Hello Bruce,
> Here's the output for all of the failing tests. Not sure what's going on
> with the "WARNING: Empty testfile detected" errors.
I've added the code below to test/RegressTests.py a little while ago:
line = self.fd.readline() # self.fd is a file object created with open()
if not line:
print >>sys.stderr, "WARNING: Empty testfile detected: %s" % (self.filename)
return False
I thought it would only fail for testfiles that are empty,
as it seems that is not the case.
I wonder what readline() returns that the `if not line:` test succeeds,
even though the file isn't empty.
In a previous message you mentioned Python 2.7.10 is used on Cygwin,
I use the same version on Mac OS X 10.11 El Capitan, yet I do not see
the issues.
Maybe you could rerun the tests with the following slight modification
of test/RegressTests.py to aid in debugging:
line = self.fd.readline() # self.fd is a file object created with open()
if not line:
print >>sys.stderr, "WARNING: Empty testfile detected: %s »%s«" %
(self.filename, line)
return False
And just to be safe please provide the SHA256 sums of those files, e.g.:
% $SHELL --version
zsh 5.0.8 (x86_64-apple-darwin15.0)
% for t in
test/baseline/{cmd-price{s,db},feat-fixated-prices,opt-{datetime-format,gain,historical,market,price,revalued,time-report}}.test;
do \
echo -n "$t: "; openssl dgst -sha256 < $t; \
done | column -t
test/baseline/cmd-prices.test:
dbee77184f7355ba172a672bd6ed21ffd3dfdc995508fdcfea8d7c61f13b4ff7
test/baseline/cmd-pricedb.test:
e6e8ca3e914ae90ca34b9f4ca32075af456b2526a41a7882bb6cf5bc11d9b45c
test/baseline/feat-fixated-prices.test:
0d7040bd96d527345c890b3844382893afecece997957291de2208c053c56fd5
test/baseline/opt-datetime-format.test:
2acaef081ad742ae1ad05544e08ad863ed539599d717e5195ef15c9beebae3b3
test/baseline/opt-gain.test:
36b11c393ee915cedcbf7c9d6eafdbce3c22c30e0cc68f0e4810926ebfec4155
test/baseline/opt-historical.test:
6a408b24917e409fdf18745f708eaa7e94b3ed95700cc9e98ae9c0194b7823bf
test/baseline/opt-market.test:
110ce3bf772460149b4832090b630fc71df55bdd54ef1fa38daf76ed1f6cb6c1
test/baseline/opt-price.test:
0dad7401c8e28bf9a604c5f2cab52a4ed9c4ea61827a519649b3979c0ddaf8f0
test/baseline/opt-revalued.test:
e8792ae0bc1b47e4d2d76851a122944be7d89eb65a2be0493b227be1c4337d72
test/baseline/opt-time-report.test:
87582cbb559735558bc9041ab53517d4f8e187bafa550fe41843bdd0935cbbcd
> - the sort order for pricedb is different (I tried changing the input
> order, but the output came out in a different order again)
I've seen this happen in Travis CI too.
Who has an idea what might be happening here?
> - the time report output has an extra line with a single *
That's odd. If timelog entries are cleared, using a capital oh ('O')
for the check-out time,they are marked with an asterisk ('*') in the output.
But none of the entries in the opt-datetime-format.test are cleared.
Basically the --time-report option is just an alias for the
--balance-format option with a pre-defined format string.
The code can be found in src/report.h:973 (ff).
I'm doing some experiments to get a better idea what might be going wrong.
Cheers,
Alexis
--
---
You received this message because you are subscribed to the Google Groups
"Ledger" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to ledger-cli+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.