build_pdf ignores some bad scandata
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
Deriver |
New
|
Undecided
|
Unassigned |
Bug Description
See Bug #210359: that was filed on scribe2, but actually should be against the deriver. Restating the gist here:
The current behavior of build_pdf is to ignore all scandata if any error, no matter how small, is found in any of it. So if some (perhaps unused) variable cannot be read, then information about what pages to drop, how to number the pages, and what ppi to use on a page is all thrown away.
So, per discussion with Hank, plan to change the code so that
(a) if scandata is specified but the file can't be found or similar, it will abort
(b) if an essential parameter (i.e., one being used) cannot be parsed, it will abort
(c) if a non-used parameter cannot be parsed, it will try to patch the data and continue
Thanks in advance for any suggestions or warnings or clarifications or additional info from anybody.