For some reason /btnold for Aero::Apply didn't have any effect on my problem, so I'll just use /CANCELTEXT with inetc.Originally Posted by LoRd_MuldeR View PostI had missed that option. Thank you for pointing me at that 🙂
Aero plug-in
109 posts
There is no support for RTL languages at the moment. I will have to look at it when I have time - I have other priorities at the moment I'm afraid.
Stu
Stu
بسیارممنون Thank you🙂
New version:
Stu
1.0.0.9 - 2nd February 2013http://nsis.sourceforge.net/File:Aero.zip
* Fixed button and branding text rendering for right-to-left languages.
* Fixed bug which could cause a crash due to a variable not being initialized to NULL (only when Aero could not be enabled).
* Fixed Aero not being applied on classic UI installers with RTL languages.
Stu
Thank you for the update!
<nitpicking> Seems like you forgot to update the version info. </nitpicking>
<nitpicking> Seems like you forgot to update the version info. </nitpicking>
Fixed.
Stu
Stu
thank u so much (to fix it for RTL Lan)
New version:
Stu
1.0.1.0 - 30th November 2013
* Fixed ampersand handling on branding text.
Stu
New version:
Stu
1.0.1.1 - 24th April 2014http://nsis.sourceforge.net/File:Aero.zip
* Fixed bug where DWM rect was not invalidated on composition change.
* Added 64-bit builds.
Stu
Any chance to fix the font color in Windows 10 when window is inactive?

(Black font on black background is hard to read)

(Black font on black background is hard to read)
Yes,and me too see this problem in windows 10,mr afrow uk if is possible solve this problem...Originally Posted by LoRd_MuldeR View PostAny chance to fix the font color in Windows 10 when window is inactive?
(Black font on black background is hard to read)
Thanks
The Windows 10 issue has been resolved in the latest version.
Note:
Note:
The plug-in will detect Windows 10 if you build your installer using:Stu
* NSIS 3.0b2 or newer
* NSIS 3.0b1 with: ManifestSupportedOS {8e0f7a12-bfb3-4fe8-b9a5-48fd50a15a9a}
* NSIS 2.46 or older with: http://nsis.sourceforge.net/File:Packhdr.zip
Afrow UK, thanks for the new version!
However, it took me a long time to figure out why my Installer (made with NSIS 2.46) always failed to start with some strange "Side-by-side" error, after modifying the Manifest with the ResHacker tool. When I inspected the Manifests in ResHacker, it would show that the Manifest of the "working" installer and the one of the "failing" installer are identical, except for the additional Windows 10 compat entry. I even tried removing the compat entry entirely, so the Manifests really were 100% identical - still not working. Bummer !!!
In the end I figured out that the problem is actually the way how NSIS' built-in !appendfile command works! In NSIS 2.46, this command apparently inserts an UTF-8 BOM right at the beginning of the file. ResHacker doesn't show this at all (it silently ignores the BOM), so we cannot see the difference in the Manifest XML code. But with a Hex Editor we can see the BOM is there, even in the embedded Manifest in the final EXE. Original EXE didn't have this. And Windows doesn't like the BOM at the beginning of the XML Manifest at all! It simply refuses to run the installer. So I'm now using a "static" file, rather than a file generated by NSIS' !appendfile command, and all is fine again.
Maybe this info should be added somewhere...
Apart from that, the new version is now working nicely in Windows 10. Great work 👍
However, it took me a long time to figure out why my Installer (made with NSIS 2.46) always failed to start with some strange "Side-by-side" error, after modifying the Manifest with the ResHacker tool. When I inspected the Manifests in ResHacker, it would show that the Manifest of the "working" installer and the one of the "failing" installer are identical, except for the additional Windows 10 compat entry. I even tried removing the compat entry entirely, so the Manifests really were 100% identical - still not working. Bummer !!!
In the end I figured out that the problem is actually the way how NSIS' built-in !appendfile command works! In NSIS 2.46, this command apparently inserts an UTF-8 BOM right at the beginning of the file. ResHacker doesn't show this at all (it silently ignores the BOM), so we cannot see the difference in the Manifest XML code. But with a Hex Editor we can see the BOM is there, even in the embedded Manifest in the final EXE. Original EXE didn't have this. And Windows doesn't like the BOM at the beginning of the XML Manifest at all! It simply refuses to run the installer. So I'm now using a "static" file, rather than a file generated by NSIS' !appendfile command, and all is fine again.
Maybe this info should be added somewhere...
Apart from that, the new version is now working nicely in Windows 10. Great work 👍
Are you using a Unicode 2.46 build?
Stu
Stu
Yes.Originally Posted by Afrow UK View PostAre you using a Unicode 2.46 build?
Stu
(I know NSIS 3.x has an /CHARSET option for the !appendfile command, but NSIS v2.46 doesn't)
Why are you still using the Unicode fork?Originally Posted by LoRd_MuldeR View PostYes.
(I know NSIS 3.x has an /CHARSET option for the !appendfile command, but NSIS v2.46 doesn't)
Simply put, it has been working very stable for me for many years.Originally Posted by Anders View PostWhy are you still using the Unicode fork?
Also, I have quite a number of projects, which are at a point where we cannot make "major" changes, such as switching the whole install system, from NSIS 2.x to 3.x. The required testing would simply generate too much cost for too little gain.
Nonetheless, switching to NSIS 3.x for future projects as well as for my personal stuff definitely is on my agenda!
!appendfile in v3 is ANSI by default unless the file already has a BOM.
We would of course prefer if as many bugs as possible are found while we are still in beta...
Also, when reporting issues in the future, please say so if you are using an unofficial fork.
The generated installers should be pretty similar, most of the differences between v3 and the fork are in the compiler.Originally Posted by LoRd_MuldeR View PostAlso, I have quite a number of projects, which are at a point where we cannot make "major" changes, such as switching the whole install system, from NSIS 2.x to 3.x. The required testing would simply generate too much cost for too little gain.
We would of course prefer if as many bugs as possible are found while we are still in beta...
Also, when reporting issues in the future, please say so if you are using an unofficial fork.