Skip to content

Possible fix for #8734 #8735

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 5 commits into from
Sep 14, 2025
Merged

Possible fix for #8734 #8735

Changes from 1 commit
Commits
Show all changes
5 commits
Select commit Hold shift + click to select a range
c6de939
Possible fix for #8734
aafemt Sep 8, 2025
38da83d
Use explicit pool
aafemt Sep 8, 2025
df1717b
Use HalfStaticArray as demanded
aafemt Sep 9, 2025
f727114
Use begin() instead of getBuffer()
aafemt Sep 10, 2025
70eaf53
Combine conditions as requested
aafemt Sep 13, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
27 changes: 18 additions & 9 deletions src/dsql/DsqlRequests.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -596,19 +596,28 @@ void DsqlDmlRequest::doExecute(thread_db* tdbb, jrd_tra** traHandle,

// if this is a singleton select that return some data, make sure there's in fact one record

if (singleton && (request->req_flags & req_active) && outMsgLength > 0)
if (singleton)
{
// Create a temp message buffer and try one more receive.
// If it succeed then the next record exists.
if (outMsgLength > 0)
Copy link
Member

@hvlad hvlad Sep 13, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I don't understand why there is check for outMsgLength > 0 - is it possible to be false when singleton == true ?
Even if out message with zero length is possible (???), what should happens for the singleton && outMsgLength == 0 case ?
This place is very unclear for me.
In v5 there was just check for singleton == true, btw

Copy link
Contributor Author

@aafemt aafemt Sep 13, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Just in case. If the function was called with outMetadata == nullptr - just assume success. The difference from v5 is that buffers here are provided by application rather than of engine itself.

Copy link
Member

@hvlad hvlad Sep 13, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Is it possible to have singleton == true && outMetadata == nullptr ?
In any case, why two nested if here ?

Copy link
Contributor Author

@aafemt aafemt Sep 13, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Initially check for request's activity was out of inner if. Then I realized that some BLR code can intentionally run to the end without entering blr_send and an application aware of it can provide no output buffer.

If you are sure that I'm wrong feel free to change this piece later.

Copy link
Member

@hvlad hvlad Sep 13, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I'm sure that this code is not clear. And you changed it this way.
Formally,

if (a) {
  if (b) {
...
  }
}

is the same as

if (a && b) {
...
}

thus I ask why you choose first one. If there is no real reason, change it.
If there is a reasons - provide it, and put in comments to avoid similar misunderstanding for the future readers.

{
// No record returned though expected
if (!(request->req_flags & req_active))
{
status_exception::raise(Arg::Gds(isc_stream_eof));
}

std::unique_ptr<UCHAR[]> message_buffer(new UCHAR[outMsgLength]);
// Create a temp message buffer and try one more receive.
// If it succeed then the next record exists.

JRD_receive(tdbb, request, message->msg_number, outMsgLength, message_buffer.get());
std::unique_ptr<UCHAR[]> message_buffer(new UCHAR[outMsgLength]);
Copy link
Member

@hvlad hvlad Sep 8, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

HalfStaticArray, pls

Copy link
Contributor Author

@aafemt aafemt Sep 8, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Size of the message is known at this point, so usage of HalfStaticArray will just waste space on stack or cause stack overflow if this size is used as InlineCapacity.

Copy link
Member

@hvlad hvlad Sep 8, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Do you think that outMsgLength could be used as InlineCapacity template argument ?
Looks like you not understand purpose and usage of HalfStaticArray.

Copy link
Contributor Author

@aafemt aafemt Sep 8, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Do you think that outMsgLength could be used as InlineCapacity template argument ?

No.

Looks like you not understand purpose and usage of HalfStaticArray.

Yes. Especially in this place which is executed exactly once per query.


// Still active request means that second record exists
if ((request->req_flags & req_active))
{
status_exception::raise(Arg::Gds(isc_sing_select_err));
JRD_receive(tdbb, request, message->msg_number, outMsgLength, message_buffer.get());

// Still active request means that second record exists
if ((request->req_flags & req_active))
{
status_exception::raise(Arg::Gds(isc_sing_select_err));
}
}
}
}
Expand Down
Loading