Quantcast

Re: Pymacs runs from python-mode

classic Classic list List threaded Threaded
17 messages Options
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Pymacs runs from python-mode

Andreas Röhler-2
Am 23.01.2012 05:50, schrieb François Pinard:
[ ... ]

> Hi, Andreas (hello, Barry!)
>
> I'm revisiting this under a new angle, as I now intend to try using
> Emacs for Python development for a while, to see how it goes.  I used to
> use Vim for a few years, and more recently, Wing IDE, which by the way,
> is nice.  Except that it requires some stunts to adapt it to my needs
> (while Vim or Emacs are more customizable), and the maintainers, who are
> quite responsive on the average, are reluctant to recognize a few bugs
> as such, so I sometimes feel stuck.  Free software has virtues! :-)
>
> Org mode brought me back to Emacs for many things, and one thing leading
> to another, here I am, a possibly fresh, reborn python-mode.el user.
>


Hi François,

really good news!

While completion now might be done alternatively, Pymacs based
py-complete feels faster.
Relates also to the other proceedings. Always in favor of Pymacs :)

> It seems that one problem you met is that Pymacs installs a Python
> package, having only pymacs.py besides __init__.py, while you prefer a
> single pymac.py file.  I thought that people might want to install their
> Pymacs specific files within that package.  There are alternate means,
> like specifying within Emacs an extra load-path for Python modules.  As
> I have no evidence that the Pymacs package is used as such, maybe that
> __init__.py is overkill.  I should ask Pymacs users.  Cc:ed! :-).
>
> Another problem you might have with Pymacs is that it is now
> pre-processed, there is no more a "single pymacs.py fits all".  Python 3
> and Jython introduced incompatible syntax to the point a single file may
> not be stretched to do for all cases, even through 2to3 (the needed
> stunts would become just too ugly).
>
> Earlier today, I compared the Pymacs copy within python-mode.el trunk
> with mine, and the python-mode.el lags slightly, nothing important yet.
> But it might become a problem, as we cannot have two copies of Pymacs at
> once in the same Emacs session.  Another problem is that python-mode
> freezes a particular pre-processing of pymacs.py, which is good for most
> recent Python 2.x versions, but not as universal as one would hope.
>
>>> If I understand correctly, I think that's the best way to go.
>>> python-mode.el itself shouldn't require pymacs,
>
>> Agree.  Nonetheless it was always required, as pycomplete relied on
>> it.  That changed only recently. Basic completion with M-TAB now works
>> out from python-mode.el alone.
>
> I do not understand python-mode.el enough yet to see how it uses Pymacs.
> If Pymacs launches a pymacs-helper, and if Emacs launches an "inferior
> Python shell", and M-x pdb its own as well, I presume that all these
> Python processes do not inter-operate nicely, while they should, at
> least if they all execute the same pieces of Python code.  Moreover, in
> my case, these "pieces" may sometimes be fairly big packages.  While
> Pymacs may be useful to attempt completion or inspection, and other
> things, *not* having it around reduces the size of the above puzzle.
>

What about to maintain and release a one-file Python-mode and a fule
scale integrated project? I.e having two download-files. In earlier
times that happened AFAIK.

We may take a Pymacs-branch which contains a synced python-mode.el
I'll remove the mini-Pymacs install from python-mode.el

[ ... ]

>
> It might worth aiming that python-mode's Pymacs and the Pymacs under Git
> be almost kept in sync.  If some (simple) changes on my side could make
> this easier without hurting other Pymacs users, let's ponder them.
>
> It would not be much fun, I guess, if python-mode has to repeat the
> configuration and make machinery of the Pymacs distribution, while I do
> not see how it could be avoided without creating problems here or there.
> Expecting users to explicitly, separately install Pymacs is no fun
> either.  I do not know where lies the wiser wisdom!
>
> François
>
> P.S. Feel free to copy this discussion on the python-mode mailing list
> if you think it would be a good thing to do.
>

A question around Make:

If we provide the next Python-mode/Pymacs release using Make as Pymacs
does, the top directory should contain only the minimal requiring, i.e,
INSTALL, README etc.

Cheers,

Andreas




_______________________________________________
Python-mode mailing list
[hidden email]
http://mail.python.org/mailman/listinfo/python-mode
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Pymacs runs from python-mode

François Pinard
Andreas Röhler <[hidden email]> writes:

>> Org mode brought me back to Emacs for many things, and one thing leading
>> to another, here I am, a possibly fresh, reborn python-mode.el user.

