Hi,

I am a bit confused with the FPC version numbers in SubVersion. Obviously I just don't know or understand how the version numbers in FPC work. I do know even numbers are stable and odd numbers unstable. The version increments is still a mystery.

Why has the /Fixes_2_2 branch got a version number 2.1.3? The name is very misleading. Why not 2.1.2 or 2.2.0? Or is 2.2.0 reserved for the day it actually gets released?

What is the version number of /trunk now? 2.2.1? 2.1.4?

Graeme Geldenhuys wrote

Why has the /Fixes_2_2 branch got a version number 2.1.3? The name is very misleading. Why not 2.1.2 or 2.2.0?

Even version numbers are reserved for releases (either internal/ external releases), i.e. they correspond to a single source tree which is tagged. 2.1.2 has been tagged already for internal release building testing. 2.1.4 will be the first public beta for 2.2

Or is 2.2.0 reserved for the day it actually gets released?

Yes.

What is the version number of /trunk now? 2.2.1? 2.1.4?

2.3.1

Jonas

Graeme Geldenhuys wrote:

Hi,

I am a bit confused with the FPC version numbers in SubVersion. Obviously I just don't know or understand how the version numbers in FPC work. I do know even numbers are stable and odd numbers unstable. The version increments is still a mystery.

;-) Actually, I believe the scheme is described on our WWW pages (FAQ?).

Why has the /Fixes_2_2 branch got a version number 2.1.3? The name is very misleading. Why not 2.1.2 or 2.2.0? Or is 2.2.0 reserved for the day it actually gets released?

Yes, 2.2.0 will be released later based on sources in that branch (after proper testing/stabilization). 2.1.2 is being prepared as a beta release right now, so the continuing updates are getting to 2.1.3.

What is the version number of /trunk now? 2.2.1? 2.1.4?

2.3.1 (i.e. further progress beyond 2.2.x).

Tomas

Ah, thanks for that. I didn't know about internal releases, which is what threw me off. ;-)

> What is the version number of /trunk now? 2.2.1? 2.1.4? 2.3.1

Not 2.2.1? ;-) I guess that's to many even numbers to justify a unstable release.

Thanks for the information!

Op Tue, 27 Mar 2007, schreef Graeme Geldenhuys:

> 2.3.1 Not 2.2.1? ;-) I guess that's to many even numbers to justify a unstable release.

When 2.2 is out, the fixes_2_2 branch will be named 2.2.1. Trunk will ultimately become FPC 2.4.

Daniël

2.2.1 is the version number of the fixes_2_2 branch between the 2.2.0 and 2.2.2 releases :)

Anyway, I'm looking forward to beta/stable versions from the 2.2.x branch, and in particular generics support, which will be enabled in 2.2.x I presume?

Regards,

Bram

Bram Kuijvenhoven wrote

Anyway, I'm looking forward to beta/stable versions from the 2.2.x branch, and in particular generics support, which will be enabled in 2.2.x I presume?

Yes, but it still contains several known bugs such as cryptic error messages, context sensitivity (if procedure Foo is declared in units A and B, and the unit declaring the generic which calls Foo uses A while a unit declaring a specialisation uses B, then the specialisation will call B.Foo instead of A.Foo), and a number of tests do not yet work.

Jonas

Indeed, we should make it clear that we reserve the right to change/tweak the semantics a bit ? depending on how we solve that context sensitivity problem, and perhaps others.

Micha

Maybe simply note that this feature is still in alpha stage, that covers the state pretty well (not yet finalised feature set/ semantics, known bugs).

Jonas

FYI: added the context sensitivity test to svn (example how you could get "bitten"):

http://www.freepascal.org/cgi-bin/viewcvs.cgi/trunk/tests/test/tgeneric14.pp?view=markup http://www.freepascal.org/cgi-bin/viewcvs.cgi/trunk/tests/test/ugeneric14.pp?view=markup

The test should print "Foo", but does not.

Micha