Discussion:
3.7 cannot read aliases ?
Marinos Koutsomichalis
2014-09-26 13:34:24 UTC
Permalink
Hello,

the newest 3.7 (compiled yesterday) fails to load libraries that I’ve load as aliases at Extensions..

I get this error:

Error: Target missing for alias at ...

in 3.6 and in previous 3.7 builds this is no problem

m
_______________________________________________
sc-users mailing list

info (subscription, etc.): http://www.beast.bham.ac.uk/research/sc_mailing_lists.shtml
archive: http://www.listarc.bham.ac.uk/marchives/sc-users/
search: http://www.listarc.bham.ac.uk/lists/sc-users/search/
felix
2014-09-26 13:49:21 UTC
Permalink
is it a symlink or an OS X alias ?

try symlinking




On Fri, Sep 26, 2014 at 3:34 PM, Marinos Koutsomichalis <
Post by Marinos Koutsomichalis
Hello,
the newest 3.7 (compiled yesterday) fails to load libraries that I’ve load
as aliases at Extensions..
Error: Target missing for alias at ...
in 3.6 and in previous 3.7 builds this is no problem
m
_______________________________________________
sc-users mailing list
http://www.beast.bham.ac.uk/research/sc_mailing_lists.shtml
archive: http://www.listarc.bham.ac.uk/marchives/sc-users/
search: http://www.listarc.bham.ac.uk/lists/sc-users/search/
--
..
http://soundcloud.com/crucialfelix
http://github.com/crucialfelix
.
Marinos Koutsomichalis
2014-09-26 14:28:30 UTC
Permalink
yes, it works with symlinking - thx

m
Post by felix
is it a symlink or an OS X alias ?
try symlinking
Hello,
the newest 3.7 (compiled yesterday) fails to load libraries that I’ve load as aliases at Extensions..
Error: Target missing for alias at ...
in 3.6 and in previous 3.7 builds this is no problem
m
_______________________________________________
sc-users mailing list
info (subscription, etc.): http://www.beast.bham.ac.uk/research/sc_mailing_lists.shtml
archive: http://www.listarc.bham.ac.uk/marchives/sc-users/
search: http://www.listarc.bham.ac.uk/lists/sc-users/search/
--
..
http://soundcloud.com/crucialfelix
http://github.com/crucialfelix
.
Nick Collins
2014-09-27 14:31:58 UTC
Permalink
It really should be fixed to work with aliases on OS X though? Standard users don't want to create symlinks in terminal, they'd want to just use normal aliases... aside from backwards compatibility for people's Extensions already set up?

Sorry, not got a most recent build to test right now, but anticipating upgrade hassle and teaching use of third party libraries et al. suddenly getting trickier again...

Quarks also use normal aliases on OS X don't they?

best,
N
Post by Marinos Koutsomichalis
yes, it works with symlinking - thx
m
Post by felix
is it a symlink or an OS X alias ?
try symlinking
Hello,
the newest 3.7 (compiled yesterday) fails to load libraries that I’ve load as aliases at Extensions..
Error: Target missing for alias at ...
in 3.6 and in previous 3.7 builds this is no problem
m
_______________________________________________
sc-users mailing list
info (subscription, etc.): http://www.beast.bham.ac.uk/research/sc_mailing_lists.shtml
archive: http://www.listarc.bham.ac.uk/marchives/sc-users/
search: http://www.listarc.bham.ac.uk/lists/sc-users/search/
--
..
http://soundcloud.com/crucialfelix
http://github.com/crucialfelix
.
felix
2014-09-27 15:15:19 UTC
Permalink
Symlinks are cross platform. OS X's thingies are special case and
proprietary.

