Re: [GNC-dev] GnuCash 3 on Linux

2019-02-27 Thread David Cousens
Wm

Diagrams are up https://wiki.gnucash.org/wiki/Configuration_Locations. The
wkiki markup is crude but works. I will try later to implement a scroll box
to better accommodate narrow width monitors and mobile devices. Geert
pointed out a few corrections re what are merely labels for locations and
what are actual definable environmental variables which are now
incorporated.

The original text on the wiki did not distinguish too clearly between what
were labels for locations and what can be actually set as locations. It was
there but obscure. In the diagrams these are distinguished as clearly as I
am able to.

On Linux at least those environment variables (not the ones which are merely
labels) are not created on installation. They can be and then GnuCash should
use the locations defined by them (I haven't verified this mainly because I
don't have the need to use it). In principle at least you could define
GNC_DATA_HOME and GNC_CONFIG_HOME to point to the book location or a
subdirectories of the directory containing the book. 

>The problem with the gnc code for transition is that it placed no value 
>on who accessed it first.

Why would it? GC is not designed as a multiuser program for a multiuser
environment. It has no user structure to define authorization levels and no
code to restrict access to specific functionality. While there is perhaps an
ambition to move in this direction longer term, it is nowhere near there
yet.  

Admittedly the release notes for V3 only obliquely mention the relocation of
the user configuration location and that in hindsight should perhaps have
been highlighted more.  But it came up pretty quickly in the forum
http://gnucash.1415818.n4.nabble.com/GNC-Saved-Report-Configurations-Missing-After-Upgrade-to-v3-td4700786.html#a4700793
along with the cure of renaming saved-reports-2.4 to saved-reports-2.8 in
the new directory. 

>The place was identifiable relative to the book. 3.x changed that.

To a place which is also identifiable relative to the book. 

The changes going to V3 also make it possible (at least in principle) to
co-locate the configuration with the book by defining  GNC_DATA_HOME and
GNC_CONFIG_HOME which was not possible under V2.6. 

Again you missed the point that the Save and Save As options in the reports
toolbar and menu don't or at least shouldn't save the report per se, but 
save only the configuration information for a report. What level of personal
information gets saved in the configuration is not clear to me.

I can conceive of a situation where a number of books use a stock standard
account heirarchy and one configures a report to depend on that heirarchy
but if account guids are incorporated this becomes of limited usefulness wrt
the ability to transfer reports.

The report configuration should not contain the information directly and
only have pointers to locate the information which is contained only in the
book. Is that not the case?


David Cousens



-
David Cousens
--
Sent from: http://gnucash.1415818.n4.nabble.com/GnuCash-Dev-f1435356.html
___
gnucash-devel mailing list
gnucash-devel@gnucash.org
https://lists.gnucash.org/mailman/listinfo/gnucash-devel


Re: [GNC-dev] Building 3.4 on Mint 18.3

2019-02-27 Thread Jacob Larsen
Thanks, I missed the specific installation page and just tried to 
install from the generic dependency page, trying to guess the actual 
package names. I had missed a few, but it looks like that was not the 
issue though.


Turns out that I had to switch from the nouveau graphics driver to the 
proprietary driver, and now it works. So while the updated dependencies 
might be required for other things, they were not actually my issue, but 
it seems a broken graphics driver. Which is not broken with the stock 
gnucash 2.6 that comes with Mint 19.1 Mate.


Something I probably should have seen sooner given that a complete 
system freeze normally points to something in kernel space


/Jacob

On 27/02/2019 00.38, David Cousens wrote:

Jacob,

The reports use Scheme which is implemented in the Guile libraries so they
are a likely candidate.

The dependencies for V3 are listed
https://wiki.gnucash.org/wiki/Dependencies which is a breakout from the
general instructions for building at  https://wiki.gnucash.org/wiki/Building
and the Linux instructions at
https://wiki.gnucash.org/wiki/Building#Building_on_Linux_Distributions.
Installation of the dependencies is described
https://wiki.gnucash.org/wiki/Installing_Dependencies. These instructions
were originally formulated on Linux Mint 18.3/Ubuntu 16.04 with v 2.6 but
have been updated for V3 and also for Linux Mint 19.1/Ubuntu 18.04 but have
also been used for Ubuntu 18.10 as well.

Check the Guile library version you have installed. The Cmake command should
be checking this and stopping if the required library versions are not
present.

