diff options
author | Jeremy Evans <code@jeremyevans.net> | 2022-05-26 13:25:22 -0700 |
---|---|---|
committer | Jeremy Evans <code@jeremyevans.net> | 2022-07-21 08:02:32 -0700 |
commit | 423b41cba77719b4f62aa530593ad36a990f7c74 (patch) | |
tree | 546b853d2bd8396719d426262436c0e4215b9a5b /string.c | |
parent | cdbb9b8555b4ddcc4c557f25ad785cae6209478d (diff) | |
download | ruby-423b41cba77719b4f62aa530593ad36a990f7c74.tar.gz |
Make String#each_line work correctly with paragraph separator and chomp
Previously, it was including one newline when chomp was used,
which is inconsistent with IO#each_line behavior. This makes
behavior consistent with IO#each_line, chomping all paragraph
separators (multiple consecutive newlines), but not single
newlines.
Partially Fixes [Bug #18768]
Diffstat (limited to 'string.c')
-rw-r--r-- | string.c | 9 |
1 files changed, 7 insertions, 2 deletions
@@ -8967,6 +8967,7 @@ rb_str_enumerate_lines(int argc, VALUE *argv, VALUE str, VALUE ary) const char *eol = NULL; subend = subptr; while (subend < pend) { + long chomp_rslen = 0; do { if (rb_enc_ascget(subend, pend, &n, enc) != '\r') n = 0; @@ -8974,7 +8975,10 @@ rb_str_enumerate_lines(int argc, VALUE *argv, VALUE str, VALUE ary) if (rb_enc_is_newline(subend + n, pend, enc)) { if (eol == subend) break; subend += rslen; - if (subptr) eol = subend; + if (subptr) { + eol = subend; + chomp_rslen = -rslen; + } } else { if (!subptr) subptr = subend; @@ -8983,8 +8987,9 @@ rb_str_enumerate_lines(int argc, VALUE *argv, VALUE str, VALUE ary) rslen = 0; } while (subend < pend); if (!subptr) break; + if (rslen == 0) chomp_rslen = 0; line = rb_str_subseq(str, subptr - ptr, - subend - subptr + (chomp ? 0 : rslen)); + subend - subptr + (chomp ? chomp_rslen : rslen)); if (ENUM_ELEM(ary, line)) { str_mod_check(str, ptr, len); } |