I've been working on a system that just installs quarks in your working
folder and adds that folder to the compile path.

this is a more common system these days and much easier to keep track of.
you always check out the quarks you want, into the project you are working
on.
no more symlinks or broken installs buried somewhere mysterious in an
application support directory.
Post by Nick Collins
It really should be fixed to work with aliases on OS X though? Standard
users don't want to create symlinks in terminal, they'd want to just use
normal aliases... aside from backwards compatibility for people's
Extensions already set up?
Sorry, not got a most recent build to test right now, but anticipating
upgrade hassle and teaching use of third party libraries et al. suddenly
getting trickier again...
Quarks also use normal aliases on OS X don't they?
best,
N
On 26 Sep 2014, at 15:28, Marinos Koutsomichalis <
yes, it works with symlinking - thx
m
is it a symlink or an OS X alias ?
try symlinking
On Fri, Sep 26, 2014 at 3:34 PM, Marinos Koutsomichalis <
Post by Marinos Koutsomichalis
Hello,
the newest 3.7 (compiled yesterday) fails to load libraries that I’ve
load as aliases at Extensions..
Error: Target missing for alias at ...
in 3.6 and in previous 3.7 builds this is no problem
m
_______________________________________________
sc-users mailing list
http://www.beast.bham.ac.uk/research/sc_mailing_lists.shtml
archive: http://www.listarc.bham.ac.uk/marchives/sc-users/
search: http://www.listarc.bham.ac.uk/lists/sc-users/search/
--
..
http://soundcloud.com/crucialfelix
http://github.com/crucialfelix
.
--
..
http://soundcloud.com/crucialfelix
http://github.com/crucialfelix
.
Edward Nixon
2014-09-27 17:38:40 UTC
Permalink
Symlinks are cross platform. OS X's thingies are special case and proprietary.
I've been working on a system that just installs quarks in your working folder and adds that folder to the compile path.
this is a more common system these days and much easier to keep track of.
you always check out the quarks you want, into the project you are working on.
no more symlinks or broken installs buried somewhere mysterious in an application support directory.
Sorry. I don’t think this behavioural or workflow assumption is valid. And at first blush it seems like a confusion and a hassle for those whose working method is more random, exploratory or experimental. Unless OS X ‘thingies’ — could there be some condescension in that term? — are changing in their implementation, I’d vote — if this were a democracy or, as it is, the Users list rather than the Dev list — to go with Nick’s recommendation/suggestion, my plea, that the thingy capability be fixed in 5.7.

…edN
It really should be fixed to work with aliases on OS X though? Standard users don't want to create symlinks in terminal, they'd want to just use normal aliases... aside from backwards compatibility for people's Extensions already set up?
Sorry, not got a most recent build to test right now, but anticipating upgrade hassle and teaching use of third party libraries et al. suddenly getting trickier again...
Quarks also use normal aliases on OS X don't they?
best,
N
Post by Marinos Koutsomichalis
yes, it works with symlinking - thx
m
Post by felix
is it a symlink or an OS X alias ?
try symlinking
Hello,
the newest 3.7 (compiled yesterday) fails to load libraries that I’ve load as aliases at Extensions..
Error: Target missing for alias at ...
in 3.6 and in previous 3.7 builds this is no problem
m
_______________________________________________
sc-users mailing list
info (subscription, etc.): http://www.beast.bham.ac.uk/research/sc_mailing_lists.shtml
archive: http://www.listarc.bham.ac.uk/marchives/sc-users/
search: http://www.listarc.bham.ac.uk/lists/sc-users/search/
--
..
http://soundcloud.com/crucialfelix
http://github.com/crucialfelix
.
--
..
http://soundcloud.com/crucialfelix
http://github.com/crucialfelix
.
felix
2014-09-27 18:25:11 UTC
Permalink
Post by Edward Nixon
And at first blush it seems like a confusion and a hassle for those
whose working method is more random, exploratory or experimental.

Quite the opposite I would say. It makes it much easier to just try a
package out.

for instance today I installed a package (
https://github.com/dbushong/async-q ), tried something for 15 minutes and
realized it wouldn't solve my issue and removed it. Quite painless.

npm install async-q
npm remove async-q

In any case, if something works for you then you should do it that way.
I'm very much against forcing or requiring people to do anything in a
particular way and I'm not proposing any mainline changes at all.
Post by Edward Nixon
Unless OS X ‘thingies’ — could there be some condescension in that term?
— are changing in their implementation,