David Cousens



-
David Cousens
--
Sent from: http://gnucash.1415818.n4.nabble.com/GnuCash-Dev-f1435356.html
___
gnucash-devel mailing list
gnucash-devel@gnucash.org
https://lists.gnucash.org/mailman/listinfo/gnucash-devel

___
gnucash-devel mailing list
gnucash-devel@gnucash.org
https://lists.gnucash.org/mailman/listinfo/gnucash-devel


Re: [GNC-dev] GnuCash 3 on Linux

2019-02-27 Thread Geert Janssens
Op woensdag 27 februari 2019 20:36:57 CET schreef Wm via gnucash-devel:
> On 24/02/2019 17:18, Geert Janssens wrote:
> > Op zondag 24 februari 2019 17:19:09 CET schreef Wm via gnucash-devel:
> > 
> > Looks like you are now lying in public... (using your own conversation
> > style here).
> > 
> > Nothing gets deleted by the migration so there can't be data loss.
> 
> We are talking well on the bug list, Geert.  What you say is untrue,
> don't spoil the illusion until you know the details and truth of what
> can be lost.
> 
You know what Wm ?

I care about a constructive conversation, but this isn't one in my book and it 
saddens me we didn't manage to make it one. I've run tired of this dancing 
around and so I'll end it right here.

I have repeatedly asked about specific and objective details about your issue 
in an effort to help you solve it. All I get is "you're wrong, I'm right, 
you'll see" and vague allusions to design flaws. I don't care about right or 
wrong, but about solving problems.

You on the other hand seem to enjoy grand talk more than having your problem 
solved. So I have nothing more to contribute here. I feel I have wasted enough 
time on this.

In short: you don't pay me enough to be entitled to insult or belittle me and 
at the same time expect me to continue to help you.

Enjoy your life and goodbye,

Geert

PS I will ignore your mails future mails until I find you have learned how to 
respectfully and constructively communicate with other people.


___
gnucash-devel mailing list
gnucash-devel@gnucash.org
https://lists.gnucash.org/mailman/listinfo/gnucash-devel


Re: [GNC-dev] GnuCash 3 on Linux

2019-02-27 Thread Wm via gnucash-devel

On 24/02/2019 01:50, Chris Good wrote:


Chris,

If I save reports from GnuCash 3.4 on Linux Mint Tara (Ubuntu 18.04), the
reports are saved in
/home//.local/share/gnucash/saved-reports-2.8 and I have verified this
contains the report config which I saved.


yay


This is where I expected to find them based on the description in
https://wiki.gnucash.org/wiki/Configuration_Locations.


yay


I am not sure what the 2.4 in your case and 2.8 in my case actually refer
to. The wiki says  they are related to GnuCash versions but not obviously
the version which is running.


2.8 and 2.4 make no difference, wastrel :)


Hi David,

Geert answered this 6/6/2018:

2.8, not 2.4. If not 2.8 file is found, gnucash 3 and up will search for a
saved-reports-2.4, but it will only save to saved-reports-2.8.


this is a spit mouth reply as it doesn't say *where* the report file 
will be saved.


I expected more of you, ChrisG


I haven't updated the wiki as it was recently changed to say:
saved-reports-x.y
presumably to ensure it doesn't quickly become outdated.

Personally, I would prefer if it was more specific.


I'm hoping I'll prevail or you or someone with some sense.


I am finally about to release my BackupGnuCash java app 1.3.0, now updated
for GnuCash 3.


Scary stuff, brave you.


As it is difficult to tell if a configuration file location has been
overridden by an environment variable somewhere, I've taken the tack of only
looking for configuration files to backup in the default locations.



Regards,
Chris Good


May I have a look at your script?

Love, Wm



___
gnucash-devel mailing list
gnucash-devel@gnucash.org
https://lists.gnucash.org/mailman/listinfo/gnucash-devel


Re: [GNC-dev] GnuCash 3 on Linux

2019-02-27 Thread Wm via gnucash-devel

On 25/02/2019 06:52, David Cousens wrote:


I admit freely I do not have a clear understanding of how the reports (and
much else inside GnuCash) do work in detail and I doubt if I am alone in
that apart from maybe the core development team.


You are expected to be the foil for those that do know, sir :)


GnuCash for whatever historical reasons is only sparsely documented which
increases the difficulty of those of us with some development experience,
but not long term on GnuCash, to contribute effectively. There is a big
learning curve anytime you dig into the code.


That is an understatement, the code is largely opaque comparative to the 
relatively simple transactions that occur.



Having worked in development
teams which had very limited resources in the past


[snip] I hear you, I am not that boss here


GnuCash doesn't have the luxury of a systems analysis team and management
structure with oversight either. There is just a minimal team of unpaid
developers, often wearing multiple hats, doing their best. Despite that it
is far more flexible and adaptable than many programs which do have
extensive development teams - hence the escapees from MYOB and Quicken where
cost becomes a very significant factor.


Hint: I am part of a systems analysis team and manage software.  I just 
don't get used much.


Since I know that I often find myself taking another view.


Does gnc know where it took stuff from and placed it when 3.x was run
for the first time?


Looking at the code which did this, should answer this question. Not easily
though, as the code generates scripts which perform the gconf->gsettings
,dconf migration (forced by the deprecation of g conf) as well as the
changes in the settings location.


some stuff got lost there, continue


My brief excursion into it indicates that
it is as described on the wiki Configuration Locations page. What was
migrated is defined in /usr/local/share/gnucash/migratable-prefs.xml and
processed by /usr/local/share/gnucash/make-prefs-migration-script.xsl to
produce the gsettings from the original gconf settings.


