Re: switch to C99 breaks DFA tests

2017-01-10 Thread Paul Eggert
On 01/10/2017 01:27 PM, Assaf Gordon wrote: > dfa.c:727:3: error: ‘for’ loop initial declarations are only \ > allowed in C99 mode >for (int i = 0; i < CHARCLASS_WORDS; i++) >^ > Yes, the DFA module uses C99 statement-after-declarati

switch to C99 breaks DFA tests

2017-01-10 Thread Assaf Gordon
Hello all, I noticed that some standalone tests now fail because they use c99 constructs. For example: ./gnulib-tool --create-testdir --dir test1 dfa cd test1 ./configure make fails with: dfa.c: In function ‘fillset’: dfa.c:727:3: error: ‘for’ loop initial declarations a

update DEPENDENCIES file

2017-01-10 Thread Bruno Haible
An update to the DEPENDENCIES file: * Don't recommend protocols like http and ftp that are vulnerable to man-in-the-middle attacks. * Update minimum autoconf version: Some m4/*.m4 files require 2.60 or newer. * Update minimum automake version. (Guesswork of mine.) * Update recommended gperf ver

Re: [PATCH 2/2] maint: time stamp -> timestamp

2017-01-10 Thread Jim Meyering
On Mon, Jan 9, 2017 at 11:58 AM, Paul Eggert wrote: > Use the spelling "timestamp", as that is what POSIX uses. Whenever making a global change like this (for which regression is probable), I like to add something to help ensure consistency. We already have a "syntax-check" rule to prohibit thing

[PATCH 1/4] dfa: omit unnecessary allocation

2017-01-10 Thread Paul Eggert
* lib/dfa.c (dfaanalyze): Do not allocate follow set, since an all-zero follow set works just fine. --- ChangeLog | 4 lib/dfa.c | 2 -- 2 files changed, 4 insertions(+), 2 deletions(-) diff --git a/ChangeLog b/ChangeLog index f0cd277..beeefb7 100644 --- a/ChangeLog +++ b/ChangeLog @@ -1,5 +

[PATCH 3/4] dfa: shrink constraints from 4 bits to 3

2017-01-10 Thread Paul Eggert
* lib/dfa.c (newline_constraint, letter_constraint) (other_constraint, prev_newline_dependent) (prev_letter_dependent, NO_CONSTRAINT, BEGLINE_CONSTRAINT) (ENDLINE_CONSTRAINT, BEGWORD_CONSTRAINT, ENDWORD_CONSTRAINT) (LIMWORD_CONSTRAINT, NOTLIMWORD_CONSTRAINT): Constraints need only 3 bits, not 4. U

[PATCH 2/4] dfa: omit unnecessary ptrdiff_t check

2017-01-10 Thread Paul Eggert
* lib/dfa.c (alloc_position_set): Do not worry about ptrdiff_t overflow, since xnmalloc does that now. --- ChangeLog | 4 lib/dfa.c | 2 -- 2 files changed, 4 insertions(+), 2 deletions(-) diff --git a/ChangeLog b/ChangeLog index beeefb7..91cdc6d 100644 --- a/ChangeLog +++ b/ChangeLog @@ -1,

[PATCH 4/4] dfa: minor simplification with emptyset

2017-01-10 Thread Paul Eggert
* lib/dfa.c (build_state): Simplify by using emptyset. --- ChangeLog | 5 + lib/dfa.c | 5 + 2 files changed, 6 insertions(+), 4 deletions(-) diff --git a/ChangeLog b/ChangeLog index e0b73b5..8fda293 100644 --- a/ChangeLog +++ b/ChangeLog @@ -1,3 +1,8 @@ +2017-01-10 Paul Eggert + +

Re: bug#25390: Segfault with sed 4.3

2017-01-10 Thread Paul Eggert
Assaf Gordon wrote: I'm also considering duplicating the test - once with and once without valgrind. Is this warranted or an overkill ? Sounds like overkill to me. If it works with valgrind, it'll work without.