Estimated Go 1.0 release date?
Erik Unger
precise than "early 2012" ?
Will there be another point release before 1.0?
-Erik
Brad Fitzpatrick
tux21b
Graham Anderson
> When do the holidays end? I am missing the huge amount of CLs and
> improvements every day :)
Which holidays? *ducks*
Graham Anderson
On that note, do you think exp/sql will be in the main tree for v1? I'm not
assuaged either way but us packagers are busy beavers and we try to be
advocates for the lang. Right now we have a "package everything", i'd
personally like to be more selective and exp/sql + drivers seems like it might
be going places.
Kyle Lemons
André Moraes
The library should mature a little more before going into the stdlib
and I think that the time for go 1 release will not give the library
that time.
--
André Moraes
http://andredevchannel.blogspot.com/
Brad Fitzpatrick
André Moraes
receive more feedback from other people.
If the library goes to "Go 1" changes could become more complex,
that's what I meat by "mature a little more".
Maxi
Go into a C program! (I have a C webserver, but I want to use Go for
"scripting". If I were to use FastCGI, I could just as well completely
switch to Go, which is not possible for me.)
ps
It's nice to see a serious answer to this question, as opposed to the
"When it's ready" sort of answers seen in other communities.
What about the notion of keeping the core of Go lean and clean,
similar to Node? It's a slippery slope and there's always something
else that "absolutely must be" included, based on personal
preferences. A steady hand keeps the project on course. Ignore the
screamers, they're like junkies crying for a fix.
http://www.sporcic.org/2011/10/nodejs-is-the-new-java/
John Asmuth
On Monday, January 2, 2012 2:03:07 PM UTC-5, ps wrote:
What about the notion of keeping the core of Go lean and clean,
similar to Node?
Florian Weimer
> How do you think it needs to mature?
What does this mean?
| DB is a database handle. It's safe for concurrent use by multiple
| goroutines.
If Exec() &c are potentially blocking, then this safety is not very
useful. Such blocking will not be known to the database, and the
wait-for graph it determines will lack edges, so deadlock detection
will fail.
If these operations do not block, you cannot use some PostgreSQL
features API because it seems to me that DB needs to be backed by a
connection pool, and some PostgreSQL features are not available inside
transactions.
The API does not offer a way to tell serialization failures from other
types of errors. Correct operation is not possible without
transaction retry on serialization failure.
Florian Weimer
> All the major Go 1 language & library changes are now in. Bug fixing &
> polish begins after the holidays, then some release candidates will start
> going out. The latest weekly, 2011-12-22 is basically what Go 1 will be,
> if you want to start playing.
But will there be another release (in the "r61" sense) before that, as
some sort of release candidate? Presumably, this would be packaged
for Debian unstable/testing and other distribution, and result in more
widespread testing.
