> This works well for the first use-case identified. Instead of hard-
> coding the tranche name in the callback, the name can be
> retrieved as the segment name set in GetNamedDSMSegment.
>
> The caller could still pass this name via the extra callback args, but
> it's better to separate things a bit here, and reserve the extra callback
> arguments for more complex data.
>
> What do you think?
As I did not hear back, I went ahead and prepared a patch with the above.
I went back-forth on if it makes sense to provide the name as an
extra argument and decided it provides more flexibility. For example
I can use the same init callback and arguments for different segments.
Also, the name provides a guarantee of the name of the segment that
this callback is initializing.
Overall, I felt it was a better approach.
I updated the code comments and documentation.
Also, in test_dsa.c, i updated the name of the segments to
reflect the name of the functions creating the segments, like
below:
```
@@ -38,8 +38,8 @@ test_dsa_basic(PG_FUNCTION_ARGS)
dsa_area *a;
dsa_pointer p[100];
- tranche_id = GetNamedDSMSegment("test_dsa", sizeof(int),
-
init_tranche, &found);
+ tranche_id = GetNamedDSMSegment("test_dsa_basic", sizeof(int),
+
init_tranche, NULL, &found);
@@ -79,8 +79,8 @@ test_dsa_resowners(PG_FUNCTION_ARGS)
ResourceOwner oldowner;
ResourceOwner childowner;
- tranche_id = GetNamedDSMSegment("test_dsa", sizeof(int),
-
init_tranche, &found);
+ tranche_id = GetNamedDSMSegment("test_dsa_resowners", sizeof(int),
+
init_tranche, NULL, &found);
```
This is good for showing the same init callback being re-used
for different named segment initizations.
I also updated the commit message.
--
Sami Imseih
Amazon Web Services (AWS)