Issues with 2.6.0 release candidates
This page is supposed to contain all issues discovered while testing release candidates and their status. Make sure to include information about the platform where the problem appears (if applicable). Once the issue gets resolved, you should strike it over and add number of SVN revision fixing it. If the revision number refers to a SVN repository different from "fpc" (i.e. "fpcbuild" or "fpcdocs"), include information about the SVN repository too.
cpu-os: IE209101221 while compiling XYZ.pas- r123456
- docs: missing description of XYZ
readme.txt: wrong version number- r654321 (fpcbuild)
Merge requests for post RC1 fixes so far
strikken means merged, bold means decisions/review/actions needed.
r19515 already merged (typo in winunits-base shlobj.pp) r19380 bug in exesearch r19485 fpunits.conf->.cfg r19486 safecall on win64 flag in i_win.pas
- r19525 Fix problem with QueryHook leading to crashes
r19526,r19531 haiku (confirmed by Olivier) r19342 nativent system unit (nativent dir only)
- Jonas and Florian would discuss what to do with r18961/r19312/r19313 ? -> still ongoing
r19497 check environment variable LOCALAPPDATA on windows as a possible cache location for instantfpc r19339 (Francesco) Updated libogcfpc to ver.1.8.8 r19493 deprecation msgs of tdbf. Can starting compiler handle this? Does that matter at all for non- compiler/RTL? -> 2.4.4 can
- r19328 byte typecast in set declaration gives syntax error (Jonas remarked it has a mantis item target 2.6.0 Mantis mantis 20380 by Paul) conflict on merge due to "not typeonly" introduced by r19148
r19343 mantis 19975 (classhelper, no commit msg) -> looks complex, so initially I'd say deny; Comments? (mostly under ifdef extended recorddef) 19342: http://svn.freepascal.org/cgi-bin/viewvc.cgi?view=rev&revision=19292 ) 19343: http://svn.freepascal.org/cgi-bin/viewvc.cgi?view=rev&revision=19281 19343: http://svn.freepascal.org/cgi-bin/viewvc.cgi?view=rev&revision=19282
- Class helpers are the new feature added by Sven. In my opinion it is bad to make a release with the new feature and leave the bugs unfixed if they already fixed somewhere else--Paul Ishenin 12:23, 30 November 2011 (CET)
- Merging is always a tradeoff of risks vs benefits.
OS/2 merging, many revs, see  but hinges on r18138 causes many conflicts. (also r19551,r19552) 18138 partially merged, but not marked as merged. Merging 18136 was also needed. 19551-19552 merged
Paul: Fixed today a serious bug which may cause compiler crashes: r19563 Without the change any record member symbol was treated as a tfieldvarsym and tfieldvarsym members where accessed without a check. fpcbuild r918,926,928
- open tickets with 2.6.0 as target
19684: fixes internal error in handling of dynamic arrays 19694: fixes compiler crash due to accessing potentially unassigned field 19605 (issue 0020638), 19641 (issue 0020690, issue 0020038), 19681 (issue issue 0020721): issues with nested types in classes. Looks more complicated, so I committed it separately( different day ->testsuite)
Issues with 2.6.0-rc1
- (potential) the usage of the heapmanager in a thread before it is initialized (Sergei, fpc-devel, nov 10th "Threads crash on freebsd")