Re: set -e vs. #!/bin/sh -e

Manoj Srivastava wrote: > I see no benefit in making it easy to turn error detection > off. It should be done as a matter of last resort. In that case, they > can edit (a copy of) the script. > > So, I guess I object to anything like this ever getting into > the Policy standard.

Re: set -e vs. #!/bin/sh -e

Hi, The following is quite subjective. I have grown to distrust #!/bin/sh -e; cause 1) Some systems there was a length contraint on the command name that could be put on the line (/mnt/usr/group/share/mips3000/2.12/bin/sh would fail, for example). This is probably not relevant f

Re: set -e vs. #!/bin/sh -e

Oliver Elphick wrote: > > There are cases where one wants to sandwich a section of code with `set -e' > `set +e'. Policy should not be so rigid as to exclude the possibility > of doing this. Hmm, let's forget about what I've said for a sec. I don't think that any script should do the above in a

Re: set -e vs. #!/bin/sh -e

Herbert Xu wrote: >Hi: > Up till now there has been no policy on whether one should use >#!/bin/sh -e or do set -e in a script. I think we should gradually deprecat >e >the latter in favour of #!/bin/sh -e because with it one can override the -e >option by doing something like

set -e vs. #!/bin/sh -e

Hi: Up till now there has been no policy on whether one should use #!/bin/sh -e or do set -e in a script. I think we should gradually deprecate the latter in favour of #!/bin/sh -e because with it one can override the -e option by doing something like "/bin/sh script". This would help in