Bugs item #2535084, was opened at 2009-01-25 01:59
Message generated for change (Comment added) made by paulmcq
You can respond by visiting:
https://sourceforge.net/tracker/?func=detail&atid=616200&aid=2535084&group_id=96864
Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: None
Group: None
Status: Closed
Resolution: Fixed
Priority: 5
Private: No
Submitted By: Den Ivanov (dsdiv)
Assigned to: Paul McQuesten (paulmcq)
Summary: re2c 0.13.5 not compile with Sun C
Initial Comment:
re2c 0.13.5 not compile with Sun C compiler at Solaris 10 x64:
CC -DHAVE_CONFIG_H -I. -g -c -o parser.o parser.cc
"parser.y", line 98: Warning (Anachronism): Assigning std::pair<unsigned,
re2c::RuleOp*> to std::pair<int, re2c::RegExp*>.
"parser.y", line 98: Error: Cannot cast from std::pair<unsigned, re2c::RuleOp*>
to std::pair<int, re2c::RegExp*>.
1 Error(s) and 1 Warning(s) detected.
Compiler version: "Sun C 5.8 Patch 121016-07 2007/10/03"
----------------------------------------------------------------------
>Comment By: Paul McQuesten (paulmcq)
Date: 2011-09-25 19:49
Message:
Code in revision 900.
----------------------------------------------------------------------
Comment By: Paul McQuesten (paulmcq)
Date: 2011-09-25 19:48
Message:
Sun compiler rightly objected to assigning a size_t to an int. With default
x86_64 memory model that is storing a uint64_t into an int32_t. OTHO, 2
billion spec entries is quite sufficient for this decade. Structure changed
to uint32_t, and code changed to do explicit truncation.
Could someone with a Solaris compiler please test this?
----------------------------------------------------------------------
Comment By: Paul McQuesten (paulmcq)
Date: 2011-05-24 00:57
Message:
Note that the patch proposed by davidmarker on 2010-10-04 is dangerous: it
assumes that an 'int' is as big as a 'size_t', which is not true for the
default x86_64 memory models on Linux where int is 4 bytes but size_t is 8.
----------------------------------------------------------------------
Comment By: Paul McQuesten (paulmcq)
Date: 2011-01-09 21:52
Message:
I do not have your compilers, so cannot really check, but I think the
proposed patch just hides a symptom. I think the actual fix should be on
line 495 of re.h. Change first member of NRegExp to size_t:
from: typedef std::pair<int, RegExp*> NRegExp;
to: typedef std::pair<size_t, RegExp*> NRegExp;
The second member is okay as is, because RuleOp is a derived class of
RegExp.
----------------------------------------------------------------------
Comment By: David Marker (davidmarker)
Date: 2010-10-04 13:35
Message:
Hit same issue, resolved it with this patch (but haven't tested yet, just
got it to build):
--- parser.cc-orig Mon Oct 4 14:22:39 2010
+++ parser.cc Mon Oct 4 14:21:17 2010
@@ -198,7 +198,8 @@
else
{
size_t nIndex = specMap.size() + 1; // 0 is
reserved for "0"-spec
- specMap[*it] = std::make_pair(nIndex, rule);
+ //specMap[*it] = std::make_pair(nIndex, rule);
+ specMap[*it] = std::make_pair((int)nIndex,
static_cast<RegExp*>(rule));
}
}
----------------------------------------------------------------------
Comment By: Thomas Wagner (tom68)
Date: 2009-02-13 05:41
Message:
it is the same with sparc and "Sun Ceres C 5.10 SunOS_sparc 2008/10/22"
(SunStudioExpress):
pkgbuild: /opt/SUNWspro/bin/CC -DHAVE_CONFIG_H -I. -g -c -o
parser.o parser.cc
pkgbuild: "parser.y", line 98: Warning (Anachronism): Assigning
std::pair<unsigned, re2c::RuleOp*> to std::pair<int, re2c::RegExp*>.
pkgbuild: "parser.y", line 98: Error: Cannot cast from std::pair<unsigned,
re2c::RuleOp*> to std::pair<int, re2c::RegExp*>.
pkgbuild: 1 Error(s) and 1 Warning(s) detected.
----------------------------------------------------------------------
You can respond by visiting:
https://sourceforge.net/tracker/?func=detail&atid=616200&aid=2535084&group_id=96864
------------------------------------------------------------------------------
All the data continuously generated in your IT infrastructure contains a
definitive record of customers, application performance, security
threats, fraudulent activity and more. Splunk takes this data and makes
sense of it. Business sense. IT sense. Common sense.
http://p.sf.net/sfu/splunk-d2dcopy1
_______________________________________________
Re2c-general mailing list
Re2c-general@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/re2c-general