Sorry for the confusion, this has nothing to do with whether sclang
compilation follows aliases or not. I just meant that if you have all your
code in your working directory then you don't have any symlinks or aliases
to worry about.
Post by Edward Nixon
Post by felix
Symlinks are cross platform. OS X's thingies are special case and
proprietary.
Post by felix
I've been working on a system that just installs quarks in your working
folder and adds that folder to the compile path.
Post by felix
this is a more common system these days and much easier to keep track of.
you always check out the quarks you want, into the project you are
working on.
Post by felix
no more symlinks or broken installs buried somewhere mysterious in an
application support directory.
Sorry. I don’t think this behavioural or workflow assumption is valid. And
at first blush it seems like a confusion and a hassle for those whose
working method is more random, exploratory or experimental. Unless OS X
‘thingies’ — could there be some condescension in that term? — are changing
in their implementation, I’d vote — if this were a democracy or, as it is,
the Users list rather than the Dev list — to go with Nick’s
recommendation/suggestion, my plea, that the thingy capability be fixed in
5.7.

edN
Post by felix
It really should be fixed to work with aliases on OS X though? Standard
users don't want to create symlinks in terminal, they'd want to just use
normal aliases... aside from backwards compatibility for people's
Extensions already set up?
Post by felix
Sorry, not got a most recent build to test right now, but anticipating
upgrade hassle and teaching use of third party libraries et al. suddenly
getting trickier again...
Post by felix
Quarks also use normal aliases on OS X don't they?
best,
N
On 26 Sep 2014, at 15:28, Marinos Koutsomichalis <
Post by Marinos Koutsomichalis
yes, it works with symlinking - thx
m
Post by felix
is it a symlink or an OS X alias ?
try symlinking
On Fri, Sep 26, 2014 at 3:34 PM, Marinos Koutsomichalis <
Hello,
the newest 3.7 (compiled yesterday) fails to load libraries that I’ve
load as aliases at Extensions..
Post by felix
Post by Marinos Koutsomichalis
Post by felix
Error: Target missing for alias at ...
in 3.6 and in previous 3.7 builds this is no problem
m
_______________________________________________
sc-users mailing list
http://www.beast.bham.ac.uk/research/sc_mailing_lists.shtml
Post by felix
Post by Marinos Koutsomichalis
Post by felix
archive: http://www.listarc.bham.ac.uk/marchives/sc-users/
search: http://www.listarc.bham.ac.uk/lists/sc-users/search/
--
..
http://soundcloud.com/crucialfelix
http://github.com/crucialfelix
.
--
..
http://soundcloud.com/crucialfelix
http://github.com/crucialfelix
.
--
..
http://soundcloud.com/crucialfelix
http://github.com/crucialfelix
.
Marinos Koutsomichalis
2014-09-27 18:30:23 UTC
Permalink
Sorry for the confusion, this has nothing to do with whether sclang compilation follows aliases or not. I just meant that if you have all your code in your working directory then you don't have any symlinks or aliases to worry about.
The thing is that classes should be put in the Extensions folder which is located deeply inside the Library folder - this is not a proper place to store one’s code, I prefer having it in a folder of my choice where I can easily find it or back it up and use aliases in the Extensions folder instead

m
Symlinks are cross platform. OS X's thingies are special case and proprietary.
I've been working on a system that just installs quarks in your working folder and adds that folder to the compile path.
this is a more common system these days and much easier to keep track of.
you always check out the quarks you want, into the project you are working on.
no more symlinks or broken installs buried somewhere mysterious in an application support directory.
Sorry. I don’t think this behavioural or workflow assumption is valid. And at first blush it seems like a confusion and a hassle for those whose working method is more random, exploratory or experimental. Unless OS X ‘thingies’ — could there be some condescension in that term? — are changing in their implementation, I’d vote — if this were a democracy or, as it is, the Users list rather than the Dev list — to go with Nick’s recommendation/suggestion, my plea, that the thingy capability be fixed in 5.7.
…edN
It really should be fixed to work with aliases on OS X though? Standard users don't want to create symlinks in terminal, they'd want to just use normal aliases... aside from backwards compatibility for people's Extensions already set up?
Sorry, not got a most recent build to test right now, but anticipating upgrade hassle and teaching use of third party libraries et al. suddenly getting trickier again...
Quarks also use normal aliases on OS X don't they?
best,
N
Post by Marinos Koutsomichalis
yes, it works with symlinking - thx
m
Post by felix
is it a symlink or an OS X alias ?
try symlinking
Hello,
the newest 3.7 (compiled yesterday) fails to load libraries that I’ve load as aliases at Extensions..
Error: Target missing for alias at ...
in 3.6 and in previous 3.7 builds this is no problem
m
_______________________________________________
sc-users mailing list
info (subscription, etc.): http://www.beast.bham.ac.uk/research/sc_mailing_lists.shtml
archive: http://www.listarc.bham.ac.uk/marchives/sc-users/
search: http://www.listarc.bham.ac.uk/lists/sc-users/search/
--
..
http://soundcloud.com/crucialfelix
http://github.com/crucialfelix
.
--
..
http://soundcloud.com/crucialfelix
http://github.com/crucialfelix
.
--
..
http://soundcloud.com/crucialfelix
http://github.com/crucialfelix
.
Daniel Mayer
2014-09-27 18:46:58 UTC
Permalink
Sorry for the confusion, this has nothing to do with whether sclang compilation follows aliases or not. I just meant that if you have all your code in your working directory then you don't have any symlinks or aliases to worry about.
The thing is that classes should be put in the Extensions folder which is located deeply inside the Library folder - this is not a proper place to store one’s code, I prefer having it in a folder of my choice where I can easily find it or back it up
You have also the choice to include paths with LanguageConfig.


