Drugwash, on Dec 12 2006, 12:58 PM, said:
Really? That's very bad, Riched20.dll 4.0 (5.40.x.x) is required for Metapad and many other programs to work with non-ASCII codepages, like East/Central European codepage 1250.
This topic was alaredy discussed here: http://www.msfn.org/...showtopic=72448 and the conclusion was that it is safe to use riched20.dll version 5.40.11.2218 because it does not break anything.
Here you can see the difference:
Richedit 3.0 (all versions):

Richedit 4.0 (all versions):

This is also noted here: http://www.liquidnin...ad/faq.html#Q18
This problem si common for Windows 98, Windows Me, Windows 2000, Windows XP, Windows Server 2003 and even Windows Vista.
Therefore I have opened the incident # SRQ060526601424 with Microsoft and asked them to correct this bug but after several months of e-mailing and phone calls the result was:
Quote
I regret to inform you that Peter Constable, Program Manager for Windows Globalization maintains that this is a “by design” behavior. Here’s some more info from Peter:
Please note some things regarding versions of the RichEdit control:
- When you see “5.30.23.1221”, ignore the initial “5.”; this is version 3.0.
- The system has never shipped version 5.0 of RichEdit. RichEdit is developed by Office, and they are currently developing version 6.x for Office 12. Version 5.0 would only have shipped with Office.
- Version 4.0 of riched20.dll has never been shipped by the system; it has only been shipped by Office. The system ships version 4.x of RichEdit using the file name msftedit.dll.
Furthermore we have discussed the possibility of msftedit.dll being used instead of riched20.dll when building a new application.
For this issue to be further pursued I will need you to send me a business impact plan so that Microsoft can decide whether this should be pursued or not.
I’ve already arranged to have a conference call with my manager to discuss this issue if you’d like and I’m also fully available and happy to help you with building the business impact plan if needed.
Quote
Therefore, if we can be of any further assistance on this case or if you have any suggestions for improving our service please do not hesitate to contact either myself or my manager Bruno Ribeiro on brunor@microsoft.com so we can ensure you receive our immediate attention.
As discussed with Bruno, I have forwarded the info and documentation regarding riched20.dll you provided, plus your workaround suggestion, to the program group in order for them to further study the issue and come up with a KB article. This, unfortunately, might take a couple of weeks.
As this case was related to a design issue, I have closed this case as non decrement for you.
It was a pleasure working with you. I only wished all our customers documented their cases as well as you did. Thank you!
Should you need any further assistance with this matter, please feel free to contact me directly and I will be happy to assist.
Very polite response - but nothing was done yet, even the bug was not documented in MSKB yet.
Back to RICHED20.DLL. Are you able to document your problems with Richedit 4.0 and Miranda? I don't use Miranda so I'd like to test them. The serious problem with Richedit 3.0 and 3.1 and Windows with default codepage 1250 (defualt in Czech, Hungarian, Polish, Slovak, Slovenian versions of Windows 98) has to be solved some way. Maybe there is some other workaround? Or someone will be able to patch riched20.dll 3.0?
My understanding is so that Richedit does a test if the character is supported in the currently selected font. If it is missing, then it switches to different font that contains it - just to be able to display the character and not empty rectangle only. Unfortunately there is a bug in Richedit 3.0, 3.1 and even 5.0 that causes font switching even when the character is supported by the currently selected font.
Regards,
Petr



Help



Back to top









