Errors in apitest

[Charlie Hull]

Several errors here, not sure if this is one problem or several. All occur on each type of database.

Running test: qterminfo1... .\api_anydb.cc:1262: ((mymset1a.get_termfreq(term1)) == (mymset2a.get_termfreq(term1))) Expected mymset1a.get_termfreq(term1)' and mymset2a.get_termfreq(term1)' to be equal: were 1 and 3

Running test: collfreq1... .\api_db.cc:1082: ((db1.get_collection_freq("this")) == (15)) Expected db1.get_collection_freq("this")' and 15' to be equal: were 11 and 15

Running test: multidb2... .\api_anydb.cc:286: ((mymset1.size()) == (mymset2.size())) Expected mymset1.size()' and mymset2.size()' to be equal: were 2 and 3

Running test: rsetmultidb1... DocNotFoundErrorDocNotFoundError: REMOTE:No termlist for document 4

Running test: rsetmultidb2... DocNotFoundErrorDocNotFoundError: REMOTE:No termlist for document 4

Running test: multiexpand1... DocNotFoundErrorDocNotFoundError: REMOTE:No termlist found for document 7

[Olly Betts]

All these tests open a database with two source files, so this would seem to be a duplicate of bug#217.

Except there you said an exception was thrown - did you patch around that or something?

[Charlie Hull]

No I didn't; it would seem best to try and fix 217 first. I'll accept this one.

[Charlie Hull]

Apitest output after fixing #217

[Richard Boulton]

The relation to #217 is that if only some of the source files get indexed, this would result in incomplete databases, and could lead to the kind of errors listed here. I suspect that your proposed fix to #217 doesn't actually fix it fully, somehow.

[Olly Betts]

Try whatever the equivalent of this is, then rerun apitest:

cd tests rm -rf .flint .quartz .multi

[Richard Boulton]

According to Charlie, Olly's suggestion combined with removing the old test databases fixes the problem.

This was essentially a duplicate of #217 - just symptoms of the problems with #217 failing to build the test databases correction. Resolving it as such.

* This bug has been marked as a duplicate of 217 *