Clarion 9 apps using Report to PDF to require C++ redistributable?
There's a thread in the Clarion 9 newsgroup that seems to confirm that if you're using the ReportToPDF functionality in Clarion 9, any such apps will need to also have the Microsoft Visual C++ 2005 Redistributable installed on that machine.Â
The tricky thing is that your apps may run fine on your test machines because you have the 2005 Redistrutable, but then things blow up when you ship product to your customers.Â
Here's a post by L Jack Wilson on the subject:
In reply, Diego said:
If it's the PNG support that triggers this, then perhaps any use of PNG files will require the C++ redistributable.Â
Graham Dawson wondered how to check for an existing installation of the redistributable. Richard Rose pointed out a couple of registry keys (the 2005 package is the one Clarion apps need):
- VC++2005 (SP1): HKLM\Software\Microsoft\Windows\CurrentVersion\Installer\UserData\S-1-5-18\Products\B25099274A207264182F8181ADD555D0
- VC++2008 (sp1): HKLM\SOFTWARE\Microsoft\Windows\CurrentVersion\Uninstall\{3C3D696B-0DB7-3C6D-A356-3DB8CE541918}
- VC++2008 (original): HKLM\SOFTWARE\Microsoft\Windows\CurrentVersion\Uninstall\{FF66E9F6-83E7-3A3E-AF14-8DE9A809A6A4}
The easiest way to deal with this seems to be to build your installs with LinderSoft's SetupBuilder, which can install the 2005 redistributable if it isn't already present. Yes, SetupBuilder is an additional cost but if you ship product it's almost certainly a wise investment even without this feature.Â
Update: Confirmation by Bob Z
Bob Zaunere has confirmed the above requirements:
Note the phrase "current c9 release" - in future releases the runtime will only need to be installed if you are using the PDF Report Generator and specifically using the new "Use Scan Copy Mode".Â
Update: More details on the "scan copy" innards, from Diego:
Deeply inside is converting the report to a png of some resolution and color deep to be able to keep the information of the report (font,images, resolution, content ,etc) exactly the same. With out hte need of embebing the Fonts or to support any other images formats and adding security to the content (a PDF is just text when not compressed and encripted) Estern char sets now can be used in the PDF with out having to store the fonts or have the Font in the user. With out that option the PDF engine process the report and re-create the report in PDF format, but not always the result is 100% exactly but using this option the print out of one and the other are 99.98% a match.
Graham Dawson has confirmed that when you use Scan Copy you do in fact get a PDF that is composed of images only and contains no searchable text.Â
Update: ClaRunExt.dll now dynamically loaded
As of build 10309 ClaRunExt.dll is no longer statically linked. From a post by Bob Z:
That .lib file is not needed because there is no static linking to Clarunext.dll - its been changed to a runtime dynamic load that only happens _if you are using the new 'Scan copy mode', so if your .app/project is looking for it then either; - its a leftover project reference from when it was a static link, and you can just delete it - you are using the IP driver in this .app (an update will be out today that removes its linking to clarunext - you don't have or are not using the latest templates; core (EE) or PDF if running PE