> Hi François, really good news!

Are you sure? ;-) I'll likely begin to ramble and moan about everything!
You might prefer me /silent/! :-)

>> * Pymacs installs a Python package [... rather than a single file ...]
>> * Pymacs is [...] now pre-processed [...]
>> * the Pymacs within python-mode.el trunk lags slightly [...]

> What about to maintain and release a one-file Python-mode and a full
> scale integrated project?

I'm not sure I understand your vision.  What would that be?  How could
we make things simple and dependable?

>> If Pymacs launches a pymacs-helper, and if Emacs launches an "inferior
>> Python shell", and M-x pdb its own as well, I presume that all these
>> Python processes do not inter-operate nicely [...]

That seems like a serious problem to me.  Regardless of Pymacs, I wonder
if (and hope that) pdb and an inferior Python share data and state.  It
seems to me that it might become fairly tedious otherwise.  I do not
know yet, I have to try and explore python-mode usage for real :-).

If Pymacs is merely used for completion because it is speedier (cannot
be *that* speedy anyway!), maybe that Pymacs tricks maybe repeated
without it?  That would yield a tighter integration between the
remaining Python processes, so python-mode stays more useful.

Currently, Pymacs starts its helper as a standalone Python program,
which is also quite disconnected from data and state of an inferior
Python or pdb interaction buffer.  Not much useful.

An idea, maybe!  I do not know how doable it would be: maybe Pymacs
should have some operational mode by which it starts a new thread in an
already existing program, and as such, opening a window into it.  It
surely cannot install itself on the main thread then, which is often
reserved for GUI interactions and such things.  Maybe threaded programs
want full control over their threads, and would squash Pymacs at the
first opportunity?  Maybe a Pymacs thread would mess program
termination?  Maybe this is a big can of worms, I do not know.

François

P.S. There are interesting and tempting problems in there, but I have so
little free time.  Sigh!

_______________________________________________
Python-mode mailing list
[hidden email]
http://mail.python.org/mailman/listinfo/python-mode
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Pymacs runs from python-mode

Andreas Röhler-2
Am 23.01.2012 18:40, schrieb François Pinard:

> Andreas Röhler<[hidden email]>  writes:
>
>>> Org mode brought me back to Emacs for many things, and one thing leading
>>> to another, here I am, a possibly fresh, reborn python-mode.el user.
>
>> Hi François, really good news!
>
> Are you sure? ;-) I'll likely begin to ramble and moan about everything!
> You might prefer me /silent/! :-)
>

if you could manage to moan into the bug-tracker, that will be great :)


>>> * Pymacs installs a Python package [... rather than a single file ...]
>>> * Pymacs is [...] now pre-processed [...]
>>> * the Pymacs within python-mode.el trunk lags slightly [...]
>
>> What about to maintain and release a one-file Python-mode and a full
>> scale integrated project?
>
> I'm not sure I understand your vision.  What would that be?  How could
> we make things simple and dependable?
>

Existing python-mode should not disturb installed Pymacs.
Does it? If yes, please make a report at

https://bugs.launchpad.net/python-mode

If not, next step might be a repo, keeping/installing Pymacs alongside
with our stuff.


>>> If Pymacs launches a pymacs-helper, and if Emacs launches an "inferior
>>> Python shell", and M-x pdb its own as well, I presume that all these
>>> Python processes do not inter-operate nicely [...]
>
> That seems like a serious problem to me.  Regardless of Pymacs, I wonder
> if (and hope that) pdb and an inferior Python share data and state.  It
> seems to me that it might become fairly tedious otherwise.  I do not
> know yet, I have to try and explore python-mode usage for real :-).

Fine.

>
> If Pymacs is merely used for completion because it is speedier (cannot
> be *that* speedy anyway!), maybe that Pymacs tricks maybe repeated
> without it?  That would yield a tighter integration between the
> remaining Python processes, so python-mode stays more useful.
>
> Currently, Pymacs starts its helper as a standalone Python program,
> which is also quite disconnected from data and state of an inferior
> Python or pdb interaction buffer.  Not much useful.
>
> An idea, maybe!  I do not know how doable it would be: maybe Pymacs
> should have some operational mode by which it starts a new thread in an
> already existing program, and as such, opening a window into it.  It
> surely cannot install itself on the main thread then, which is often
> reserved for GUI interactions and such things.  Maybe threaded programs
> want full control over their threads, and would squash Pymacs at the
> first opportunity?  Maybe a Pymacs thread would mess program
> termination?  Maybe this is a big can of worms, I do not know.

try M-x python-dedicated for a better sleep :)

Well, have some visions indeed. Some advices from you, resp. some bug
reports will make it advance.

