Closed GitMensch closed 1 year ago
For the PostgreSQL driver, this should be solved in the v1.0.20dev branch with b859affdff7aa5769ed7a834a5c57aa221ec86ff and ad62ffb1bd3658fdce118076362944f570096f6f
Still working on the implementation for the other drivers.
Solved for the other drivers in the v1.0.20dev branch and in the post-merge main branch, at least for binary data.
UTF16 was not tested. Given that I wanted to add and/or test specific support for correctly handling different charsets I am probably testing and adding specific support after the release of v1.0.20. I will open a new charset-specific issue and close this one.
Any idea when the first 1.0.20 preview release is available? BTW: GC's 3.2rc1 should be today.
Any idea when the first 1.0.20 preview release is available?
I am re-running the packaging and test scripts, there are a few things that need fixing. I guess a pre-release (packages) could be coming by the beginning of next week. Then I need to update the documentation and write the new docs for the test suite, that will require a few more days, depending, as usual, on my day-job.
The v1.0.20dev branch itself is now building from source on Linux, I will complete this round of build-related fixes in the v1.0.20dev branch and, as soon as binary packages can be successfully built and the tests pass inside my Github Actions workflow, I will re-merge this set of modifications into the main branch: unfortunately is a time-consuming process, with many small adjustments here and there.
unfortunately is a time-consuming process, with many small adjustments here and there
Yes, I totally know how that feels :-)
This is nice :smile:
Using a code like
Can not work in libgixsql-pg (possibly also in others) because
DbInterfacePGSQL::_pgsql_exec_params
usespvals[i] = (char*)paramValues.at(i).c_str();
- where the return would be be an empty string because of the leading low-value.This would also be very bad for utf16 content like n"35" - which is x'00330035`.
Note: I do not need that for a current project (actually
PIC N / PIC U
is to be tested "soon" but I guess that's not supported either at all [it seems to be supported "in general" in gixpp and libgixsql'sSqlVar
] or in the PG driver), but stumbled over it during debugging.If this isn't likely to come soon I'd suggest to leave a note in the README about the current support of:
It would be totally reasonable to only allow binary data with a special
USAGE
orEXEC SQL VAR
definition. That would obviously be given forPIC U
andPIC N
.