Bug #1502
SystemCommand in Microsoft version
Description
Andraschko reports that SystemCommand
does not work properly in the Microsoft release.
Related issues
History
#1 Updated by John Abbott over 3 years ago
- Related to Feature #1231: system command added
#2 Updated by John Abbott over 3 years ago
Our code looks to be correct (in agreement with the description given on website cpprefeence.com
).
At the moment I have no idea to proceed. A quick internet search did not reveal any other people reporting similar problems.
Emacs on Microsoft appeared to be able to send CLI commands (via M-!
).
Puzzled.
#3 Updated by John Abbott over 3 years ago
With luck Andraschko will make some tests this week.
By chance I happened to speak to a friend (Nick) who suggested that it might be a good idea to ensure that output from any command is sent to a file (since Microsoft cmd.exe
may not have a notion like "standard output").
#4 Updated by John Abbott over 2 years ago
- Status changed from New to In Progress
- Assignee set to John Abbott
- % Done changed from 0 to 50
Bernhard and I made several tests yesterday: well, he made the tests, I just came up with "clever"(??) ideas.
SUMMARY- so that
SystemCommand
works the code must be compiled in Cygwin with the flag-D_GNU_SOURCE
, and the executable must reside in a subdirectory ofC:/cygwin
- if either condition is not satsfied then
SystemCommand
appears to do nothing, and returns a non-zero code; strangely the test inside CoCoA-5 does not detect that the low-levelsystem
command fails... I wonder why?
Anyway, The build scripts now set the compilation flag _GNU_SOURCE
if the platform is Cygwin.
We must document the requirement to place the executable in a subtree of C:/cygwin
. Where should this advice be?
PS Many thanks to Bernhard for his patience and assiduity.
#5 Updated by John Abbott over 2 years ago
- Related to Feature #1373: GFan also for Microsoft? added
#6 Updated by John Abbott over 2 years ago
- Target version changed from CoCoA-5.4.0 to CoCoA-5.4.2