Cheers,

Andreas
>
> François
>
> P.S. There are interesting and tempting problems in there, but I have so
> little free time.  Sigh!
>

_______________________________________________
Python-mode mailing list
[hidden email]
http://mail.python.org/mailman/listinfo/python-mode
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Pymacs runs from python-mode

François Pinard
Andreas Röhler <[hidden email]> writes:

> if you could manage to moan into the bug-tracker, that will be great
> :)

Don't get me started on bug trackers! :-) I much prefer humans to
robots.  But OK, I'll ponder how usable is your tracker.  It better
*behaves*! :-)

> Existing python-mode should not disturb installed Pymacs.  Does it?

I do not know.  I've no problem currently, because the Python mode
version lags only a bit, and for tiny things without consequences.

> https://bugs.launchpad.net/python-mode

I've seen much worse than this tracker, could be better.  Trackers are
also devices favoring bad maintainer attitudes.  Some of my ramblings:

https://plus.google.com/110898546836957350133/posts/NLwkdraQeay

> If not, next step might be a repo, keeping/installing Pymacs alongside
> with our stuff.

Whatever works best!

> try M-x python-dedicated for a better sleep :)

Wrong type argument: commandp, python-dedicated

> Cheers,

Keep happy!

> Andreas

François :-)
_______________________________________________
Python-mode mailing list
[hidden email]
http://mail.python.org/mailman/listinfo/python-mode
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Pymacs runs from python-mode

Andreas Röhler-2
Am 24.01.2012 04:28, schrieb François Pinard:
> Andreas Röhler<[hidden email]>  writes:
>
>> if you could manage to moan into the bug-tracker, that will be great
>> :)
>
> Don't get me started on bug trackers! :-) I much prefer humans to
> robots.

I'm your man! - LC :)

   But OK, I'll ponder how usable is your tracker.  It better

> *behaves*! :-)
>
>> Existing python-mode should not disturb installed Pymacs.  Does it?
>
> I do not know.  I've no problem currently, because the Python mode
> version lags only a bit, and for tiny things without consequences.
>
>> https://bugs.launchpad.net/python-mode
>
> I've seen much worse than this tracker, could be better.  Trackers are
> also devices favoring bad maintainer attitudes.  Some of my ramblings:
>
> https://plus.google.com/110898546836957350133/posts/NLwkdraQeay
>

"Bug trackers and other maintainer tools are not so meant to ease
communications, than to give maintainers some feel they are good at
processing bug reports. The original idea was to solve programming
problems, but it rather quickly shifted into a starve to get rid of the
report, or even the user. I observed this a lot of times."

Ahh, you got me completely.

https://bugs.launchpad.net/python-mode provides a kind of wellness.

And indeed: it helps to forget. Thus living relaxed with all the fuss
stored savely until the end of times  - :)


>> If not, next step might be a repo, keeping/installing Pymacs alongside
>> with our stuff.
>
> Whatever works best!
>
>> try M-x python-dedicated for a better sleep :)
>
> Wrong type argument: commandp, python-dedicated

bzr branch lp:python-mode

should deliver it, afterwards a

bzr pull

to update


Let's go on,

Andreas

>
>> Cheers,
>
> Keep happy!
>
>> Andreas
>
> François :-)
>

_______________________________________________
Python-mode mailing list
[hidden email]
http://mail.python.org/mailman/listinfo/python-mode
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Pymacs runs from python-mode

Andreas Röhler-2
In reply to this post by François Pinard
Am 24.01.2012 04:28, schrieb François Pinard:

>
> https://plus.google.com/110898546836957350133/posts/NLwkdraQeay
>

Hi François,

had quite similar apprehensions when started. Changed my mind.

Now use it as a kind of notebook, sometimes splitting reports into
pieces still.

Delivers a good reference when filtered. Relaxing and animating likewise.