Greetings

Daniel

-----------------------------
www.daniel-mayer.at
-----------------------------


_______________________________________________
sc-users mailing list

info (subscription, etc.): http://www.beast.bham.ac.uk/research/sc_mailing_lists.shtml
archive: http://www.listarc.bham.ac.uk/marchives/sc-users/
search: http://www.listarc.bham.ac.uk/lists/sc-users/search/
felix
2014-09-27 19:16:56 UTC
Permalink
yes, this is exactly what I meant.

you can include or exclude folders at will using a language config.

having all the classes deep inside the library folder is not very
user-centric. its hard to find, and when things break then the user is
helpless.

countless times I've seen people post that something broke and they go and
reinstall all of supercollider, or simply give up using supercollider
entirely.

if they only thing you can mess up is your working directory then its much
more relaxing.
Post by Marinos Koutsomichalis
Post by Marinos Koutsomichalis
The thing is that classes should be put in the Extensions folder which
is located deeply inside the Library folder - this is not a proper place to
store one’s code, I prefer having it in a folder of my choice where I can
easily find it or back it up
You have also the choice to include paths with LanguageConfig.
Greetings
Daniel
felix
2014-09-27 19:20:12 UTC
Permalink
yes, this is exactly what I meant.

you can include or exclude folders at will using a language config.

having all the classes deep inside the library folder is not very
user-centric. its hard to find, and when things break then the user is
helpless.

countless times I've seen people post that something broke and they go and
reinstall all of supercollider, or simply give up using supercollider
entirely.

if they only thing you can mess up is your working directory then its much
more relaxing.
Am 27.09.2014 um 20:30 schrieb Marinos Koutsomichalis <
Post by Marinos Koutsomichalis
Post by felix
Sorry for the confusion, this has nothing to do with whether sclang
compilation follows aliases or not. I just meant that if you have all your
code in your working directory then you don't have any symlinks or aliases
to worry about.
Post by Marinos Koutsomichalis
The thing is that classes should be put in the Extensions folder which
is located deeply inside the Library folder - this is not a proper place to
store one’s code, I prefer having it in a folder of my choice where I can
easily find it or back it up
You have also the choice to include paths with LanguageConfig.
Greetings
Daniel
-----------------------------
www.daniel-mayer.at
-----------------------------
_______________________________________________
sc-users mailing list
http://www.beast.bham.ac.uk/research/sc_mailing_lists.shtml
archive: http://www.listarc.bham.ac.uk/marchives/sc-users/
search: http://www.listarc.bham.ac.uk/lists/sc-users/search/
--
..
http://soundcloud.com/crucialfelix
http://github.com/crucialfelix
.
Daniel Mayer
2014-09-27 21:39:18 UTC
Permalink
Though I must say, being on OSX I don't mind
being Library (one of them) the place of Extensions.
Added permanently to the window sidebar of finder
you're there immediately.
Nevertheless LanguageConfig is very useful.
It can solve the problem of handling
different versions of extensions and help files
with different versions of SC.

