, it appears there is some other character beneath or on top of the final "2" - as if the 2 was printed on top of some other character (slash?).

fhub · 08-01-2015, 07:07 PM

(08-01-2015 05:39 PM)rprosperi Wrote: I'm running the windows version on Win-7, x64.

Have you changed the default DPI scaling (96dpi) in Windows to a higher value (e.g. 120dpi)?
This might be one possible reason that this version text is too big (with 120dpi) to fit on the WP34s alpha display!?

Franz

**rprosperi** · 08-01-2015, 09:08 PM

(08-01-2015 07:07 PM)fhub Wrote:
(08-01-2015 05:39 PM)rprosperi Wrote: I'm running the windows version on Win-7, x64.
Have you changed the default DPI scaling (96dpi) in Windows to a higher value (e.g. 120dpi)?
This might be one possible reason that this version text is too big (with 120dpi) to fit on the WP34s alpha display!?

Franz

No, still set to default of 96dpi. If using Windows, can you see the odd defect when using the Modern skin?

fhub · 08-01-2015, 10:08 PM

(08-01-2015 09:08 PM)rprosperi Wrote: If using Windows, can you see the odd defect when using the Modern skin?

Yes, strange.
Now I've also tried the 'Medium' skin (usually I use 'Compact'), and I can confirm the 380_ with this bigger skin.

I've now changed the internal version number with a hex-editor, and after using all possible digits it's clear now that only the digit 2 has this problem (and only with the big size in 'Medium' or 'Modern' skins). So I guess there's something wrong with the character 2 in the font file (at least for bigger sizes).

Franz

fhub · (This post was last modified: 08-01-2015 11:32 PM by fhub.)

(08-01-2015 10:08 PM)fhub Wrote: So I guess there's something wrong with the character 2 in the font file (at least for bigger sizes).

Well, there's indeed something wrong with the font DejaVuSans that is used for the alpha display in the Qt-emulator!
I've now searched on the internet for this font-file, and found a good version here:
http://www.fontsquirrel.com/fonts/dejavu-sans

Download this font (button 'DOWNLOAD OTF'), unzip DejaVuSans.ttf from this package and copy it over the file in the Qt-emulators 'fonts' folder.

With this font-file everything is displayed correctly (in every size, i.e. skin). Smile

Franz

**rprosperi** · (This post was last modified: 08-02-2015 01:36 PM by rprosperi.)

(08-01-2015 11:30 PM)fhub Wrote: With this font-file everything is displayed correctly (in every size, i.e. skin).

I can confirm Franz' solution works fine for all skins.

Nice detective work Franz! Thanks for that!

Indeed it appears the font bundled in the install pkg. is corrupt. Probably the happiest Pascal could be about a bug being reported/investigated/fixed all in a few hours. Smile

pascal_meheut · 08-02-2015, 07:28 AM

Yes, I like it when someone finds a solution while I'm asleep. Thanks for the help.
The font did not changed since 2013 and the problem occurs only on Windows with the Medium Skin... So this bug was a well hidden one.
I've uploaded new emulators with the better font fhub found.

fhub · (This post was last modified: 08-02-2015 03:51 PM by fhub.)

(08-02-2015 07:28 AM)pascal_meheut Wrote: I've uploaded new emulators with the better font fhub found.

Additional infos about this DejaVu font:

1) there's even a more recent version (2.35) on SourceForge (the official site of the font developers):
http://sourceforge.net/projects/dejavu/f...javu/2.35/
But since this bug with "2" was already fixed in the previous version, there's no need to update your package again - just as information for future updates.

2) and there's still one more small glitch with this DejaVuSans font:
if you use the option "Use real menus to display catalogs", all characters which go below the ground line are cut-off in this popup menu!
The worst example is the letter 'J' which is displayed in fact as 'I' (e.g. 'D->J' looks like 'D->I'), but also a few other characters like e.g. '"gjpqyQ".
It seems that the height for the entries in these popup menus is too small to fully display those characters.

I don't know if this can be fixed in the Qt-emulator, but at least I've found a solution for me:
I've used the font DejaVuSansMono from the full SourceForge font package, renamed it (just removing the 'Mono' from the name), and use this one instead of the original font - at least for 'J' vs. 'I' this mono-font is indeed looking better.

Franz

pascal_meheut · 08-03-2015, 09:59 AM

Thanks. I'll fix it when I'm back from holidays.

Claudio L. · 08-03-2015, 10:09 PM

(07-25-2015 12:02 AM)Paul Dale Wrote:
(07-24-2015 12:50 PM)Claudio L. Wrote: This is great! No rounding, no problems with negatives, handles any input (like 1h 65').
newRPL has found its HMS commands implementation.

Why won't the /150 and /1.5 steps cause rounding problems?
3 is not a factor of the base (10).

- Pauli

Sorry, I didn't mean rounding problems. I don't even know why I called it "rounding" but I was thinking of "carry propagation" actually. I meant the rounding issues like 0.5960, where you need to "round" the seconds (though it's not rounding what you are doing).
I meant it doesn't need any manual carry propagation, the formula takes care of everything, even if the input is invalid. It's a really good implementation.