@deadalnix : Can you please get specific about what you mean by 'wild guesses' in the current patch .
- Queries
- All Stories
- Search
- Advanced Search
- Transactions
- Transaction Logs
Advanced Search
Jul 3 2017
The test framework is using the default value of '.' for SRCDIR .
I've tested this using a build with Qt4 and it works on my system.
I'll try building on older Ubuntu to see if its ok there too.
Please update the TEST PLAN , that was the point of my initial request.
Fixes the test on my system (which is not plain Debian 8).
Jul 2 2017
If there is no intention to further test on nolnet (expressed by further comments / planning discussion on this ticket), I will withdraw (abandon) this Diff by Tue 4 Jul 12:00AM UTC .
I'm ok with the simplification of the regex and nailing it down to use std::regex::basic .
Jul 1 2017
Jun 30 2017
Repurpose this diff to make it into better error handling for misconfiguration.
Jun 29 2017
Changes look good to me.
I do not have any issue with this change, but the Test Plan is inadequate since there are no unit tests covering this, hence 'make check' does not test it - it requires running at least the Python tests.
Jun 28 2017
Accepted - this is a revert of 05fdbe78b and back to existing Core code .
Agreed, D284 code change is simpler and appears to be just as stable.
Minor things only.
Travis tests results are green:
In D282#4611, @deadalnix wrote:No. If the argument passed down is wrong, you want to report it, not do wild guesses that may or may not be what the caller intended.
Jun 25 2017
Jun 23 2017
Taking this to close / land as there seems to be no outstanding action required.
If, when we format everything before release, we could solve the problem of single-line if(some_condition) { do_something(); } contraction, that would be great.
I'd much prefer the body to be on a different line.
Jun 22 2017
Updated after the initial stderr checkers are no longer
passed to initialize_chain. It is transformed instead into
a test which calls initialize_chain directly to set up
a test chain.
Ok, some doxygen comment misery, but that's for fixing later.
Resolved with D243
Ignore previous comment - I checked and there's no linter diff on my tree for transaction.h, so I guess it's good.
No diff against transaction.h ?
Accepted, nothing major noticed.
More cleanups after review
Accepted this on account of argument against semantic change.
If there was such a semantic change it would need separate testing to show it is ok.
Simplify test as suggested by review
I'm ok with these changes and have convinced myself (by a little math) that the block stuffing works as advertised.
Initial batch of fixes from deadalnix review comments
In case of changing compiled files source tree, make check should be added to Test Plan.
Jun 21 2017
I didn't test this yet. But you wrote
"In practice it does not seem to work right despite -Wimplicit-fallthrough=3"
Note that you used "FALLTHROUGH" but on this page you linked, https://developers.redhat.com/blog/2017/03/10/wimplicit-fallthrough-in-gcc-7/:
it says "FALLTHRU".
Might that be the reason explaining this?