Cheers,

Andreas

_______________________________________________
Python-mode mailing list
[hidden email]
http://mail.python.org/mailman/listinfo/python-mode
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Pymacs runs from python-mode

François Pinard
In reply to this post by Andreas Röhler-2
Andreas Röhler <[hidden email]> writes:

> And indeed: it helps to forget. Thus living relaxed with all the fuss
> stored savely [within bug trackers] until the end of times  - :)

Don't sleep too comfortably!  I often saw bug trackers losing bugs, yet
I agree that they got more solid and dependable over time, but not fully.

> Now use it as a kind of notebook, sometimes splitting reports into
> pieces still.  Delivers a good reference when filtered.  Relaxing and
> animating likewise.

Exactly.  bug trackers are tools for *maintainers*.  Whatever makes them
comfortable and happy.  They can refile, tag, sort, split, as much as
they want.  This may be their choice and sandboxes.  These tools may
*also* be convenient for users too, at least as a way, among others, to
let them know what is (not) happening.  I do understand they may be
useful for everybody.  As long as maintainers do not become fascist
about them, I've no problem. :-)

François



_______________________________________________
Python-mode mailing list
[hidden email]
http://mail.python.org/mailman/listinfo/python-mode
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Pymacs runs from python-mode

Andreas Röhler-2
Am 24.01.2012 16:11, schrieb François Pinard:

> Andreas Röhler<[hidden email]>  writes:
>
>> And indeed: it helps to forget. Thus living relaxed with all the fuss
>> stored savely [within bug trackers] until the end of times  - :)
>
> Don't sleep too comfortably!  I often saw bug trackers losing bugs, yet
> I agree that they got more solid and dependable over time, but not fully.
>
>> Now use it as a kind of notebook, sometimes splitting reports into
>> pieces still.  Delivers a good reference when filtered.  Relaxing and
>> animating likewise.
>
> Exactly.  bug trackers are tools for *maintainers*.  Whatever makes them
> comfortable and happy.  They can refile, tag, sort, split, as much as
> they want.  This may be their choice and sandboxes.  These tools may
> *also* be convenient for users too, at least as a way, among others, to
> let them know what is (not) happening.  I do understand they may be
> useful for everybody.  As long as maintainers do not become fascist
> about them, I've no problem. :-)
>
> François

started a Blueprint delivering the reasons for the kind of proceeding
tried currently

https://blueprints.launchpad.net/python-mode/+spec/pymacs

Think such a  Blueprint might be a good place for listing pros and cons.
Using Make has it's merits too, bien sure.

Andreas
_______________________________________________
Python-mode mailing list
[hidden email]
http://mail.python.org/mailman/listinfo/python-mode
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Pymacs runs from python-mode

François Pinard
In reply to this post by Andreas Röhler-2
[hidden email] (François Pinard) writes:

> It seems that one problem you met is that Pymacs installs a Python
> package, having only pymacs.py besides __init__.py, while you prefer a
> single pymacs.py file.

Hi Andreas, and gang. :-)

This detail should be over.  Pymacs now installs a single Pymacs.py file
(after pre-processing it out of Pymacs.py.in), and it does not install a
package anymore.  I kept the capital (it is Pymacs.py instead of
pymacs.py) merely so the documented API does not change at all.

While doing this, I saw that the distinction between a file and a
package matters at many little places in my code.  python-mode has been
distributing a single file, while Pymacs code was really written for a
package.  So, I wonder if Pymacs *ever* worked in python-mode...

The commit is now on GitHub (https://github.com/pinard/Pymacs).
Hopefully, no Pymacs users will complain about the disappearance of a
Pymacs package.  It is always possible that someone will reveal a blunt
I do not see right now; if this happens, more thinking will be needed.

This simplification in Pymacs should ease its usage in python-mode.

François

P.S. I am a bit sad.  After replacing the python-mode bundled in Ubuntu
by the one bleeding within launchpad, I did *not* succeed in having it
to work for me, earlier today.  And my debugging abilities in Emacs are
rather dusty.  I have to put all of this aside for a little while,
because of other duties.  As Arnold S. used to say: "I'll be back!"

