[root6] Fix call to memset in TpcAvgCurrent.C

[plexoos]

StWarning: StDbSql::QueryDb(table,time) line=439 TpcAvgCurrent  has No data for query
LoadTable: .L /star-sw/StarDb/Calibrations/tpc/TpcAvgCurrent.C
In file included from input_line_983:1:
/star-sw/StarDb/Calibrations/tpc/TpcAvgCurrent.C:6:3: error: no matching function for call to 'memset'
  memset(row, 0, sizeof(TpcAvgCurrent_st));
  ^~~~~~
/usr/include/string.h:62:14: note: candidate function not viable: no known conversion from 'TpcAvgCurrent_st' to 'void *' for 1st argument; take the address of the argument with &
extern void *memset (void *__s, int __c, size_t __n) __THROW __nonnull ((1));
             ^
root4star: .sl79_gcc485/OBJ/StRoot/St_db_Maker/St_db_Maker.cxx:932: virtual TDataSet* St_db_Maker::LoadTable(TDataSet*): Assertion `!ee' failed.
Aborted

[genevb]

I can't even view this directory on github because of all the old files that came and went:
https://github.com/star-bnl/star-sw/tree/main/StarDb/Calibrations/tpc
message from github (not my browser): "This page is taking too long to load."
:-(

[klendathu2k]

I think github must be having a problem at the moment. I was able to review this earlier... but now I see the unfriendly unicorn.

…

On 2023-01-30 11:42, Gene Van Buren wrote: I can't even view this directory on github because of all the old files that came and went: https://github.com/star-bnl/star-sw/tree/main/StarDb/Calibrations/tpc message from github (not my browser): "This page is taking too long to load." :-( -- Reply to this email directly, view it on GitHub [1], or unsubscribe [2]. You are receiving this because you commented.Message ID: ***@***.***> Links: ------ [1] #487 (comment) [2] https://github.com/notifications/unsubscribe-auth/ANL4LVA2OOC3TWZXKGQRIWDWU7VPVANCNFSM6AAAAAAUKR2FKI

[iraklic]

I have a question about this, how did it ever work if it needed &row and had row for many years till now?

[genevb]

On Jan 30, 2023, at 11:54 AM, Irakli Chakaberia ***@***.***> wrote: I have a question about this, how did it ever work if it needed &raw and had raw for many years till now?

Yeah, I was wondering the same, which is why I was trying to look at the history. Indeed, it's been wrong since the initial version: https://www.star.bnl.gov/cgi-bin/protected/viewvc.cgi/cvsroot/StarDb/Calibrations/tpc/TpcAvgCurrent.C?revision=1.1&view=markup Hm. BTW, the old CVS browser also takes a little while to view that directory too, but eventually comes through. https://www.star.bnl.gov/cgi-bin/protected/viewvc.cgi/cvsroot/StarDb/Calibrations/tpc/

…

-Gene

[klendathu2k]

ROOT/CINT is not a C++ compiler. It is a C++ flavored interpreter, laced with several usability hacks... one of which blurs the distinction between objects and pointers. This code runs happily under ROOT5/CINT... struct A_t { int a; int b; int c; }; void testdum() { a.a = 1; a.b = 2; a.c = 3; std::cout << "before questionable memset ... " << std::endl << a.a << " " << a.b << " " << a.c << std::endl; memset( a, 0, sizeof(A_t) ); std::cout << " after questionable memset ... " << std::endl << a.a << " " << a.b << " " << a.c << std::endl; }; And produces the output before questionable memset ... 1 2 3 after questionable memset ... 0 0 0 So... the original code is not correct C++. But under CINT it should work as the author intended. Jason

…

On 2023-01-30 12:31, Gene Van Buren wrote: > On Jan 30, 2023, at 11:54 AM, Irakli Chakaberia ***@***.***> wrote: > I have a question about this, how did it ever work if it needed &raw and had raw for many years till now? Yeah, I was wondering the same, which is why I was trying to look at the history. Indeed, it's been wrong since the initial version: https://www.star.bnl.gov/cgi-bin/protected/viewvc.cgi/cvsroot/StarDb/Calibrations/tpc/TpcAvgCurrent.C?revision=1.1&view=markup Hm. BTW, the old CVS browser also takes a little while to view that directory too, but eventually comes through. https://www.star.bnl.gov/cgi-bin/protected/viewvc.cgi/cvsroot/StarDb/Calibrations/tpc/ -Gene -- Reply to this email directly, view it on GitHub [1], or unsubscribe [2]. You are receiving this because you commented.Message ID: ***@***.***> Links: ------ [1] #487 (comment) [2] https://github.com/notifications/unsubscribe-auth/ANL4LVFYHF5M7URDRCCGL4TWU73HPANCNFSM6AAAAAAUKR2FKI

[plexoos]

The old code worked because in reality a pointer was passed to memset regardless of what the source code said. This is the result of CINT not following the C++ standard but with ROOT6 cling we can detect such leeways and fix them.

@iraklic you can still approve this PR if you feel like it. I was not sure it was worth waiting

[iraklic]

Thank you. That is kind of what I suspected, but I was thinking since we are compiling the code it would pass through the compilator C++ rules and not CINT, but I guess there are details that are not as clear to me in this regard.

[plexoos]

This particular file (actually all .C files in StarDB/Calibrations/) is "interpreted" by CINT/Cling

[iraklic]

This particular file (actually all .C files in StarDB/Calibrations/) is "interpreted" by CINT/Cling

Oh, right, that is the answer that puts things in place. Yes I forgot for a moment about our way with some cint files. Thanks.