Wow!  I admire your faith!

problem is:

the variables are just that, variable

I don't think anyone knows where some of the details went


My experience with the transition from 2.6.19 in my case to v3.0 was that I
lost all of the training for the import matcher - no great problem, a couple
of imports and it was largely back and that solved some historical import
problems  which were throwing up accounts which were no longer appropriate.
There is some value in occasionally retraining the import matcher in any
case.


Your report is occasional not substantive. You will hopefully get a 
better report.  The importer farts when I throw new stuff at it.  Expected.



I get by with the basic reports when i need them so i personally had
no problems with reports, but I do appreciate that others may/did/do. I kept
the 2.6 .gnucash preferences folder from 2.6.19 for some time after i
migrated to 3.0in case I went back but unfortunately I decided around 3.3.
that for my purposes things had become stable enough that i no longer needed
to keep it.


Uh oh


If I remember correctly from looking at it at the time I had problems with
the importer, what was in

/home//.gnucash

  was moved to

  /home//.local/share/gnucash

on Linux systems as described in the Wiki page on configuration locations. I
am presuming this is also he case for Windows systems  but I don't really
use Windows so can't be sure.


The problem with the gnc code for transition is that it placed no value 
on who accessed it first.


This is slightly specious but if the cleaning lady accessed the file 
first she'd have all the reports in her private space on her phone 
forever!  It is that weird.



It is a lot more complicated than you think.
  
Wm,

2.6 didn't store the saved reports with the book file either, it just stored
them in a different location, so that is clearly not the problem.


The place was identifiable relative to the book. 3.x changed that.


What are the other changes that were introduced going from 2.6 to 3.x which
are causing you problems now in backing up your files that didn't exist with
2.6? You are going to have to be specific if anyone is to diagnose and fix
it.


Other technical people already know.  I'll bore you by saying some 
reports should be part of or very close to a "book". For the simple 
reason that (in the case of my charities for example) only the trustees 
are allowed to report on the accounts.



I also agree completely that if a report configuration is tied to a specific
book/set of accounts, i.e. it has been customised or otherwise been modified
to be specific to that book, then its storage location should be with the
data file. What defines this level of customisation for you and for me and
any other user? Could GnuCash  this on the basis of the configuration file
contents perhaps decide where a particular report configuartion file should
be located?


Yes, 

Re: [GNC-dev] GnuCash 3 on Linux

2019-02-27 Thread Wm via gnucash-devel

On 24/02/2019 17:18, Geert Janssens wrote:

Op zondag 24 februari 2019 17:19:09 CET schreef Wm via gnucash-devel:



Looks like you are now lying in public... (using your own conversation style
here).

Nothing gets deleted by the migration so there can't be data loss.


We are talking well on the bug list, Geert.  What you say is untrue, 
don't spoil the illusion until you know the details and truth of what 
can be lost.



If the
migration was actually triggered all files from DOT_GNUCASH_DIR have been
*copied* to either GNC_DATA_HOME or GNC_CONFIG_HOME. You can look up the exact
locations per platform on https://wiki.gnucash.org/wiki/
Configuration_Locations. The original files are still in DOT_GNUCASH_DIR.


The problem is all of those VARIABLES were different for each person so 
we don't know where the bits and pieces of gnc have gone for everyone.


I think this is an irrevocable error.

If anyone is interested Geert and I are talking about this on the bug 
list in detail.


This will, I think, turn out to be a mistake.
--
Wm



___
gnucash-devel mailing list
gnucash-devel@gnucash.org
https://lists.gnucash.org/mailman/listinfo/gnucash-devel


Re: [GNC-dev] gnucash maint: Multiple changes pushed

2019-02-27 Thread Wm via gnucash-devel

On 26/02/2019 11:56, Christopher Lam wrote:

Hi Geert,
Sharp eye indeed.
I haven't used the new gnc:gui-error function because report.scm is being
simulateously attacked:
(1) refactoring in maint-scheme-progress
(2) slowly creating nearly 100% coverage for tests
(3) fixing an invalid code path introduced about 8 months ago, which was
returning #f to signify failure for a report-definition-without-guid, when
in reality the report-definition-without-guid was correctly handled by
autogenerating guid.
I considered gnc:gui-error to be a proper refactoring job belonging to (1)
in my unpushed maint-scheme-progress branch but needed to create tests
during (2) first to ensure the refactoring was safe, and the tests needed
to handle non-gui code properly.
I felt that completing (2) was more important than (1).
But you're right, we could easily have used gnc:gui-error here.


I love your confidence, Cristopher Lam.

--
Wm


___
gnucash-devel mailing list
gnucash-devel@gnucash.org
https://lists.gnucash.org/mailman/listinfo/gnucash-devel