1 |
On Wed, May 14, 2003 at 03:47:45PM +0200, foser wrote: |
2 |
> On Wed, 2003-05-14 at 11:14, Robin H.Johnson wrote: |
3 |
> > I've almost completed the new PHP eclasses for mod_php and CLI php. |
4 |
> > I just wanted to discuss a possible new global USE flag. |
5 |
> > |
6 |
> > informix - Provides support for IBM Informix server |
7 |
> > |
8 |
> > This would compliment the current 'oci8' for Oracle. |
9 |
> > Once we also get other packages for informix, then will greatly increase |
10 |
> > our appeal to the enterprise market :-). |
11 |
> |
12 |
> In my opinion USE flags should stay local as long as there's not |
13 |
> multiple uses for it currently available (or pending) in the tree . I |
14 |
> don't know if that's the case here (taking into consideration php and |
15 |
> mod_php are basically the same package). |
16 |
Informix is supported by |
17 |
dev-java/jdbc-informix |
18 |
net-www/cocoon |
19 |
|
20 |
But neither of those need the informix USE flag, however Informix is |
21 |
supported in general. |
22 |
|
23 |
> > We want to push this to a USE flag as it does cause a visible |
24 |
> > performance hit for the increase in security it provides. |
25 |
> > |
26 |
> > From a little research on Google, it seems that both TeX and PostGres |
27 |
> > support memory limit stuff to some degree (I did not dig much beyond the |
28 |
> > first page linked off google). |
29 |
> wouldn't be a general security type of flag better ? In general it seems |
30 |
> like the sort of thing flagged. |
31 |
Any suggested names? Postgres and TeX seem to have it because they can |
32 |
get big easily, not for security reasons. |
33 |
|
34 |
> > Given the utter lack of response to Tal's previous email to -core, if I |
35 |
> > don't get any answer on these by saturday, I'll look at including them |
36 |
> > myself as they are needed for the new PHP ebuilds. |
37 |
> see it like this, other devs don't deal with your packages and may be |
38 |
> careful to give their opinion about something they haven't looked into |
39 |
> (i know i am). |
40 |
I don't expect that they have looked into my packages deeply, but I was |
41 |
looking for response regarding these flags and their packages (which I |
42 |
haven't looked deeply at). |
43 |
|
44 |
> In general i was looking trough USE flags list, things like icc and |
45 |
> jikes and memlimit maybe, do they really belong in use.desc ? Aren't USE |
46 |
> flags defined as compile time enabling or disabling of certain features |
47 |
> in an application, i guess memlimit fits the description, but icc and |
48 |
> jikes don't, theyre compile time options that have no direct effect on |
49 |
> the functionality of the app. Shouldnt they have their own file, |
50 |
> removing some of the burden of the use.desc file ? |
51 |
Is the performance of the app not part of it's functionality? If so, |
52 |
then icc has a use. Likewise for jikes. |
53 |
|
54 |
-- |
55 |
Robin Hugh Johnson |
56 |
E-Mail : robbat2@××××××××××××××.net |
57 |
Home Page : http://www.orbis-terrarum.net/?l=people.robbat2 |
58 |
ICQ# : 30269588 or 41961639 |
59 |
GnuPG FP : 11AC BA4F 4778 E3F6 E4ED F38E B27B 944E 3488 4E85 |