_______________________________________________
Python-mode mailing list
[hidden email]
http://mail.python.org/mailman/listinfo/python-mode
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Pymacs runs from python-mode

François Pinard
In reply to this post by Andreas Röhler-2
Andreas Röhler <[hidden email]> writes:

> started a Blueprint delivering the reasons for the kind of proceeding
> tried currently

> https://blueprints.launchpad.net/python-mode/+spec/pymacs

> Think such a Blueprint might be a good place for listing pros and cons.
> Using Make has it's merits too, bien sure.

Hello, Andreas, and gang.

After a quick visit to the link above, its purpose is not clear.

Is that a device you happen to like for organizing your own thoughts on
the matter, and where you can later direct people when you feel useful
to so?  If yes : nice, good, great!  Enjoy, and have a lot of fun! :-)

Am I expected to follow what's in it?  No problem I get an email
whenever it gets edited.  Otherwise, as I really do not feel like
having-yet-another-thing-to-check-daily.  I have far too much already.

Do you expect me to share ideas in there?  Do you really expect me to
use Web editors for speaking to a "blueprint"?  I much prefer humans.
And moreover, what's wrong with Emacs, email, and this mailing list?

As a grown up, you surely not need me to jump into your chosen sandbox
for organizing thoughts in your place, are you?  *You* do that! ;-) If
you want to help me organizing my own ideas, thanks, but I can manage.

François

P.S. Who hates the useless multiplication of maintainer toys, and who
honestly needs a good incentive before taking the time it takes to study
what each is meant to do, and how (badly) it does (not do) it...


_______________________________________________
Python-mode mailing list
[hidden email]
http://mail.python.org/mailman/listinfo/python-mode
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Pymacs runs from python-mode

Andreas Röhler-2
Am 27.01.2012 03:27, schrieb François Pinard:

> Andreas Röhler<[hidden email]>  writes:
>
>> started a Blueprint delivering the reasons for the kind of proceeding
>> tried currently
>
>> https://blueprints.launchpad.net/python-mode/+spec/pymacs
>
>> Think such a Blueprint might be a good place for listing pros and cons.
>> Using Make has it's merits too, bien sure.
>
> Hello, Andreas, and gang.
>

Bon jour François,

> After a quick visit to the link above, its purpose is not clear.
>
> Is that a device you happen to like for organizing your own thoughts on
> the matter, and where you can later direct people when you feel useful
> to so?

consider it a kind of organizer for a certain title resp. task


  If yes : nice, good, great!  Enjoy, and have a lot of fun! :-)
>
> Am I expected to follow what's in it?  No problem I get an email
> whenever it gets edited.  Otherwise, as I really do not feel like
> having-yet-another-thing-to-check-daily.  I have far too much already.
>
> Do you expect me to share ideas in there?  Do you really expect me to
> use Web editors for speaking to a "blueprint"?  I much prefer humans.

it's just a possibility, some editable notice. Always listening :)

> And moreover, what's wrong with Emacs, email, and this mailing list?
>

Sending Mails onto the Blueprint works too

> As a grown up, you surely not need me to jump into your chosen sandbox
> for organizing thoughts in your place, are you?  *You* do that! ;-) If
> you want to help me organizing my own ideas, thanks, but I can manage.
>

it's rather to make public some project, discussion with results it.
a kind of public reflecting.

> François
>
> P.S. Who hates the useless multiplication of maintainer toys, and who
> honestly needs a good incentive before taking the time it takes to study
> what each is meant to do, and how (badly) it does (not do) it...

okay

So let's come to the very issue of
https://blueprints.launchpad.net/python-mode/+spec/pymacs
not mentioned yet :)

BTW did you see
https://blueprints.launchpad.net/~fpinard

Isn't that nice ;)

Cheers,

Andreas
_______________________________________________
Python-mode mailing list
[hidden email]
http://mail.python.org/mailman/listinfo/python-mode
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Pymacs runs from python-mode

Andreas Röhler-2
In reply to this post by François Pinard
Am 25.01.2012 04:28, schrieb François Pinard:

