You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
In #9686, it was reported that gh start-up time could be improved.
In #9686 (comment), it was discovered that a factor causing slow gh start-up time is related to tcell's init process, and that setting TCELL_MINIMIZE=1 to bypass that process could improve gh start-up time.
This issue is to investigate and identify any other factors involved in a less than optimal gh start-up time and provide a recommendation of what can be done to improve the situation
Expected output
An investigation into factors slowing down gh start-up time
A recommendation on the work required to reduce gh start-up time
The issue is that init functions are executed deepest first, with different import branches having no guarantees between them. Thus, the only way to have tcellinit run after our init would be to have a direct dependency on our init from tcell i.e. forking it. At least that's the only way I can think of it working.
Activity
williammartin commentedon Oct 7, 2024
Before I forget, the existing attempt to set
TCELL_MINIMIZEis just not going to consistently work:cli/internal/build/build.go
Lines 21 to 27 in d645fd4
The issue is that
initfunctions are executed deepest first, with different import branches having no guarantees between them. Thus, the only way to havetcellinitrun after ourinitwould be to have a direct dependency on our init fromtcelli.e. forking it. At least that's the only way I can think of it working.