Greetings

Daniel

-----------------------------
www.daniel-mayer.at
-----------------------------


_______________________________________________
sc-users mailing list

info (subscription, etc.): http://www.beast.bham.ac.uk/research/sc_mailing_lists.shtml
archive: http://www.listarc.bham.ac.uk/marchives/sc-users/
search: http://www.listarc.bham.ac.uk/lists/sc-users/search/

Edward Nixon
2014-09-27 18:46:11 UTC
Permalink
Post by Edward Nixon
And at first blush it seems like a confusion and a hassle for those whose working method is more random, exploratory or experimental.
Quite the opposite I would say. It makes it much easier to just try a package out.
for instance today I installed a package ( https://github.com/dbushong/async-q ), tried something for 15 minutes and realized it wouldn't solve my issue and removed it. Quite painless.
npm install async-q
npm remove async-q
Not to perpetuate or increase my confusion here: a question or two:

* is this a user tool you’re talking about or something used for compiling from source?

* if it’s a user tool, does everything one needs to run ’npn’ come with the binary install of SC, in my case for OS X? Or do I have to go elsewhere to download and perhaps learn some other software? As would be the case, for example, with SVN.

If the later, I could see it as an optional functionality; I think anything that takes people away from a particular (early & narrow) orientation toward SC, in this case the idea that you’re just writing SC code and not venturing into more obscure or more sophisticated territory should be in some sort of second tier. (I’m not expressing myself well here. I suppose what I’m trying to get at is that the whole ball of wax should work ‘out of the (binary) box’ with a suitable amount of swatting up within the SC system and not depend on second or third party tools.

…edN
In any case, if something works for you then you should do it that way. I'm very much against forcing or requiring people to do anything in a particular way and I'm not proposing any mainline changes at all.
Post by Edward Nixon
Unless OS X ‘thingies’ — could there be some condescension in that term? — are changing in their implementation,
Sorry for the confusion, this has nothing to do with whether sclang compilation follows aliases or not. I just meant that if you have all your code in your working directory then you don't have any symlinks or aliases to worry about.
Post by Edward Nixon
Symlinks are cross platform. OS X's thingies are special case and proprietary.
I've been working on a system that just installs quarks in your working folder and adds that folder to the compile path.
this is a more common system these days and much easier to keep track of.
you always check out the quarks you want, into the project you are working on.
no more symlinks or broken installs buried somewhere mysterious in an application support directory.
Sorry. I don’t think this behavioural or workflow assumption is valid. And at first blush it seems like a confusion and a hassle for those whose working method is more random, exploratory or experimental. Unless OS X ‘thingies’ — could there be some condescension in that term? — are changing in their implementation, I’d vote — if this were a democracy or, as it is, the Users list rather than the Dev list — to go with Nick’s recommendation/suggestion, my plea, that the thingy capability be fixed in 5.7.
…edN
Post by Edward Nixon
It really should be fixed to work with aliases on OS X though? Standard users don't want to create symlinks in terminal, they'd want to just use normal aliases... aside from backwards compatibility for people's Extensions already set up?
Sorry, not got a most recent build to test right now, but anticipating upgrade hassle and teaching use of third party libraries et al. suddenly getting trickier again...
Quarks also use normal aliases on OS X don't they?
best,
N
Post by Marinos Koutsomichalis
yes, it works with symlinking - thx
m
Post by felix
is it a symlink or an OS X alias ?
try symlinking
Hello,
the newest 3.7 (compiled yesterday) fails to load libraries that I’ve load as aliases at Extensions..
Error: Target missing for alias at ...
in 3.6 and in previous 3.7 builds this is no problem
m
_______________________________________________
sc-users mailing list
info (subscription, etc.): http://www.beast.bham.ac.uk/research/sc_mailing_lists.shtml
archive: http://www.listarc.bham.ac.uk/marchives/sc-users/
search: http://www.listarc.bham.ac.uk/lists/sc-users/search/
--
..
http://soundcloud.com/crucialfelix
http://github.com/crucialfelix
.
--
..
http://soundcloud.com/crucialfelix
http://github.com/crucialfelix
.
--
..
http://soundcloud.com/crucialfelix
http://github.com/crucialfelix
.
felix
2014-09-27 19:19:52 UTC
Permalink
this would be completely optional and not anything to do with supercollider
itself.
Post by Edward Nixon
If the later, I could see it as an optional functionality;
you’re just writing SC code and not venturing into more obscure
well the move would be towards less obscure. supercollider is quite obscure.
--
..
http://soundcloud.com/crucialfelix
http://github.com/crucialfelix
.
Loading...