HepTuple Package -- Problem Tracking ================================================================== BugNumber: 008 NepNtuple member function columnDefault did not operate as advertised in the documentation. Initial Bug Report 21 Oct. 2001 ------------------ Version 2.0.9 Default values for ntuple columns were being set properly for the initial instantiation but not for a re-instantiation. columnDefault was supposed to do this but didn't. Instead of setting a new default value, it attempted to return the nonexistent "current" default value and promptly crashed on a seg fault. - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - Bug Reproduced 21 Oct. 2001 -------------- Any attempt to update an existing ntuple and supply default values for one or more columns. - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - Repair Plans ------------ Leave columnDefault as a function to *return* current default values and fix it so it doesn't crash. In addition, define a new function, setColumnDefault to *supply* new default values, making good on our original design. - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - Status of repair 25 Oct. 2001 ---------------- Done in jm's test release. Will appear in the cvs repository at the same time this notice does. - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - Problem Rectified 25 Oct. 2001 ----------------- version 2.0.10 - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - Ongoing Caveats --------------- None beyond noting that the code now behaves as had been advertised the whole time. ================================================================== ================================================================== BugNumber: 007 Odd behavior in the Root file manager destructor Initial Bug Report 1 Oct. 2001 ------------------ Version 2.0.5 This started while testing for the problem described in under BugNumber 006. If one releases some object prior to the end of the job, the output file gets screwed up such that on subsequent attempts to update the file messes up the cycle numbers. - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - Bug Reproduced 2 Oct. 2001 -------------- Any attempt to update a Root output file when any object is released prior to job end. The problem also occurs on the second update attempt even without any release. - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - Repair Plans ------------ Restructure the Root file manager destructor. This is complicated because the implementation went somewhat beyond the original design. - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - Status of repair ---------------- The Root file manager has been restructured but still needs work because most objects are being written out twice. All of this also revealed a bug in Root that must be fixed as well. - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - Status of repair ---------------- Done. Code now behaves as the documentation says and objects are now being written only once. - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - Problem Rectified 10 Oct. 2001 ----------------- version 2.0.6 - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - Ongoing Caveats --------------- None beyond noting that the code now behaves as it should have the whole time. Anybody depending on the former (broken) behavior is going to get a surprise. Since nobody complained, I doubt that this will happen. ================================================================== ================================================================== HepTuple 006: Removed an un-needed test in HBookNameTag.cc. HBook column names are *not* limited to c chars. Initial Bug Report 08/14/01 ------------------ Version 2.0.5 No specific report. This possibility turned up while investigating a request from Avi Yagil about duplicate column names in different blocks. - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - Bug Reproduced 08/14/01 -------------- Yes, in a private version of hepToyz - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - Repair Plans ------------ Remove a test in isValidColumnName() in HBookNameTag.cc - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - Status of repair ---------------- Done and checked in. - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - Problem Rectified 08/15/01 ----------------- version 2.0.6 - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - Ongoing Caveats --------------- ================================================================== HepTuple 005: Put in a test to make sure index variables do not exceed their advertised span. ZMthrow if they do. Initial Bug Report 07/24/01 ------------------ Version 2.0.5 No specific report. This possibility turned up while investigating HepTuple 004, reported below. - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - Bug Reproduced 07/25/01 -------------- Yes, in a private version of hepToyz - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - Repair Plans ------------ Put a test in captureValue in ColumnT. - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - Status of repair ---------------- Done and checked in. - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - Problem Rectified 07/26/01 ----------------- version 2.0.6 - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - Ongoing Caveats --------------- ================================================================== HepTuple 004: Fix a logic error in HepHBookFileManager::cd( target ) Program failed when target was // and we were already in //. Initial Bug Report 07/18/01 ------------------ Version 2.0.5 User reported very strange error messages from ZMthrow which suggested a very different problem! Go figure. - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - Bug Reproduced 07/19/01 -------------- Sitting in some subdirectory, execute manager->cd("//"); twice. On the second call, the program ZMthrew with the misleading error message. - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - Repair Plans ------------ When sitting in // and being asked to cd("//"), replace "//" with "//" + (manager->rootDir()) - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - Status of repair ---------------- Done and checked in. - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - Problem Rectified 07/20/01 ----------------- version 2.0.6 - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - Ongoing Caveats --------------- ================================================================== HepTuple 003: Trap missing span for index variables Initial Bug Report 07/08/01 ------------------ Version 2.0 User reported very strange entries in Root histograms made from ntuple. - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - Bug Reproduced 07/16/01 -------------- Omitted call to span(,,) for variable-length column array. While strictly a user error, jobs appear to run normally but put trash into ntuple buffer. - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - Repair Plans ------------ Insert test for no-span arrays and ZMthrow. - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - Status of repair ---------------- Done - (for the most part) There are numerous capture methods - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - Problem Rectified ----------------- version 2.0.1 - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - Ongoing Caveats --------------- The error may happen again under obscure circumstances because of the multiplicity of capture methods. ================================================================== HepTuple 002: Warning message when compiling with ROOT 3.xx Initial Bug Report 06/06/2001 ------------------ Version 2.0.1 In ROOT 3.xx a few methods have been made 'const' and are over-ridden by HepTuple. This now provokes a justified warning messages. - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - Bug Reproduced ---------- -------------- Yes. - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - Repair Plans ------------ Update the header and source file to add 'const' when ROOT 3.xx is detected. - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - Status of repair ---------------- Done. - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - Problem Rectified 06/11/2001 ----------------- version 2.0.3 ================================================================== ================================================================== HepTuple 001: Incorrect version of LIBDEPS file Initial Bug Report 04/24/2001 ------------------ Version 2.0.0 Dave Ritchie and Paul Russo reported that the LIBDEPS files that are used for DZero package dependencies tracking were stale. - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - Bug Reproduced ---------- -------------- Bug not reproduced per se. Those files are tested and maintained by DZero. - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - Repair Plans ------------ Import files provided by Paul Russo. - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - Status of repair ---------------- Done. - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - Problem Rectified 04/24/2001 ----------------- version 2.0.1 ================================================================== ================================================================== Template for each bug: ================================================================== BugNumber: BugName Initial Bug Report ------------------ Version x.y * * * - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - Bug Reproduced -------------- circumstances * * - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - Repair Plans ------------ * * (or technical problems) * - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - Status of repair ---------------- * * * - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - Status of repair ---------------- * * (in long-running problems might need more than one) * - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - Problem Rectified ----------------- version x.z - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - Ongoing Caveats --------------- * * * ==================================================================