[LH4758] Table Crash

Extract This (linked) Scrivener Project
Open in Scrivener
Right-click inside table bounds and click ‘Table->Table…’:

Select all cells in the table:

Right-click on the table and select ‘Merge Cells’:

In Table properties increase the column count:

Click Apply.
Table is corrupted:

If scrivener doesn’t crash right away, add another row and click Apply again. It will crash.

What version of Scrivener for Windows is this in? If it’s in the beta, this should be posted in the Windows Beta Test forum.

Well I feel silly, for some reason I thought that’s the forum I was on.

Is there any way it can be moved, or do I have to post this all over again?

Send a DM to a moderator.

For reference, this is on RC11.

I can confirm that following the directions as posted causes Scrivener to crash on my system.

I believe that the error is that the properties are not updating after the “Merge cells” operation. If I close the table properties page before performing the “Merge Cells,” then unselect the table, then reselect the table and re-open Table Properties, I would expect to see that it shows me a 1x1 table. Instead, it is showing me a 2x3 table, which is stale data.

Thanks for the report! It’s been logged. To confirm, are you only getting this crash when merging all the cells in the table? I experimented with some other variations but the table in that case updated correctly with the adjusted row/column data and I didn’t get the crash. I’d like to know if that matches your experience.

It doesn’t seem to. I opened the OP’s attached test project, accepted the option to Continue, then added two more rows for a total of 4 rows x 3 columns. I then merged the top two rows and opened the table properties. I should have a 3x3 table now, but it’s still showing as 4x3,

[attachment=0]tablemergebug.jpg[/attachment]

There certainly are other scenarios that lead to a crash. My experience has lead me to play with merging cells and using the table properties to modify it from there. Sometimes it’s on add column/row. Other times it’s when I remove them.

The main point here is merged cells are poorly handled in general.

My understanding is that tables are a beast to work with from the developing standpoint, too. ;) So any reproducible steps you have for triggering a crash would be appreciated. It may be that fixing the bug with adjusting the columns after a merge will also bring stability to the other cases you’re seeing, but if you’re able to provide steps to replicate particular cases, it would allow the developers to test them directly.

Yep.

Reproduction steps take time to figure out. I lucked out with that example I posted.

I haven’t taken the time to exhaustively go through and find the exact steps for others, but in playing with it I found I was able to trigger a crash. The thing is it isn’t hard to cause, which is why my steps were simple.

If I have time I’ll try to write them down; it’s just that such steps aren’t always easy to come by, and are the result of much trial and error.