Changes-File & caching

Hello,

I reported a bug I found in
https://pharo.fogbugz.com/default.asp?10849_d0g95e46s78671vi

The issue was closed in fogbugz. I couldn't really figure out why, until I
looked at Penelope to find out that bug was considered a duplicate.

I had a hard time finding that information and the information on the
duplicate (I could figure it out yesterday night, but can't reproduce the
way to find the information today). From what I understood yesterday, this
bug is supposed to be fixed in Pharo-3.0?

The images that are offered for downloading seem to all include this bug. I
have a notebook with ubuntu-13.04 and 8GB RAM, of which around 7GB are used
for hard disk caching. As far as I understood from reading the bug tracking
entries, heavy caching triggers the bug earlier?

Anyhow, with this bug, the changes file gets garbaged (or, at least, the
pointers to it) and the image is unuseable. Abandoning the RemoteString
error message looks to the user as if the issue wasn't all that bad, but is
a nuclear attack on the image soon to come.

Correct me if I am wrong please.

About issue tracking: is the process closed to the public on purpose? I have
to admit I misused the google code bugtracker in the past, because I
reported an incident from the user point of view and later learned, that the
bug reporting & fixing process was meant to be used only from a developers
perspective.

When I have a "user" issue - am I supposed to report to the mailing list
instead and an issue will be created by the pharo core team?

Kind regards,
   Markus

On Jun 7, 2013, at 10:26 AM, Markus Fritsche <[hidden email]> wrote:

> Hello,
>
> I reported a bug I found in
> https://pharo.fogbugz.com/default.asp?10849_d0g95e46s78671vi
>
> The issue was closed in fogbugz. I couldn't really figure out why, until I
> looked at Penelope to find out that bug was considered a duplicate.
>

Normally I thought that fogbugz should send you updates on bugs you reported?
We need to check that.

> I had a hard time finding that information and the information on the
> duplicate (I could figure it out yesterday night, but can't reproduce the
> way to find the information today). From what I understood yesterday, this
> bug is supposed to be fixed in Pharo-3.0?
>
In this case, we have fixed in in 3.0 and 2.0 (the bug as a "Parent Case" for 2.0,
this was decided in the discussion and done).
We often fix things first in 3.0 to test and then fix it in 2.0 so we are sure the fix
works.

> The images that are offered for downloading seem to all include this bug.

The problem is that the 2.0 files referenced directly on pharo-project.org are pointing
to the last *release*, while the fixes we do in 2.0 are only in the nightly builds that you can
find on http://files.pharo.org/image/20/ for example

I already was thinking that this is a problem, we need to discuss what to do.

Maybe we should instead have pharo-project.org point on the latest 2.0?


>
> Anyhow, with this bug, the changes file gets garbaged (or, at least, the
> pointers to it) and the image is unuseable. Abandoning the RemoteString
> error message looks to the user as if the issue wasn't all that bad, but is
> a nuclear attack on the image soon to come.
>
So there are two problems:

1) the "RemoteString past end of file" on recent linux.

2) Something strange happens sometimes that reading the source fails.

     https://pharo.fogbugz.com/f/cases/7559/Plugins-Primitives-no-longer-work-after-a-while
     (it has a parent case for 2.0)


> About issue tracking: is the process closed to the public on purpose?

FugBugz does not allow read access for anonymous users. We *really* don't like this
but it can't be changed. That is why we added bugs.pharo.org where you can get an
account immediately *and* view all issues

Maybe this is not enough and we need to abandon Fogbugz because it just is "too closed"?

> I have
> to admit I misused the google code bugtracker in the past, because I
> reported an incident from the user point of view and later learned, that the
> bug reporting & fixing process was meant to be used only from a developers
> perspective.
>
Not really…. but it is difficult to do both in the same system. Maybe we need
two bug trackers? One for bugs and one to funnel through improvements and fixes?

> When I have a "user" issue - am I supposed to report to the mailing list
> instead and an issue will be created by the pharo core team?
>

No, normally the idea is that people should report bug entires.

But one thing that is always a problem is man-power: Not many people look at the bug
tracker. (To me it always feels as if I am the only one…).

This means in turn, we need to sometimes deal with bugs in a way that might not be perfect.
The only way to make it better is if more people share the load. I don't see any other way.

        Marcus

Hi Markus,

On 07 Jun 2013, at 10:26, Markus Fritsche <[hidden email]> wrote:

> Hello,
>
> I reported a bug I found in
> https://pharo.fogbugz.com/default.asp?10849_d0g95e46s78671vi
>
> The issue was closed in fogbugz. I couldn't really figure out why, until I
> looked at Penelope to find out that bug was considered a duplicate.

Yes, I closed the bug because it was a duplicate.

Yes, it is an important problem (but the trigger is a change in Ubuntu 13.04). I am not an Ubuntu desktop user, I can only hope that those who are will help.

Pharo is open source and developed in the open, after the switch to Fogzbug things are not yet perfect: it is too hard for casual users. This will improve slowly.

Rest assured: you are helping a lot by reporting issues from a user perspective !

Thanks, keep pushing !

Sven

Hello Sven & List,

On 07.06.2013 10:45, Sven Van Caekenberghe wrote:
> Yes, it is an important problem (but the trigger is a change in Ubuntu 13.04). I am not an Ubuntu desktop user, I can only hope that those who are will help.
Well, I have 13.04 on a few laptops (2GB Atom netbook, 8GB i7 dual core
and 6GB AMD 4Core) with different file systems (ext4, btrfs) on them, so
I have quite a few configurations at hand to test with.

