Bug 44425 - [patch] junit TestDataValidation keeps modifying file TestDataValidation.xls
Summary: [patch] junit TestDataValidation keeps modifying file TestDataValidation.xls
Status: RESOLVED WONTFIX
Alias: None
Product: POI
Classification: Unclassified
Component: HSSF (show other bugs)
Version: 3.0-dev
Hardware: Other other
: P2 minor (vote)
Target Milestone: ---
Assignee: POI Developers List
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2008-02-14 16:59 UTC by Josh Micich
Modified: 2008-02-15 03:10 UTC (History)
0 users



Attachments
svn diff of one changed test case (67.13 KB, text/plain)
2008-02-14 17:02 UTC, Josh Micich
Details
tar bz2 of one added test utility class (2.13 KB, patch)
2008-02-14 17:02 UTC, Josh Micich
Details | Diff

Note You need to log in before you can comment on or make changes to this bug.
Description Josh Micich 2008-02-14 16:59:45 UTC
TestDataValidation is currently structured to re-write TestDataValidation.xls 
every time it is executed.  Unfortunately this makes TestDataValidation.xls 
always appear to subversion as a modified file.

The test was re-structured to consider the existing copy of 
TestDataValidation.xls as the 'proof' copy. The test now generates a separate 
copy in memory, and these two are compared (ignoring timestamps etc). The 
comparison step is new.  User interaction will only be requested if the files 
differ.

The test still generates the exact same content so TestDataValidation.xls did 
not change with this patch.
Comment 1 Josh Micich 2008-02-14 17:02:03 UTC
Created attachment 21532 [details]
svn diff of one changed test case
Comment 2 Josh Micich 2008-02-14 17:02:32 UTC
Created attachment 21533 [details]
tar bz2 of one added test utility class
Comment 3 Nick Burch 2008-02-15 03:10:05 UTC
I appear to have tackled this issue at the same time you did...

I decided to just change it to output into the system temp directory, since
you're supposed to open the file up in excel and check. Re-running the test on
different computers can give slightly different files, so I don't think we want
to be testing against a given file in svn, as we might have spurious failures
that way.

So, I think we don't need this patch, but thanks for trying to fix it!