feat: Double Ctrl+C to close pgxcli

[LeeFred3042U]

Closes #75

Changes

internal/app/ui/model.go

• Added ctrlcSeq int field to Model to track the current Ctrl+C seq generated
• In Update():
  • updated the StatePending guard to also exclude Interrupt key
  • updated seqTimeoutMsg case to also check ctrlcSeq
  • updated Interrupt handler inside tea.KeyMsg block:
    • first Ctrl+C in StateInput sets StatePending and fires a tea.Tick
    • second Ctrl+C within 500ms dispatches QuitRequestMsg and returns to StateInput
    • Ctrl+C during StateExecuting remains instant cancel, no double press required

Notes

• No changes to keys.go : Interrupt is already connected to ctrl+c
• Reused StatePending and seqTimeoutMsg introduced in feat: double ESC to clear the input #73

[balajz]

Hi, since both share the same state, pressing Esc and Ctrl+C will close the pgxcli.

How about using a global sequence counter and splitting the states?

escSeq   int
ctrlcSeq int

to

pendingSeq int

const (
    StateInput State = iota
    StateExecuting
    StatePendingClear
    StatePendingQuit
)

[LeeFred3042U]

@balajz ok

[LeeFred3042U]

@balajz

Changes

internal/app/ui/model.go

• Replaced StatePending with StatePendingClear and StatePendingQuit in the State iota
• Replaced escSeq int and ctrlcSeq int with a single pendingSeq int field on Model
• In Update() I updated:
  • the pending guard to check StatePendingClear || StatePendingQuit instead of StatePending
  • the seqTimeoutMsg case which is now simplified to a single msg.seq == m.pendingSeq check since states no longer overlap
  • the Interrupt handler:
    • first Ctrl+C in StateInput sets StatePendingQuit and fires a tea.Tick
    • second Ctrl+C within 500ms dispatches QuitRequestMsg and returns to StateInput
    • Ctrl+C during StateExecuting remains instant cancel, no double press required
  • the Clear handler to use StatePendingClear instead of StatePending

Note: When running select pg_sleep(5);, when ctrl+c is pressed it stops the program, heres the statement, what should be done here?

postgres@localhost:postgres> select pg_sleep(5);                       
✗ ERROR: canceling statement due to user request (SQLSTATE 57014)

[balajz]

It looks like you might have forgotten to push the changes you mentioned in your latest comment!

[LeeFred3042U]

@balajz mb
some conflicts occurred but now it works
Should i squash the commits?

[balajz]

@balajz mb some conflicts occurred but now it works Should i squash the commits?

thanks for the contributions. I will squash the commits.