> [hidden email] (François Pinard) writes:
>
>> It seems that one problem you met is that Pymacs installs a Python
>> package, having only pymacs.py besides __init__.py, while you prefer a
>> single pymacs.py file.
>
> Hi Andreas, and gang. :-)
>
> This detail should be over.  Pymacs now installs a single Pymacs.py file
> (after pre-processing it out of Pymacs.py.in), and it does not install a
> package anymore.  I kept the capital (it is Pymacs.py instead of
> pymacs.py) merely so the documented API does not change at all.
>
> While doing this, I saw that the distinction between a file and a
> package matters at many little places in my code.  python-mode has been
> distributing a single file, while Pymacs code was really written for a
> package.  So, I wonder if Pymacs *ever* worked in python-mode...
>
> The commit is now on GitHub (https://github.com/pinard/Pymacs).
> Hopefully, no Pymacs users will complain about the disappearance of a
> Pymacs package.  It is always possible that someone will reveal a blunt
> I do not see right now; if this happens, more thinking will be needed.
>
> This simplification in Pymacs should ease its usage in python-mode.
>
> François
>
> P.S. I am a bit sad.  After replacing the python-mode bundled in Ubuntu
> by the one bleeding within launchpad, I did *not* succeed in having it
> to work for me, earlier today.  And my debugging abilities in Emacs are
> rather dusty.

all I need is the output of M-x report-emacs-bug,
maybe some example code,
and, hmm, maybe an entry into the tracker with a nice new number?

BTW if you may open an own branch at launchpad, that might be a major
help. We could play with that lp-forked Pymacs then without any risk.

OTOH could also upload a branch at your repo, if you prefer that.

At third possibility is https://github.com/pdee/pdee

Cheers,

Andreas

Cheers,

Andreas

I have to put all of this aside for a little while,
> because of other duties.  As Arnold S. used to say: "I'll be back!"
>

_______________________________________________
Python-mode mailing list
[hidden email]
http://mail.python.org/mailman/listinfo/python-mode
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Pymacs runs from python-mode

François Pinard
In reply to this post by Andreas Röhler-2
Andreas Röhler <[hidden email]> writes:

> If we provide the next Python-mode/Pymacs release using Make as Pymacs
> does, the top directory should contain only the minimal requiring,
> i.e, INSTALL, README etc.

Hi, Andreas, and gang.

Recently, I noticed that pymacs.el within python-mode (should be
Pymacs.el from now on), has a small local modification.  When Emacs
launches the Python helper, it first does:

   " sys.path.insert(1, os.getenv('PYMACS_INSTALL_DIR'));"

Let me comment a bit about this change:

* Nowhere in python-mode README nor INSTALL, this PYMACS_INSTALL_DIR
  environment variable is presented nor explained, while it really
  should.  Average Python mode users are unlikely to guess they have to
  set this environment variable prior to calling Emacs.

* The change addresses a need that python-mode has and which Pymacs
  currently does not offer, which is the need of use it without prior
  installation.  Don't shudder, I'm surely not going to ask you to go to
  GitHub and file a report about this.  If I needed it, I'd do it myself
  rather than ask others to do it in my place -- you know me! :-)

* It would be comfortable that python-mode could work with the minimum
  setup possible.  Granted that Pymacs is not the easiest thing to
  package, especially since the advent of Python 3.  Yet, if we could
  avoid the formal installation step, it would be nice.

  * I first thought about yet another Emacs variable to tell where
    exactly sits Pymacs.py, but this is not a good idea.  If Emacs is
    not able to find Pymacs.py, other Python modules written by users
    would also be unable to import from Pymacs.  It would be broken,
    just at a later stage.

  * So, if Pymacs.py is not installed or not available in the regular
    Python search path, it better be somewhere along the
    pymacs-load-path list.  So my idea would be to launch the Pymacs
    helper from any Pymacs.py found in a directory listed in
    pymacs-load-path.  If none found, then it would fall back on looking
    elsewhere in the system.  Does it sound reasonable?  Does someone
    see a flaw in this?

Have a nice evening, everybody.

François

P.S. A few more points, maybe, conveying personal opinions:

* If python-mode distributes packages copied from external sources
  (Pymacs in this example, there are a few others), it should pay
  attention at keeping its copies reasonably up-to-date.  Otherwise
  confusion, then frustration will eventually develop.

