Hi, I had taken bsc#1178478 as an opportunity to rework package a bit. After we get Apache/apache2 state into Factory, I think it will be ready for cautious testing. Notable changes: * package is now massively tested with upstream testsuite * update-alternatives are used to provide httpd command and modules under /usr/lib64/apache2 * APACHE_EXTENDED_STATUS="lua" gives something like https://httpd.apache.org/server-status Please report me any oddity you spot. Petr -- Have a lot of fun!
pgajdos wrote:
[...] * update-alternatives are used to provide httpd command and modules under /usr/lib64/apache2
Is that really necessary? update-alternatives is really a terrible tool as it is. We had conflicting subpackages for years just fine right? cu Ludwig -- (o_ Ludwig Nussel //\ V_/_ http://www.suse.com/ SUSE Software Solutions Germany GmbH, GF: Felix Imendörffer HRB 36809 (AG Nürnberg)
On Mon, Dec 14, 2020 at 05:12:46PM +0100, Ludwig Nussel wrote:
pgajdos wrote:
[...] * update-alternatives are used to provide httpd command and modules under /usr/lib64/apache2
Is that really necessary? update-alternatives is really a terrible tool as it is. We had conflicting subpackages for years just fine right?
I do not think it is strictly necessary. Proposals are welcome! Petr -- Have a lot of fun!
On Mon, 2020-12-14 at 17:12 +0100, Ludwig Nussel wrote:
pgajdos wrote:
[...] * update-alternatives are used to provide httpd command and modules under /usr/lib64/apache2
Is that really necessary? update-alternatives is really a terrible tool as it is.
I agree it's terrible, but do we have an "alternative" (sic)?
Martin
--
Dr. Martin Wilck
pgajdos wrote:
On Mon, Dec 14, 2020 at 05:12:46PM +0100, Ludwig Nussel wrote:
pgajdos wrote:
[...] * update-alternatives are used to provide httpd command and modules under /usr/lib64/apache2
Is that really necessary? update-alternatives is really a terrible tool as it is. We had conflicting subpackages for years just fine right?
I do not think it is strictly necessary. Proposals are welcome!
How about making apache2-worker and -prefork conflict? Can't run both at the same time anyways, right? cu Ludwig -- (o_ Ludwig Nussel //\ V_/_ http://www.suse.com/ SUSE Software Solutions Germany GmbH, GF: Felix Imendörffer HRB 36809 (AG Nürnberg)
On Tue, Dec 15, 2020 at 03:54:52PM +0100, Ludwig Nussel wrote:
pgajdos wrote:
On Mon, Dec 14, 2020 at 05:12:46PM +0100, Ludwig Nussel wrote:
pgajdos wrote:
[...] * update-alternatives are used to provide httpd command and modules under /usr/lib64/apache2
Is that really necessary? update-alternatives is really a terrible tool as it is. We had conflicting subpackages for years just fine right?
I do not think it is strictly necessary. Proposals are welcome!
How about making apache2-worker and -prefork conflict? Can't run both at the same time anyways, right?
They can. Petr -- Have a lot of fun!
pgajdos wrote:
On Tue, Dec 15, 2020 at 03:54:52PM +0100, Ludwig Nussel wrote:
pgajdos wrote:
On Mon, Dec 14, 2020 at 05:12:46PM +0100, Ludwig Nussel wrote:
pgajdos wrote:
[...] * update-alternatives are used to provide httpd command and modules under /usr/lib64/apache2
Is that really necessary? update-alternatives is really a terrible tool as it is. We had conflicting subpackages for years just fine right?
I do not think it is strictly necessary. Proposals are welcome!
How about making apache2-worker and -prefork conflict? Can't run both at the same time anyways, right?
They can.
And we want that? If so why do we need some kind of default expressed in the file system via /usr/sbin/httpd? Looks like apache2.service starts the mpm specific binary via the fancy /usr/sbin/start_apache2 anyways? cu Ludwig -- (o_ Ludwig Nussel //\ V_/_ http://www.suse.com/ SUSE Software Solutions Germany GmbH, GF: Felix Imendörffer HRB 36809 (AG Nürnberg)
Martin Wilck wrote:
On Mon, 2020-12-14 at 17:12 +0100, Ludwig Nussel wrote:
pgajdos wrote:
[...] * update-alternatives are used to provide httpd command and modules under /usr/lib64/apache2
Is that really necessary? update-alternatives is really a terrible tool as it is.
I agree it's terrible, but do we have an "alternative" (sic)?
Yeah, for example simply not using it unless there's a really, really good use case. /bin/sh? srsly? cu Ludwig -- (o_ Ludwig Nussel //\ V_/_ http://www.suse.com/ SUSE Software Solutions Germany GmbH, GF: Felix Imendörffer HRB 36809 (AG Nürnberg)
On 12/15/20 3:54 PM, Ludwig Nussel wrote:
How about making apache2-worker and -prefork conflict? Can't run both at the same time anyways, right?
In some use-cases you would want to run two different httpd instances with different MPM models (started by custom systemd units). Ciao, Michael.
On 12/16/20 8:47 AM, pgajdos wrote:
On Tue, Dec 15, 2020 at 04:08:12PM +0100, Ludwig Nussel wrote:
And we want that? If so why do we need some kind of default expressed in
I think yes, see [fate#317786] for example.
Could you please elaborate on this? AFAIK fate tickets are not publicly visible. So it's not helpful to only cite a ticket number on an openSUSE list. Ciao, Michael.
On Wed, Dec 16, 2020 at 09:59:42AM +0100, Michael Ströder wrote:
On 12/16/20 8:47 AM, pgajdos wrote:
I think yes, see [fate#317786] for example.
Could you please elaborate on this?
AFAIK fate tickets are not publicly visible. So it's not helpful to only cite a ticket number on an openSUSE list.
Sure, apache2@.service was introduced trough this fate. Petr -- Have a lot of fun!
participants (4)
-
Ludwig Nussel -
Martin Wilck -
Michael Ströder -
pgajdos