#8701 closed defect (fixed)
created_by submits too much personal information (operating system details)
| Reported by: | pendluuum | Owned by: | team |
|---|---|---|---|
| Priority: | normal | Milestone: | |
| Component: | Core | Version: | |
| Keywords: | privacy | Cc: |
Description
Since May (release version) JOSM fills in details of the operating system version without notifying the user about it and with out a way to turn it off (is there one?). This is no good privacy practice!
It should only include "JOSM" and "JOSM version number", if needed.
Attachments (0)
Change History (19)
comment:1 by , 12 years ago
| Keywords: | privacy added |
|---|
comment:2 by , 12 years ago
comment:3 by , 12 years ago
| Priority: | normal → critical |
|---|
This should have never been uploaded in case of privacy. Thought this information would only be used privately which still is not save as there is no secure connection available.
comment:4 by , 12 years ago
Agreed.
It is customary for email programs and browsers to include details such as operating system in the user agent. So I see no problem if JOSM does the same when JOSM sends requests to a remote host. Also for bug reports, we require this information.
However for the created_by tag in the changeset, "JOSM" and the version number should be enough.
comment:6 by , 12 years ago
@bastiK: exactly - the difference is: the info in the changeset is permanently public for everybody in the world. The info during a connection is only temporarily accessible (at least it should be) for the communication target (the server).
Thank you Don-vip.
comment:7 by , 12 years ago
| Resolution: | fixed |
|---|---|
| Status: | closed → reopened |
This fix (for a "critical" bug) should have been pushed out to release version much(!) faster. Still version 5939 is offered. Anyway, in the meantime every frequent OSM contributior who is using Webstart or is fastly manually updating has involuntary leaked his operating system details ... So it is not anymore that important to push it out now.
comment:8 by , 12 years ago
| Resolution: | → fixed |
|---|---|
| Status: | reopened → closed |
This is not so critical, and tested version is coming in a few days, have a look at DevelopersGuide/Schedule
comment:9 by , 12 years ago
Well, there should be bug fix releases of the "tested" versions ... I will not use the "tested" (which is apparently not that true) via webstart anymore but manually use a .jar and update every now and then AFTER reading the change log. Time which I also could use to contribute to OSM ...
comment:10 by , 12 years ago
pendluum, I think you are overreacting. The bug was reported and fixed quickly enough, and it never was a critical bug in the first place. OSM contributors tell the world *much* more about themselves than which operating system they use. I agree that there was a bug here and it is great that it has been fixed but now let's put the matter to rest, please.
follow-up: 12 comment:11 by , 12 years ago
thank you for your comment, framm. Yes, it indeed is very good that the bug got quickly fixed, however this fix is not quickly pushed to the users (makes the quickness of the fix nearly useless). I think it is not clever to not provide corrected versions until the next "tested" release. That way the "tested" users are stuck with the bugs.
Yes, OSM contributors are usually providing much information about themselves - however, that is voluntary and known to them. This bug caused a disclosure of information which was (likely) not known to the contributors. By the way: the OS info string is quite verbose on some operating systems (which may be in addition quite exotic) - I am not talking of Windows 7 x64 here.
Sorry for being quite annoyed by this (I think more people would be if they would know) - the JOSM devs are still doing very good work!
comment:12 by , 12 years ago
Replying to pendluuum:
thank you for your comment, framm. Yes, it indeed is very good that the bug got quickly fixed, however this fix is not quickly pushed to the users (makes the quickness of the fix nearly useless). I think it is not clever to not provide corrected versions until the next "tested" release. That way the "tested" users are stuck with the bugs.
We don't make bug-fix releases except there are major bugs in a tested. See also Releases about our release plan.
Remember that we have limited capabilities - mainly regarding OUR time.
Feel free to contribute to JOSM in any way you want, but don't expect that we change a system which works very well for several years now without real need. If someone will provide a bug-fixed version of josm-tested we have no objections, but we see no need for this at the moment. JOSM release policy may change, but ATM it is unlikely this happens in the near future.
Regarding this issue: In any case this bug does not have enough importance for special care.
comment:13 by , 12 years ago
| Priority: | critical → normal |
|---|
comment:14 by , 12 years ago
note that the reverter plugin reintroduced the bug! It was fixed 16th June 2013 at 15:04 https://trac.openstreetmap.org/changeset/29663/subversion However, the JOSM tested version does not update to a newer version than 29561 where the *bug is still active*. Not that nice ...
Could the method getAgentString be set to default to not leak the user's OS details? Maybe there are more faulty plugins?!
comment:15 by , 12 years ago
So please update to latest. This problem is fixed for good, please let it go.
follow-up: 18 comment:16 by , 12 years ago
Don-Vip, that does not answer my question.
However, yes it is okay. Thank you for finding the problem before I did (yesterday) in the plugin. I guess that with the next tested release the plugin will also update.
comment:17 by , 12 years ago
The answer is no. There was a problem (thanks for reporting it), but now it has been fixed, so we don't care anymore about this one but we focus on another ones. We have a pool of 1000+ open tickets and limited resources for them.
comment:18 by , 12 years ago
Replying to pendluuum:
Don-Vip, that does not answer my question.
However, yes it is okay. Thank you for finding the problem before I did (yesterday) in the plugin. I guess that with the next tested release the plugin will also update.
If you did not change your plugin policy preferences not to do so, the answer is yes.
comment:19 by , 12 years ago
Okay. latest versions work (I tested it with latest non-"tested" JOSM and the reverter plugin: JOSM/1.5 (6041 de);reverter_plugin/29663).
I could not find other equally problematic calls in the plugins directory:
Version.getInstance().getAgentString() without "false" is called in the following files, but apparently only used for server download connections:
The change has been introduced in r5819 on 2013-03-31 and refined in r5851 on 2013-04-14. The driving factor seems to have been "better bug reports" but I agree that operating system details should not be uploaded to OSM without involving an user decision.