* Observed flaws or developed corrections for packages included in
  python-mode from external source should be duly reported upstream.
  The GPL applies of course, one may play by the law (I recently read
  python-mode archives, and saw such games acted by David Love, which
  were not always tasteful).  My point is that symmetrical collaboration
  is more polite, and likely more fruitful in the long run.
_______________________________________________
Python-mode mailing list
[hidden email]
http://mail.python.org/mailman/listinfo/python-mode
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Pymacs runs from python-mode

Andreas Röhler-2
Am 29.01.2012 04:24, schrieb François Pinard:

> Andreas Röhler<[hidden email]>  writes:
>
>> If we provide the next Python-mode/Pymacs release using Make as Pymacs
>> does, the top directory should contain only the minimal requiring,
>> i.e, INSTALL, README etc.
>
> Hi, Andreas, and gang.
>
> Recently, I noticed that pymacs.el within python-mode (should be
> Pymacs.el from now on), has a small local modification.  When Emacs
> launches the Python helper, it first does:
>
>     " sys.path.insert(1, os.getenv('PYMACS_INSTALL_DIR'));"
>
> Let me comment a bit about this change:
>
> * Nowhere in python-mode README nor INSTALL, this PYMACS_INSTALL_DIR
>    environment variable is presented nor explained, while it really
>    should.  Average Python mode users are unlikely to guess they have to
>    set this environment variable prior to calling Emacs.
>
> * The change addresses a need that python-mode has and which Pymacs
>    currently does not offer, which is the need of use it without prior
>    installation.  Don't shudder, I'm surely not going to ask you to go to
>    GitHub and file a report about this.  If I needed it, I'd do it myself
>    rather than ask others to do it in my place -- you know me! :-)
>
> * It would be comfortable that python-mode could work with the minimum
>    setup possible.  Granted that Pymacs is not the easiest thing to
>    package, especially since the advent of Python 3.  Yet, if we could
>    avoid the formal installation step, it would be nice.
>
>    * I first thought about yet another Emacs variable to tell where
>      exactly sits Pymacs.py, but this is not a good idea.  If Emacs is
>      not able to find Pymacs.py, other Python modules written by users
>      would also be unable to import from Pymacs.  It would be broken,
>      just at a later stage.
>
>    * So, if Pymacs.py is not installed or not available in the regular
>      Python search path, it better be somewhere along the
>      pymacs-load-path list.  So my idea would be to launch the Pymacs
>      helper from any Pymacs.py found in a directory listed in
>      pymacs-load-path.  If none found, then it would fall back on looking
>      elsewhere in the system.  Does it sound reasonable?  Does someone
>      see a flaw in this?
>
> Have a nice evening, everybody.
>
> François
>
> P.S. A few more points, maybe, conveying personal opinions:
>
> * If python-mode distributes packages copied from external sources
>    (Pymacs in this example, there are a few others), it should pay
>    attention at keeping its copies reasonably up-to-date.  Otherwise
>    confusion, then frustration will eventually develop.
>
> * Observed flaws or developed corrections for packages included in
>    python-mode from external source should be duly reported upstream.
>    The GPL applies of course, one may play by the law (I recently read
>    python-mode archives, and saw such games acted by David Love, which
>    were not always tasteful).  My point is that symmetrical collaboration
>    is more polite, and likely more fruitful in the long run.

Hi François,

think that's our understanding too. Should something in the past didn't
happen that way, assume rather a lack of force than a lack of will.

Remains the question if a README is the suitable place for storing that
kind of info. See you did it at Pymacs and certainly thats a possibility.

Indeed we are severely behind writing the info. Sorry for that.
Exists a Milestone already, 6.1.1 should deliver it - maybe already 6.1?

May you open your own branch forked from trunk at lp?
Could merge than all Pymacs diffs easily.

Cheers,

Andreas





_______________________________________________
Python-mode mailing list
[hidden email]
http://mail.python.org/mailman/listinfo/python-mode
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Pymacs runs from python-mode

François Pinard
Andreas Röhler <[hidden email]> writes:
> Am 29.01.2012 04:24, schrieb François Pinard:

>> * README nor INSTALL should tell about PYMACS_INSTALL_DIR
>> * The change addresses a need which Pymacs does not offer
>> * python-mode should work with the minimum setup possible
>> * python-mode should pay attention at keeping its copies up-to-date
>> * Observed flaws should be duly reported upstream

> think that's our understanding too. Should something in the past
> didn't happen that way, assume rather a lack of force than a lack of
> will.

Yes, this is what I assume.  (Same for user documentation.)  Yet we have
to speak out about such things, because laziness being what it is (I do
know what laziness is!), those accidental misses progressively transform
into habits, habits into rights, and rights into principles.  We have to
stay careful all along, and discuss methodology problems early -- as
long, of course, as discussions do not suck up the left bits of energy!

