2017-05-15 18:10 GMT+02:00 Alistair Grant :
> And if I execute the following in a playground:
>
> | ls |
> OSSUnixSubprocess new
> shellCommand: 'ls ~';
> redirectStdout;
> runAndWaitOnExitDo: [ :process :outString |
> ls := outString
> ].
> ls.
>
>
On Mon, May 15, 2017 at 04:10:17PM +, Alistair Grant wrote:
> On Mon, May 15, 2017 at 08:14:44AM -0700, Eliot Miranda wrote:
> >
> >
> > On Mon, May 15, 2017 at 8:09 AM, Damien Pollet
> > wrote:
> >
> > My point was, if #shellCommand: accepts a single string for a whole
> > command
On Mon, May 15, 2017 at 08:14:44AM -0700, Eliot Miranda wrote:
>
>
> On Mon, May 15, 2017 at 8:09 AM, Damien Pollet
> wrote:
>
> My point was, if #shellCommand: accepts a single string for a whole
> command, then surely it passes it to a system shell already, so why nest a
> bash i
> On 15 May 2017, at 18:03, Denis Kudriashov wrote:
>
>
> 2017-05-15 16:55 GMT+02:00 K K Subbu :
> What exactly are you trying to do? If you want the list of entries in ~
> directory, you could use FileDirectory methods:
>
> (FileDirectory on: (OSProcess thisOSProcess environmentAt: 'HOME'))
2017-05-15 16:55 GMT+02:00 K K Subbu :
> What exactly are you trying to do? If you want the list of entries in ~
> directory, you could use FileDirectory methods:
>
> (FileDirectory on: (OSProcess thisOSProcess environmentAt: 'HOME'))
> entries collect: [ :f | f name ].
>
My question was only abo
Thanks!
We should produce a new version of the launcher and push it more but too
busy.
stef
On Mon, May 15, 2017 at 9:15 AM, Torsten Bergmann wrote:
> Pharo Launcher was updated for Pharo 7 already:
>
> ---
> Name: PharoLauncher-Core-Torst
On Mon, May 15, 2017 at 8:09 AM, Damien Pollet
wrote:
> My point was, if #shellCommand: accepts a single string for a whole
> command, then surely it passes it to a system shell already, so why nest a
> bash in between? (effectively running the equivalent of sh -c "bash -c \"ls
> ~\"')
>
Clearly
My point was, if #shellCommand: accepts a single string for a whole
command, then surely it passes it to a system shell already, so why nest a
bash in between? (effectively running the equivalent of sh -c "bash -c \"ls
~\"')
On 15 May 2017 at 16:28, Eliot Miranda wrote:
> Hi Damien,
>
> On May 1
On Monday 15 May 2017 05:46 PM, Denis Kudriashov wrote:
OSSUnixSubprocess new
shellCommand: 'ls ~';
redirectStdout;
runAndWaitOnExitDo: [ :process :outString |
^outString
].
What exactly are you trying to do? If you want the list of entries in ~
directory, you could use FileDirectory methods:
Hi Damien,
> On May 15, 2017, at 6:44 AM, Damien Pollet wrote:
>
>> On 15 May 2017 at 15:26, Eliot Miranda wrote:
>> Try something like
>>
>> shellCommand: 'bash -c ''ls ~''';
>
> But then that would run ls inside of bash inside of the system shell
> (/bin/sh), wouldn't it? What's th
On 15 May 2017 at 15:26, Eliot Miranda wrote:
> Try something like
>
> shellCommand: 'bash -c ''ls ~''';
>
But then that would run ls inside of bash inside of the system shell
(/bin/sh), wouldn't it? What's the point?
--
Damien Pollet
type less, do more [ | ] http://people.untyped.org/
> On May 15, 2017, at 5:16 AM, Denis Kudriashov wrote:
>
>
> 2017-05-15 14:04 GMT+02:00 Mariano Martinez Peck :
>> Denis, you can also have shortcut methods to call shell from OSSubprocess.
>> See the tests under "tests - shell" protocol.
>
> I tried this:
>
> OSSUnixSubprocess new
>
2017-05-15 14:04 GMT+02:00 Mariano Martinez Peck :
> Denis, you can also have shortcut methods to call shell from OSSubprocess.
