Hexacorn (Adam) Thu 2020-11-26 2:14 PM Hi Stuart, Happy to help. I don't envy the challenges of writing a software that is trying to make peace with everything that Microsoft designed over last 30 years. I used to code for Windows back in MFC/Delphi times and with win32 API. At that time things were a bit more predictable, or perhaps my memory is playing a trick on me and it was as terrible as it is now, if no more :-) Coding is certainly harder than reversing... I will release an update to the script and I'd like to credit you for helping in spotting / fixing the bug. How would you like to be credited? Cheers, Adam On 2020-11-26 19:27, V Stuart Foote wrote: > Adam, > > That makes sense. > > We had an issue when implementing Win7 Jump Lists for tdf#35785 [1], IIRC because LibreOffice's core is an EXE launcher of the cross platform bin (soffice.exe / soffice.bin ) we debug against the bin. Each "module" of the core receives its own EXE launcher, and receives file associations against that launcher. > > To provide Win7 Jump list handling we had to create unique AppUserIDs for each module. Not sure why they would not have ended up in CustomDestination-ms, but clearly using the "automatic destination" parsing of scalc.exe, swriter.exe launchers is not what we are doing. > > There was some dust up over "advertised" vs "non-advertised" shortcuts handling in the MSI packaging, and we left them non-advertised. But I guess it also means that since we remove the registry entries on upgrade installs our issues on tdf#76131 are self inflicted. > > Thanks again so much though for the work on your appid_calc script, it is very helpful. > > Stuart Foote > San Antonio, Texas > > [1] https://bugs.documentfoundation.org/show_bug.cgi?id=35785 > > From: Hexacorn (Adam) > Sent: Thursday, November 26, 2020 11:54 AM > To: V Stuart Foote > Subject: [EXTERNAL] Re: [EXTERNAL] Re: [EXTERNAL] Re: appid_calc.pl script broken, does it need a revised polynomial for the AppID algorithm > > **EXTERNAL EMAIL** > This email originated outside of The University of Texas at San Antonio. > Please exercise caution when clicking on links or opening attachments. > > > Hi Stuart, > > Cool, that was a silly bug - > > as for the LibreOffice's APPID the situation is a bit more complicated :-) > > I looked at Explorer creating the jumplist and this is what I saw: > >//vsf -- image of a stacktrace showing the string 'TheDocumentFoundations.LibreOffice.Calc' being passed// > > > As I mentioned in the blog post, some applications redefine the default APPID with their own -- a such is the case for LibreOffice > e.g. Calc uses "TheDocumentFoundation.LibreOffice.Calc" and Writer uses "TheDocumentFoundation.LibreOffice.Writer" > > e:\>perl.exe appid_calc_fix.pl TheDocumentFoundation.LibreOffice.Calc > Hexadecimal number > 0xffffffff non-portable at appid_calc_fix.pl line 13. > Hexadecimal number > 0xffffffff non-portable at appid_calc_fix.pl line 262. > > appid_calc.pl v0.2 (c) Hexacorn 2013-2016. All rights reserved. > Visit us at http://www.hexacorn.com > Thx to Terry O. for spotting and reporting the bug in the way APPIDs were calculated. > foo: TheDocumentFoundation.LibreOffice.Calc > > TheDocumentFoundation.LibreOffice.Calc > > APPID = 83DD64E7FA560BD5 > > e:\>perl.exe appid_calc_fix.pl TheDocumentFoundation.LibreOffice.Writer > Hexadecimal number > 0xffffffff non-portable at appid_calc_fix.pl line 13. > Hexadecimal number > 0xffffffff non-portable at appid_calc_fix.pl line 262. > > appid_calc.pl v0.2 (c) Hexacorn 2013-2016. All rights reserved. > Visit us at http://www.hexacorn.com > Thx to Terry O. for spotting and reporting the bug in the way APPIDs were calculated. > foo: TheDocumentFoundation.LibreOffice.Writer > > TheDocumentFoundation.LibreOffice.Writer > > APPID = D38A3EA7EC79FBED > > > which match the ones you have shown below. > > See references to arbitrary assigned APP IDs in the Libre Office source code > https://lists.freedesktop.org/archives/libreoffice-commits/2013-July/054490.html > https://cgit.freedesktop.org/libreoffice/core/commit/?id=6205c58e262e9b82e815855199de462775fcd35b > > Hope this helps. > > Cheers, > Adam