Re (3): HOW2 make an objectified bash script ?
in [General Linux]
|
Prev: IRQ assigning for PCI card in SMP linux system ( Dual core )
Next: Re (3): HOW2 make an objectified bash script ?
From: no.top.post on 4 Jan 2010 05:56 In article <7q97rpFl8eU1(a)mid.individual.net>, "Chris F.A. Johnson" <cfajohnson(a)gmail.com> wrote: --snip -- > >> > How would I start coding this? > >> > >> My recommendation would be to put your phrases in a file like boj.data > >> > > IMO on advantage of objectifying is that the data and its operator/s > > are together. The vehicle is never separated/lost from its driver. > > I don't want extra files to be created if possible. > > So you want to bloat all the files with the same code in each? Yes, each vehicle has got its own attatched driver. Not like my real-life nightmare where I've got to fetch my single flashlite from location B or C to use at location A. Interesting how OOD used inheritance to be more economical. My computing system 'postpones the common sharing for economy' to a very low-level: all jobs use the single/same CPU to eg. clear AX. I want to economise MY EFFORT, not file space. > What happens when you need to modify the code? You have to edit it > in multiple places. This is an important consideration, which hopefully factoring and indirection would handle. So we're back to the 1950's concepts ? > If each file is written as a script, you could do it by sourcing > the code in each file: > > data=( > .... > .... > ) > source /path/to/script I don't understand this, but it may be related to a recent optimisation: I'd evolved [by adding entries] a script m: echo ' g1277 <URL-list> <accumFetces>' echo ' <next command explanation I...> .... echo ' <next command explanation N...> So then I exercise the new-bash-trick that I'd just learned to: 'm > mOptinised' And edited mOptinised to bracket the contents in: 'cat << EOF' <the file> EOF But such cute optimisations don't increase my productivety much. I waste hours looking at bash code to try to simply:-- Read(LineNumber) Write(LineNumber,ofFileX) --- And absurdly end up using awk inside bash ! Even 'sed' seems a fraud, since the tutor-writters who just copy each other, tell what: 'sed 5q' and 'sed <all but line5> does. But NOT <print line 5 only>.
From: Chris F.A. Johnson on 4 Jan 2010 07:43 On 2010-01-04, no.top.post(a)gmail.com wrote: > In article <7q97rpFl8eU1(a)mid.individual.net>, "Chris F.A. Johnson" <cfajohnson(a)gmail.com> wrote: > --snip -- >> >> > How would I start coding this? >> >> >> >> My recommendation would be to put your phrases in a file like boj.data >> >> >> > IMO on advantage of objectifying is that the data and its operator/s >> > are together. The vehicle is never separated/lost from its driver. >> > I don't want extra files to be created if possible. >> >> So you want to bloat all the files with the same code in each? > > Yes, each vehicle has got its own attatched driver. Why? A single command can be given any filename as an argument. > Not like my real-life nightmare where I've got to fetch my single > flashlite from location B or C to use at location A. If you put all your executables in a directory in your PATH, there's no problem. > Interesting how OOD used inheritance to be more economical. > My computing system 'postpones the common sharing for economy' > to a very low-level: all jobs use the single/same CPU to eg. clear AX. > > I want to economise MY EFFORT, not file space. You seem to be doing neither. >> What happens when you need to modify the code? You have to edit it >> in multiple places. > > This is an important consideration, which hopefully factoring and > indirection would handle. So we're back to the 1950's concepts ? > >> If each file is written as a script, you could do it by sourcing >> the code in each file: >> >> data=( >> .... >> .... >> ) >> source /path/to/script > > I don't understand this, The command, source, is a synonym for .; it executes a file in the current shell environment, so that it can define variables and functions to be used in the rest of the script. > but it may be related to a recent optimisation: > I'd evolved [by adding entries] a script m: > echo ' g1277 <URL-list> <accumFetces>' > echo ' <next command explanation I...> > ... > echo ' <next command explanation N...> > > So then I exercise the new-bash-trick that I'd just learned to: > 'm > mOptinised' > And edited mOptinised to bracket the contents in: 'cat << EOF' > <the file> > EOF > But such cute optimisations don't increase my productivety much. > > I waste hours looking at bash code to try to simply:-- > Read(LineNumber) > Write(LineNumber,ofFileX) > --- > And absurdly end up using awk inside bash ! There's nothing absurd about that. For many things, especially dealing with large files, awk is the best tool to use. > Even 'sed' seems a fraud, since the tutor-writters who just > copy each other, tell what: 'sed 5q' and 'sed <all but line5> > does. But NOT <print line 5 only>. sed -n '5{p;q;}' "$file" -- Chris F.A. Johnson, author | <http://cfajohnson.com> Shell Scripting Recipes: | My code in this post, if any, A Problem-Solution Approach | is released under the 2005, Apress | GNU General Public Licence
From: Bit Twister on 4 Jan 2010 13:35
On Mon, 4 Jan 2010 10:56:48 +0000 (UTC), no.top.post(a)gmail.com wrote: > In article <slrnhjupoh.71m.BitTwister(a)wb.home.test>, Bit Twister <BitTwister(a)mouse-potato.com> wrote: > >> >> Yes. So does locate boj.data > > 'whereis + locate' is conceptually-DOUBLE 'whereis', > "boj" + "boj.data" is conceptually-DOUBLE "boj". Disregarding the switch about what I said, I think I hear what you are saying, but not really the same thing. > So you've got 4 times the cognitive load !! No way. locate makes one pass into the locate database. whereis makes several calls looking in each directory in the $PATH variable until it can find the executable or runs out of directories. |