Eric Curtin

I wonder is it worth adding a just build the tests option to the make file, I kinda hacked that option together here.

Sorry that wasn't clear at all I do this (I basically comment out the execution of tests): `sed -i "s/\$(MAKE) \$(AM_MAKEFLAGS) check-TESTS/#\$(MAKE) \$(AM_MAKEFLAGS) check-TESTS/g" Makefile` so when I run `make...

``` #!/bin/bash git clean -fdx git reset --hard origin/master FIND="find . -name *.c -o -name *.h" GREP="grep -v 'mu\|dqlite.h'" autoreconf -i ./configure CFLAGS=-O0 sed -i "s/\$(MAKE) \$(AM_MAKEFLAGS) check-TESTS/#\$(MAKE) \$(AM_MAKEFLAGS) check-TESTS/g"...

@freeekanayaka I think it's as you stated now :smile: The script that did the work below ``` #!/bin/bash FIND="find . -name *.c -o -name *.h" GREP="grep -v mu\|dqlite.h" autoreconf -i...

We now have passing tests, these are the exceptions (not including anything that broke builds or tests): `^_|^page_size|^sqlite|^foreign_keys|^journal_mode|^cache_size|^wal_autocheckpoint`

> ```diff > requestConnect__decode > ``` Yeah these tokens must be treated differently to all the others in the script. I must grep for "##__" macros and ignore everything preceding...

> Another case with mixed style is `struct dqlite_node` defined in `server.h`: > > ```c > struct dqlite_node > { > pthread_t thread; /* Main run loop thread. */ >...

The changes related to macros with ## were added

Fixed that struct with the (partial) commits. Will look for structs in the overall diff when I get some time tomorrow.

ok I think it's pretty good, got rid of any inconsistencies in structs I saw in the: `git diff origin/master` output