Summary: | LibreOffice installer doesn't add appropriate entries to TrackPoint TP4TABLE.DAT file | ||
---|---|---|---|
Product: | LibreOffice | Reporter: | MartinPC <PeterCraigMartin+LO> |
Component: | LibreOffice | Assignee: | Not Assigned <libreoffice-bugs> |
Status: | RESOLVED WONTFIX | ||
Severity: | enhancement | CC: | ilmari.lauhakangas |
Priority: | low | ||
Version: | 5.0 all versions | ||
Hardware: | x86-64 (AMD64) | ||
OS: | Windows (All) | ||
Whiteboard: | |||
Crash report or crash signature: | Regression By: | ||
Bug Depends on: | |||
Bug Blocks: | 113117 |
Description
MartinPC
2017-08-30 22:48:52 UTC
Ok, let's set to NEW. Did you test rules with all the other LibreOffice applications? I think your solution is best, if it works for you and then it's a question of someone adding the functionality to apply those settings to TP4TABLE.DAT upon install. I'll set to low priority due to it being somewhat niche and the workaround (people will find this report with web search). I don't think we can do that. The tp4table.dat is some proprietary settings file, having its own format. The scrolling device developers decided to use some per-application settings. But LibreOffice cannot edit that file: 1. That file is not a public API - the only mention I found on the official IBM site [1] does not describe the details; both on Lenovo site and on Synaptics', I was unable to find anything related. Should we have additional workload to track possible changes to that file format maybe? 2. The file does not have some standard location. It is reported to appear under C:\Windows\System32\; C:\Windows\System32\DriverStore\FileRepository\tp4track.inf_45c84db5\; C:\Program files\Synaptics\SynTP\; C:\SWTOOLS\APPS\TRACKPT\; C:\Program Files\Lenovo\TrackPoint\. It's obviously dependent on specific vendor using Synaptic solution in its product. 3. The file uses executable name to refer to the application. And guess what, there is at least one other office suite out there, which has those soffice.exe/scalc.exe/... executable names. So trying to write there, we could break that other suite settings. The comment on the line above does not allow driver to discriminate real applications - it's just a comment. 4. Even if there's no other suite on the system - there's no way to tell if user had modified that data already to own liking. So we could also break user-defined settings. 5. There's no way for LibreOffice developers to know if those settings fit any equipment that might be using that file. It might be that it's device-dependent. 6. Last thing: the scrolling device developer decided to make it application-aware - so that developer should keep track of applications out there; to them, testing LibreOffice means downloading a free application; a LibreOffice developer should *buy* some device to be able to test that somehow. My opinion: WONTFIX. [1] https://www.ibm.com/support/knowledgecenter/en/SSS9FA_12.0.0/com.ibm.hod.doc/help/mousewheel.html Thanks, as a whole that is a pretty convincing argument for wontfix, so I will close. But anyway, I suppose OP has done some great work here. It might be best to provide the results to Synaptics - they could be the best people who can include it to their products. MartinPC: Please consider sending that to them. Thank you for your interest making LibreOffice better! (In reply to Mike Kaganski from comment #2) I don't know what handle will appear with this comment, but I used to be MartinPC. I seem to recall that I stopped being able to log in with my old credentials at some point. I no longer remember the details, except that the reason was technical -- a change in hosting? -- not due to any abuse or impropriety on my part. > I don't think we can do that. And this now seems reasonable to me, in light of all the particulars you listed. I will say that it strikes me as highly questionable on the part of OEMs to *not* make their pointing-stick APIs fully public and to *not* use standard filenames and file locations for configuration files, within each major OS family. Do they think every application developer in the world is going to get in touch with them (and maybe pay them a little something, or cut them a special wholesale price on bundling?) to get their applications included in the pre-supplied pointing-stick configuration file, or to be given enough information to appropriately edit the configuration file at the time the applications are installed? For the life of me, I can't see a legitimate rationale for keeping this information private. (As someone with a modest background in antitrust, I *can* imagine an *illegitimate* rationale, namely, a combination in restraint of trade to unfairly promote some applications, e.g., Microsoft Office, and unfairly disadvantage competing applications, e.g., LibreOffice. The Document Foundation should ask the European Commission's Directorate General for Competition what *they* think. I'm a US citizen and antitrust enforcement is all but dead over here, so the ball is in LibreOffice's court on this one.) I may have already tried to contact Lenovo and Synaptics about this problem around the time I first posted this report. It was too long ago and I don't remember. I can try contacting them again. At any rate, I REALLY appreciate the effort and thoroughness Mike Kaganski put into investigating this problem and putting it to rest. THANK YOU. The "bug" is not LibreOffice's fault; it's the OEMs'. |