Commit Graph

3 Commits

Author SHA1 Message Date
Piotr Pawel Stefaniak
ec5ac89ecd indent(1): improve handling of comments and newlines between "if (...)" or
"while (...)" and "else" or "{"

* Don't flush newlines - there can be multiple of them and they can happen
before a token that isn't else or {. Instead, always store them in save_com.
* Don't dump the buffer's contents on newline assuming that there is only
one comment before else or {.
* Avoid producing surplus newlines, especially before else when -ce is on.
* When -bl is on, don't treat { as a comment (was implemented by falling
through "case lbrace:" to "case comment:").

This commit fixes the above, but exposes another bug and thus breaks several
other tests. Another commit will make them pass again.
2018-06-03 14:03:20 +00:00
Piotr Pawel Stefaniak
a3abcad0b7 indent(1): don't produce unneeded space character in function declarators. 2017-07-23 14:04:45 +00:00
Piotr Pawel Stefaniak
61fd3be0b0 indent(1): add regression test cases
These examples show expected behavior of indent(1). They are meant to be used
together with a regression test mechanism, either Kyua, a Makefile or perhaps
something else. The mechanism should in essence do this:
  indent -P${test}.pro < ${test}.0 > ${test}.0.run
and compare ${test}.0.stdout to ${test}.0.run. If the files differ or the exit
status isn't 0, the test failed.

* ${test}.pro is an indent(1) profile: a list of options passed through a file.
  The program doesn't complain if the file doesn't exist.
* ${test}.0 is a C source file which acts as input for indent(1). It doesn't
  have to have any particular formatting, since it's the output that matters.
* ${test}.0.stdout contains expected output. It doesn't have to be formatted in
  Kernel Normal Form as the point of the tests is to check for regressions in
  the program and not to check that it always produces KNF.

Reviewed by:	ngie
Approved by:	pfg (mentor)
Differential Revision:	https://reviews.freebsd.org/D9007
2017-02-10 09:31:39 +00:00