Re: Incompatibility with Tk (Was Re: More release 2.3.0 problems)
Re: Incompatibility with Tk (Was Re: More release 2.3.0 problems)
- Subject: Re: Incompatibility with Tk (Was Re: More release 2.3.0 problems)
- From: Tom Lane <email@hidden>
- Date: Mon, 28 Jul 2008 00:05:58 -0400
- Comments: In-reply-to "Jordan K. Hubbard" <email@hidden> message dated "Sun, 27 Jul 2008 17:54:54 -0700"
"Jordan K. Hubbard" <email@hidden> writes:
> Nope, it's usually neither of those. It's usually some group of (even
> commercial) engineers who absolutely know how painful API transitions
> are but, since they're giving you the code for free, also don't feel
> compelled to do a lot of extra work to make such transitions easier.
After having looked at the code a bit, I have to revise my previous
opinion: adding event types and consequently increasing LASTEvent
is a pretty obvious and predictable form of extension to the protocol.
So it really is Tk's fault to have coded in a way that was brittle
in the face of that change; they could/should have done things slightly
differently and not been vulnerable to crashing.
But the fact remains that if they did it wrong, and nobody's complained
for the fifteen-plus years that Tk's been around, then it's pretty
likely that other clients did likewise. Going slow on introducing the
change seems like a good plan.
regards, tom lane
_______________________________________________
Do not post admin requests to the list. They will be ignored.
X11-users mailing list (email@hidden)
This email sent to email@hidden