> Remains the question if a README is the suitable place for storing
> that kind of info.

In the current state of affairs with python-mode, yes, clearly,
undoubtedly.

> See you did it at Pymacs and certainly thats a possibility.

Pymacs documentation is not in README, it is a reStructured Text
formatted manual (the pre-processor has its own separate manual).  But
when Pymacs started and had nothing else, yes, it was all in README.
README is the traditional unavoidable file in a distribution, where
users seek for installation instructions and user documentation (or at
least, pointers to where it sits, when not included).

> Indeed we are severely behind writing the info. Sorry for that.
> Exists a Milestone already, 6.1.1 should deliver it - maybe already
> 6.1?

Making Milestones, Blueprints, Gantt charts about it is not effective.
Fill in the README details *right now*, when you happen to stumble on
them.  When you see yourself thinking "Hey, I'll add this information
later into the Great Documentation I'm planning to have", slap yourself!
:-) Rather *immediately* dump the information, all raw, all unformatted,
into README.  Do not even try to be proud of it, that would come later.
For now, just pour information in it at any opportunity.

> May you open your own branch forked from trunk at lp?  Could merge
> than all Pymacs diffs easily.

Hey, let's move python-mode over to GitHub instead :-) And besides, if
people really think it is hopeless to ever get python-mode integrated
into Emacs, is there much remaining reason left, sticking with Bazaar?

Keep happy!

François

P.S. Behind the above tease, I hope you see my point: maintainers should
get away from this attitude of always attempting to pull all users (or
other maintainers) in their own playground.  This is one of the aspects
I hate about trackers and other maintainer toys.  While it's comfortable
for me to get a mere "pull request" on GitHub, I'd never try to force my
users into it, as it might not be comfortable for them.  My comfort is
not *more worth* than theirs.
_______________________________________________
Python-mode mailing list
[hidden email]
http://mail.python.org/mailman/listinfo/python-mode
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Pymacs runs from python-mode

Richard Everson
In reply to this post by Andreas Röhler-2


On 29 Jan 2012, at 08:03, Andreas Röhler wrote:

> Am 29.01.2012 04:24, schrieb François Pinard:
>> Andreas Röhler<[hidden email]>  writes:
>>
>>> If we provide the next Python-mode/Pymacs release using Make as Pymacs
>>> does, the top directory should contain only the minimal requiring,
>>> i.e, INSTALL, README etc.
>>
>> Hi, Andreas, and gang.
>>
>> Recently, I noticed that pymacs.el within python-mode (should be
>> Pymacs.el from now on), has a small local modification.  When Emacs
>> launches the Python helper, it first does:
>>
>>    " sys.path.insert(1, os.getenv('PYMACS_INSTALL_DIR'));"
>>
>> Let me comment a bit about this change:
>>
>> * Nowhere in python-mode README nor INSTALL, this PYMACS_INSTALL_DIR
>>   environment variable is presented nor explained, while it really
>>   should.  Average Python mode users are unlikely to guess they have to
>>   set this environment variable prior to calling Emacs.

Hi,

This was my ugly hack to enable Python to find Pymacs when Pymacs is distributed with python-mode.el and is therefore not on sys.path.  There's nothing about it in README or INSTALL because I think it's purely local to the python-mode/pymacs interaction and the general user shouldn't need to know or do anything about it: that environment variable is set by python-mode.el.   On the other hand, it would be *much* better to use the distributed, current version of Pymacs, but it's not clear to me how to achieve that if Pymacs is not installed somewhere on the standard sys.path.

Best,
Richard.


_______________________________________________
Python-mode mailing list
[hidden email]
http://mail.python.org/mailman/listinfo/python-mode
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Pymacs runs from python-mode

Andreas Röhler-2
In reply to this post by François Pinard
[ ... ]
>> May you open your own branch forked from trunk at lp?  Could merge
>> than all Pymacs diffs easily.
>
> Hey, let's move python-mode over to GitHub instead :-)

it's mirrored here:

https://github.com/pdee/pdee/tree/master/python-modes/python-mode-el

Should you make a branch, Pymacs stuff shaped as you want, I can merge
from there.


Andreas

[ ... ]
_______________________________________________
Python-mode mailing list
[hidden email]
http://mail.python.org/mailman/listinfo/python-mode
Loading...