I used to upload diagnostic reports here following a crash, but I appreciate that the team is not always able to respond given the demands on their time. It may well be that these reports are reviewed even when no reply is possible; however, not knowing whether that is the case does make me somewhat less inclined to submit them.
That said, I wonder whether it might be useful to give users an opportunity to provide some additional context when a crash report is automatically generated following project recovery, before it is forwarded to the Dorico team. For example, users could briefly describe what they were doing immediately before the crash occurred, or note any circumstances they believe may have contributed to it. Such information could potentially help identify patterns or narrow down the causes of issues that are not readily apparent from the diagnostic data alone.
I appreciate that these automatically generated reports are anonymous by design, but it might also be worth offering users the option of including contact details should the team wish to follow up for further information. I suspect many users would be happy to do so if it helped with diagnosing and resolving problems.
I think it’s been said that the automatic crash reports come to the team indirectly, and so while they may yield some general observations about things that need fixing, they are not as useful as the Diagnostic Reports that individual users supply directly to them.
Yes, I believe you are correct. If I recall correctly, Daniel also mentioned that the team uses an automated process to help prioritise fixes based, at least in part, on the number of reports received.
There is, of course, no downside to sending a diagnostic report via PM or email. I am simply suggesting that it might be easier and more effective to invite users to submit a report when recovering a project after a crash, as the events leading up to the crash are likely to be much clearer in their minds at that point.
If there were an option to generate and attach a diagnostic report automatically, together with a field in which users could briefly describe what they were doing when the crash occurred, I would find that the most efficient way to proceed. Others may feel differently, but I suspect it could encourage more users to submit reports and provide potentially useful contextual information.
Absolutely, Jim, but that wouldn’t be accompanied by the automatically generated report that users are invited to send to Steinberg when a project is recovered after a crash. I have no idea what information is actually included in those reports, but on a number of occasions, after emailing the team separately, I’ve been asked to provide additional details about the circumstances surrounding the crash.
That is really the basis of my suggestion: it might be useful to provide an opportunity to include that contextual information at the point a report is submitted, while the events are still fresh in the user’s mind. In any case, it was only a suggestion, one that I’m beginning to regret making…
When you restart Dorico after some (all?) crashes, you are prompted to send a crash report to Steinberg. This is different from uploading a crash report or diagnostics here to the forum; the former method results in anonymized reports which seem to be less prioritized by the dev team.
The OP is requesting the ability to add context to the post-crash auto-uploads, to make them slightly more helpful to the team.
@rraby , @asherber , Thank you both!
I haven’t had to deal with that (any crashes I’ve had were due to my own stupidity and not worth sending on to Steinberg upon restart)! I’m much more used to seeing reports asked for, and uploaded, here in the forum. Now I get it!
Could be, but I believe that sending the report is under Dorico’s control, in which case they might be able to collect further info from the user before sending.
Perhaps my experience is atypical, but I encounter crashes on a reasonably frequent basis. I use the software for up to ten hours a day, and it is not uncommon for Dorico to crash once or twice during that time. Without exception, however, the team has been extremely responsive and helpful, and fixes have often arrived in the very next update.
Earlier today, after a couple of crashes that the team have not yet been able to track down, I found myself wondering how many users actually click the button to send the automatically generated report, let alone take the additional step of uploading a diagnostic report (and a spindump, if the application hangs) to the forum, together with some contextual information about what they were doing at the time. How would a user even know to do this, unless they had been fortunate enough to find this place?
My suggestion is simply that there may be a way to make this process easier for users and, perhaps, more useful for the team. If users were given the opportunity, at the point of project recovery, to generate and submit a diagnostic report along with a brief description of the circumstances leading up to the crash, I suspect more reports would be submitted and the developers would receive richer, more actionable information to help identify and resolve issues.
Of course, I may be completely wrong, but it seemed to me a potentially useful enhancement rather than an unreasonable feature request.
The Version History for Dorico 6 lists many situations that have the potential to cause crashes, which demonstrates both that such issues do arise and that the development team is highly effective at identifying and resolving them. My suggestion was simply that, when crashes do occur, it might be possible to make it easier for users to provide the contextual information that can help the team track them down more quickly.
The report Dorico invites users to send following a crash is anonymous. As Daniel has explained in the past, these reports do not always provide sufficient information to diagnose a problem. Because they are anonymous, there is no way for the development team to associate a report with a particular user or to request additional information if it is needed.
At the point where users are invited to submit a report following a crash, there is nothing to indicate that it may be helpful to generate and submit a full diagnostic report (with a spindump, where appropriate) via the forum or directly to a member of the development team.
These logs do not contain the contextual information that is so often requested when investigating a crash: what the user was doing at the time, the steps that led up to the problem, whether it is reproducible, and so on.
My suggestion was simply that there may be an opportunity to streamline this process and capture some of that information at the point of recovery. Nothing more, nothing less.