The RemoteString-Error during updates has been fixed; as I see the files
referenced by the download links have been replaced. The only "old"
image left I found was the one titled "Pharo 2.0 image" on
http://www.pharo-project.org/pharo-download which raises the error in
ScriptLoaderupdate>>#20605

On 07 Jun 2013, at 11:07, Markus Fritsche <[hidden email]> wrote:

> Hello Sven & List,
>
> On 07.06.2013 10:45, Sven Van Caekenberghe wrote:
>> Yes, it is an important problem (but the trigger is a change in Ubuntu 13.04). I am not an Ubuntu desktop user, I can only hope that those who are will help.
> Well, I have 13.04 on a few laptops (2GB Atom netbook, 8GB i7 dual core and 6GB AMD 4Core) with different file systems (ext4, btrfs) on them, so I have quite a few configurations at hand to test with.

Great, we need more Linux Desktop users, the community it too Mac centric ;-)

> The RemoteString-Error during updates has been fixed; as I see the files referenced by the download links have been replaced. The only "old" image left I found was the one titled "Pharo 2.0 image" on http://www.pharo-project.org/pharo-download which raises the error in ScriptLoaderupdate>>#20605

We need to fix that then.

Thanks for the feedback.

Sven

--
Sven Van Caekenberghe
Proudly supporting Pharo
http://pharo.org
http://association.pharo.org
http://consortium.pharo.org

Hello Marcus & List,

On 07.06.2013 10:44, Marcus Denker wrote:
> Normally I thought that fogbugz should send you updates on bugs you
> reported? We need to check that.
I got one mail with an URL that I could check for updates...
> In this case, we have fixed in in 3.0 and 2.0 (the bug as a "Parent Case" for 2.0,
> this was decided in the discussion and done).
> We often fix things first in 3.0 to test and then fix it in 2.0 so we are sure the fix
> works.
Thank you for explaining. I admit it was late yesterday ;-)
> So there are two problems:
>
> 1) the "RemoteString past end of file" on recent linux.
>
> 2) Something strange happens sometimes that reading the source fails.
>
>       https://pharo.fogbugz.com/f/cases/7559/Plugins-Primitives-no-longer-work-after-a-while
>       (it has a parent case for 2.0)
Can you link to Penelope about the second issue? I'd like to help hunt
these two since ubuntu-13.04 is my primary system on a few machines, big
to small.
> FugBugz does not allow read access for anonymous users. We *really* don't like this
> but it can't be changed. That is why we added bugs.pharo.org where you can get an
> account immediately *and* view all issues
Penelope only lists active bugs?
As far as I see, it doesn't have a full text search - that would
probably make things more simple.
> No, normally the idea is that people should report bug entires.
Ok, thanks for clarification. I wasn't sure after poking around a bit.

Kind regards, Markus

On Jun 7, 2013, at 11:20 AM, Markus Fritsche <[hidden email]> wrote:

>>      https://pharo.fogbugz.com/f/cases/7559/Plugins-Primitives-no-longer-work-after-a-while
>>      (it has a parent case for 2.0)
> Can you link to Penelope about the second issue? I'd like to help hunt these two since ubuntu-13.04 is my primary system on a few machines, big to small.

Sadly not because of

10157 http://bugs.pharo.org is not in sync with FogBugz
        https://pharo.fogbugz.com/f/cases/10157/http-bugs-pharo-org-is-not-in-sync-with-FogBugz


>> FugBugz does not allow read access for anonymous users. We *really* don't like this
>> but it can't be changed. That is why we added bugs.pharo.org where you can get an
>> account immediately *and* view all issues
> Penelope only lists active bugs?

Normally that is the idea, yes.

> As far as I see, it doesn't have a full text search - that would probably make things more simple.

Yes, the problem is that man power is limited and even though we really really want all these things,
the sad reality is that they are only done when they are done. I have not found any other solution than
to do, one tiny trivial step at a time...

        Marcus

On 2013-06-07, at 11:20, Markus Fritsche <[hidden email]> wrote:

you can register on the penelope page.
avoid the dot in your gmail address (BTW: gmail ignores the dots anyway. you can put as many dots as you want and still receive mails).

On Jun 7, 2013, at 11:28 AM, Camillo Bruni <[hidden email]> wrote:

We should add a bug tracker entry for that, too.

        Marcus

On 07.06.2013 11:27, Camillo Bruni wrote:
> you can register on the penelope page.
> avoid the dot in your gmail address (BTW: gmail ignores the dots anyway. you can put as many dots as you want and still receive mails).
Ah, thanks - I think I used that form before with the dot.

When should I consider myself being a monkey?

Kind regards,
   Markus

On Jun 7, 2013, at 11:47 AM, Markus Fritsche <[hidden email]> wrote:

> On 07.06.2013 11:27, Camillo Bruni wrote:
>> you can register on the penelope page.
>> avoid the dot in your gmail address (BTW: gmail ignores the dots anyway. you can put as many dots as you want and still receive mails).
> Ah, thanks - I think I used that form before with the dot.
>
> When should I consider myself being a monkey?
>
When you are a machine… the idea is that is a trivial captcha.  Maybe too complex even for humans;-)

        Marcus

Marcus Denker-4 wrote

Normally I thought that fogbugz should send you updates on bugs you reported?
We need to check that.

There's a plugin [1] but it requires the user to enable the preference. Maybe there's a way to enable globally by default...

[1] http://forum.world.st/Auto-Subscribe-in-Fogbugz-td4690614.html

On Jun 7, 2013, at 11:14 AM, Sven Van Caekenberghe <[hidden email]> wrote:

+1
I'm waiting that mac gets totally evil to switch :)

On 07 Jun 2013, at 19:45, Stéphane Ducasse <[hidden email]> wrote:

If you would switch and force a good UX, all Pharo users on Linux will thank you for it ;-)