Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Repo info
Activity
adeshr
@adeshr

@vrajat doubt about test:
If we try to connect server, without migrating and setting up the table , the connection wont be created, since there is no data in db. But if we do migration before using flyway, then we wont be able to check if server succesfully ran the migration or not.

So, what I think is, we can call DriverManager.connect in a try/catch block and catch the error, since there is no data in db it is bound to fail when we create DataSourceSchema. But Migration would have run by then and we can test the schema-version table.

is the approach ok?

vrajat
@vrajat
It should be created
If the da
if the db is empty, the correct behaviour is to migrate and create all the tables. But the there will be no rows
So quark can't do anything. That's okay in the test
Connect should succeed. If it's throwing an exception then it's worth an investigation
adeshr
@adeshr
but connection wont get created, since there is no data in db
server would be up, but DriverManager.connect will fail
vrajat
@vrajat
Because there are no rows in the data source table ?
I still don't understand why connection should fail just because there is no data in the catalog
adeshr
@adeshr
Because there is no data in data source table
Amogh Margoor
@amoghmargoor
I think DriverManager.connect should not fail even if there is no data in catalog ... might be something we missed ... we need to fix it
after DDL support ...most people will start empty itself and are expected to fill data later
so we will have to fix it eventually
vrajat
@vrajat
Yes
It's valid to start with no data and even no tables
if there are no tables , migration should take care
if there is no data DDL statements will be used to add it
vrajat
@vrajat
@adeshr still confused by the expected behaviour ?
adeshr
@adeshr
no, understood it...but thinking about if I can do it together with server or we need to create a new issue?
vrajat
@vrajat
write the test. lets see
adeshr
@adeshr
wrote the test already, connection is failing, since we dont have data. Maybe we can use @ignore on test, and will remove it once db catalog have the required functionality?
vrajat
@vrajat
add ignore and push it.
I’ll check why
then we can decide if we want to take it up in a follow up jira or not.
adeshr
@adeshr
sure
vrajat
@vrajat
though I think it has to go in this sprint.
or with DDL statements.
adeshr
@adeshr
@vrajat pushed the test.
vrajat
@vrajat
ok
adeshr
@adeshr
@vrajat json catalog is also supported with server
vrajat
@vrajat
oh right
remove that sentence
adeshr
@adeshr
sure
adeshr
@adeshr
#32 is ready for merge.
adeshr
@adeshr
@vrajat @amoghmargoor need to understand how should I start #36 ?
vrajat
@vrajat
@amoghmargoor is your best bet @adeshr :)
adeshr
@adeshr
sure, will ask him
vrajat
@vrajat
@adeshr @amoghmargoor feed back on the current git flow of rebase and “merge” from PR ?
should we go back to squash merge ?
vrajat
@vrajat
@amoghmargoor #43 is finally ready for a merge
adeshr
@adeshr
:+1: for rebase, no need for command line
Amogh Margoor
@amoghmargoor
@vrajat Will merge #43
+1 for merge using PR... That is easier
vrajat
@vrajat
ok. thx
vrajat
@vrajat
@adeshr are these files required ? https://github.com/qubole/quark/tree/master/server/src/main/resources
for e.g. dbCatalog.json is not maintained
adeshr
@adeshr
dbCatalog.json is not required. must have missed removing it.
Logback.xml can also be removed, the server will pick it from fatjdbc/main/resources
vrajat
@vrajat
ok. I’ll clean them up
adeshr
@adeshr
sure