> See the tests under "tests - shell" protocol.
I tried this:
OSSUnixSubprocess new
shellCommand: 'ls ~';
redirectStdout;
runAndWaitOnExitDo: [ :process :outString |
Denis, you can also have shortcut methods to call shell from OSSubprocess.
See the tests under "tests - shell" protocol.
Cheers,
On Mon, May 15, 2017 at 8:00 AM, Denis Kudriashov
wrote:
>
> 2017-05-15 12:13 GMT+02:00 K K Subbu :
>
>> ~ is a shell language synonym for $HOME. Both are expanded by
2017-05-15 12:13 GMT+02:00 K K Subbu :
> ~ is a shell language synonym for $HOME. Both are expanded by the shell
> before calling file syscall in the kernel. Kernel's filename resolver does
> not do ~ or $HOME expansions.
>
Thanks Subbu :)
2017-05-15 12:10 GMT+02:00 teso...@gmail.com :
> Hi Denis,
>the ~ (tilde) is expanded by the shell, a running process should expand
> it. In the case of the command line tools they don't expand it, as they
> expect the shell to expand it before calling.
>
> Check https://www.gnu.org/software/b
2017-05-15 12:05 GMT+02:00 K K Subbu :
> Hi,
>
> I am tempted ;-) by the following message on Pharo's download page:
> ---
> The minimal image is a reduced version of Pharo 7.0 (alpha): You have the
> most basic environment and you can interact with it via command line.
> WARNING: For hackers only
On Monday 15 May 2017 03:31 PM, Denis Kudriashov wrote:
It returns empty string and console report:
ls: cannot access ~: No such file or directory
Any idea why it not works properly? (from command line "ls ~" returns
expected files).
~ is a shell language synonym for $HOME. Both are expa
Hi Denis,
the ~ (tilde) is expanded by the shell, a running process should expand
it. In the case of the command line tools they don't expand it, as they
expect the shell to expand it before calling.
Check
https://www.gnu.org/software/bash/manual/html_node/Tilde-Expansion.html
and http://www.da
Hi,
I am tempted ;-) by the following message on Pharo's download page:
---
The minimal image is a reduced version of Pharo 7.0 (alpha): You have
the most basic environment and you can interact with it via command
line. WARNING: For hackers only!
---
What kind of hacking has been done with th
Little problem which I not understand. Maybe you can explain. Path '~' is
not recognizied in ls command:
OSSUnixSubprocess new
command: '/bin/ls';
arguments: #('~');
redirectStdout;
runAndWaitOnExitDo: [ :process :outString |
outString
]
I
I just checked. It works perfectly.
One notice for API:
OSSUnixSubprocess new
command: '/bin/ls';
arguments: #('-la' '/Users');
redirectStdout;
runAndWaitOnExitDo: [ :process :outString |
outString
]
Why you are not return result of block?
Hi,
Currently extending the UI of the debugger is not as easy as extending the
inspector. (It cannot just be done using annotations)
We should work on improving this for Pharo 7.
The actual presentations at the bottom are created
in GTDebuggerVariablesBrowser>>#compose.
To change it in a debugger
Hi
2017-05-15 8:22 GMT+02:00 :
> I've looked at it, but I can't get to reproduce it. I have trouble finding
> an entry point to start understanding.
Starting point is method GTMoldableDebugger>>compose with code:
...
self tabulator with: [:browser |
self debuggerStructureIn: browser.
self debu
Yeah, in any case, it is not clear what we should do:
- can't we extend the default debugger presentation?
- or should we create a new presentation?
- maybe we can do both, but then we have the question: is there a criteria
to say "if X, extend the default presentation, if Y, create a new one"?
Pharo Launcher was updated for Pharo 7 already:
---
Name: PharoLauncher-Core-TorstenBergmann.105
Author: TorstenBergmann
Time: 15 May 2017, 9:12:21.430226 am
UUID: 9b8faeb3-da25-49e9-bdf4-0c2b130ed357
Ancestors: PharoLauncher-Core-CyrilFerli
26 matches
Mail list logo
