BCP defaults#703
Conversation
37c7156 to
a1a3ae8
Compare
Signed-off-by: Matt McNabb <matthew.mcnabb@vmssoftware.com>
…cked format Signed-off-by: Matt McNabb <matthew.mcnabb@vmssoftware.com>
Signed-off-by: Matt McNabb <matthew.mcnabb@vmssoftware.com>
Signed-off-by: Matt McNabb <matthew.mcnabb@vmssoftware.com>
Signed-off-by: Matt McNabb <matthew.mcnabb@vmssoftware.com>
Signed-off-by: Matt McNabb <matthew.mcnabb@vmssoftware.com>
Signed-off-by: Matt McNabb <matthew.mcnabb@vmssoftware.com>
Previously defaulted to a zero-length value, which is not even valid data for some column types. Signed-off-by: Matt McNabb <matthew.mcnabb@vmssoftware.com>
…etime MSSQL does allow NULL to be sent for non-nullable VARCHAR. Signed-off-by: Matt McNabb <matthew.mcnabb@vmssoftware.com>
The validation code didn't support multiple hints,
and it truncated hints with arguments (e.g. ORDER).
The server will tell us anyway if a hint is invalid
so it doesn't seem necessary to validate at compile-time.
Signed-off-by: Matt McNabb <matthew.mcnabb@vmssoftware.com>
Signed-off-by: Matt McNabb <matthew.mcnabb@vmssoftware.com>
Signed-off-by: Matt McNabb <matthew.mcnabb@vmssoftware.com>
Signed-off-by: Matt McNabb <matthew.mcnabb@vmssoftware.com>
Signed-off-by: Matt McNabb <matthew.mcnabb@vmssoftware.com>
Simulates use of a Format File with missing columns. Signed-off-by: Matt McNabb <matthew.mcnabb@vmssoftware.com>
a1a3ae8 to
d23d94d
Compare
|
Rebased to master. One line of code had to be changed as Note, the first four commits are not to do with BCP defaults; they are fixes to ASE datarows-locking support, including fixing a buffer overflow. Could be merged first before considering the rest of this PR. |
| * in a normal (non-bcp) query, which holds unpacked row data. | ||
| */ | ||
| bindinfo->current_row = NULL; | ||
| bindinfo->row_size = 0; |
There was a problem hiding this comment.
I would just remove the initialization line above.
There was a problem hiding this comment.
OK. In that case perhaps my additional comment lines 202-205 would be better suited to a commit message than a code comment then
There was a problem hiding this comment.
I would like to close this pull request.
I suppose you can post another one about this.
|
Merged first 2 commits (minor style changes). |
|
Merged most of the PR. Just 4 smaller commit left (see open conversations above). |
Cool. I'm away the next two weeks, go ahead as you see fit on the remaining commits if you like. |
This reverts commit 406fc3b. (Alternative commit will follow)
bindinfo->current_row is a buffer for packing a TDS5 BCP out row.
We will allocate space later once we have computed the required size.
Not to be confused with resinfo->current_row which holds unpacked row data.
Signed-off-by: Matt McNabb <matthew.mcnabb@vmssoftware.com>
This reverts commit 9a0d716. Replacement follows.
Signed-off-by: Matt McNabb <matthew.mcnabb@vmssoftware.com>
…pes" This reverts commit 24afcc7. Update follows.
Signed-off-by: Matt McNabb <matthew.mcnabb@vmssoftware.com>
|
Re. the four outstanding commits:
|
At the moment I used a new |
| if (!IS_TDS7_PLUS(tds->conn)) { | ||
| bindinfo->current_row = tds_new(unsigned char, bindinfo->row_size); | ||
| if (!bindinfo->current_row) | ||
| goto cleanup; | ||
| bindinfo->row_free = tds_bcp_row_free; | ||
| } | ||
|
|
There was a problem hiding this comment.
This change was causing a straight forward crash. Fixed in the commit.
|
If there are no objections I would close this PR. |
|
All changes merged |
Changes related to Sybase/ASE BCP and BCP Defaults