Fix minimum length bug for patterns containing (*ACCEPT).

This commit is contained in:
Philip.Hazel 2019-06-18 16:07:43 +00:00
parent 3b2fa4dff2
commit ef79b978a6
6 changed files with 23 additions and 8 deletions

View File

@ -31,7 +31,13 @@ minimum is potentially useful.
9. Some changes to the way the minimum subject length is handled:
* When PCRE2_NO_START_OPTIMIZE is set, no minimum length is computed;
pcre2test omits this item instead of showing a value of zero.
pcre2test now omits this item instead of showing a value of zero.
* An incorrect minimum length could be calculated for a pattern that
contained (*ACCEPT) inside a qualified group whose minimum repetition was
zero, for example /A(?:(*ACCEPT))?B/, which incorrectly computed a minimum
of 2. The minimum length scan no longer happens for a pattern that
contains (*ACCEPT).
* When no minimum length is set by the normal scan, but a first and/or last
code unit is recorded, set the minimum to 1 or 2 as appropriate.

View File

@ -10041,6 +10041,7 @@ if (cb.had_accept)
{
reqcu = 0; /* Must disable after (*ACCEPT) */
reqcuflags = REQ_NONE;
re->flags |= PCRE2_HASACCEPT; /* Disables minimum length */
}
/* Fill in the final opcode and check for disastrous overflow. If no overflow,

View File

@ -517,6 +517,7 @@ bytes in a code unit in that mode. */
#define PCRE2_HASBKPORX 0x00100000 /* contains \P, \p, or \X */
#define PCRE2_DUPCAPUSED 0x00200000 /* contains (?| */
#define PCRE2_HASBKC 0x00400000 /* contains \C */
#define PCRE2_HASACCEPT 0x00800000 /* contains (*ACCEPT) */
#define PCRE2_MODE_MASK (PCRE2_MODE8 | PCRE2_MODE16 | PCRE2_MODE32)

View File

@ -1607,13 +1607,13 @@ if ((re->flags & (PCRE2_FIRSTSET|PCRE2_STARTLINE)) == 0)
}
/* Find the minimum length of subject string. If the pattern can match an empty
string, the minimum length is already known. If there are more back references
than the size of the vector we are going to cache them in, do nothing. A
pattern that complicated will probably take a long time to analyze and may in
any case turn out to be too complicated. Note that back reference minima are
held as 16-bit numbers. */
string, the minimum length is already known. If the pattern contains (*ACCEPT)
all bets are off. If there are more back references than the size of the vector
we are going to cache them in, do nothing. A pattern that complicated will
probably take a long time to analyze and may in any case turn out to be too
complicated. Note that back reference minima are held as 16-bit numbers. */
if ((re->flags & PCRE2_MATCH_EMPTY) == 0 &&
if ((re->flags & (PCRE2_MATCH_EMPTY|PCRE2_HASACCEPT)) == 0 &&
re->top_backref <= MAX_CACHE_BACKREF)
{
int backref_cache[MAX_CACHE_BACKREF+1];

2
testdata/testinput2 vendored
View File

@ -5623,4 +5623,6 @@ a)"xI
/((?=a))[abcd]/I
/A(?:(*ACCEPT))?B/info
# End of testinput2

View File

@ -17026,6 +17026,11 @@ Capture group count = 1
First code unit = 'a'
Subject length lower bound = 1
/A(?:(*ACCEPT))?B/info
Capture group count = 0
First code unit = 'A'
Subject length lower bound = 1
# End of testinput2
Error -70: PCRE2_ERROR_BADDATA (unknown error number)
Error -62: bad serialized data