Episode: 2060
Title: HPR2060: Introduction to sed - part 5
Source: https://hub.hackerpublicradio.org/ccdn.php?filename=/eps/hpr2060/hpr2060.mp3
Transcribed: 2025-10-18 13:51:08

---

This is HPR episode 2016 titled Introduction to Send Part 5 and in part on the series Learning Send.
It is hosted by Dave Morris and in about 48 minutes long.
The summer is finishing covering send commands, looking at some enample crypts.
This episode of HPR is brought to you by an honesthost.com.
Get 15% discount on all shared hosting with the offer code HPR15, that's HPR15.
Better web hosting that's honest and fair at an honesthost.com.
Hello everyone, this is Dave Morris.
Now today I'm finishing the series on said.
It's taken 5 episodes to cover it in some depth.
I get the impression that quite a few people have tuned out by now.
But there's some quite interesting features in here and I thought there would be at least a small number of listeners who would be interested in delving into them.
So in the last episode we looked at pretty much everything about the way that said works.
And we looked at the hold and the pattern buffers.
We looked at some of the commands that we hadn't seen at that particular point and started to see what could be done with them.
Now there are a few more remaining commands, we're not going to cover them all.
But some of them are, well I think they're all pretty obscure, possibly even very obscure as I put in the notes here.
But I wanted to cover them because I'd like to go over a few of the example said scripts that are in the said, the GNU said manual.
The reason for doing this is that I went looking at this manual when I was trying to work out how to use said in a slightly more advanced way.
And looking at the examples, I could not make head or tail of the majority of them.
I wanted to fairly straightforward but the majority of them they're just so so so obscure.
And I thought well if I can shed a little bit of light on them that might be helpful to others then it's worth having to go out.
There's quite a number of them in the GNU manual but I'm only covering three in this show.
So if you want to get deeper then you're on your own or you can call on me and I'll do another show.
But I doubt that's going to happen.
So we looked at some of the less frequently used commands last time.
We skipped a few so I'm going to fill in those gaps today or some of them anyway.
There are a bunch of commands for inserting text instead.
So far we've just been modifying stuff and printing it out and so forth.
But we haven't got anything for inserting stuff.
And the first command is the C command and it's rather weird.
It's normally written out as if you're typing your said commands into a file.
So you see the letter C on the first line and it can have addresses associated with it.
All of the address types that we saw in episode two and the C is then followed by a backslash.
Then on the next line is the first line that you want to insert followed by a backslash if there's more lines.
And then it carries on to the last line which doesn't have a backslash on the end.
So it's all the lines including the C line have to have backslashes on the ends of them except for the last line.
So what it does it deletes the lines from the input stream or the file that you're working on.
It deletes the lines that match the addresses and it replaces them by the lines that follow the C command.
Since it deletes the pattern space when you issue a C command a new cycle begins.
So another line is brought in and whatever.
All the things that we talked about in earlier episodes.
Now the C command can be used on the command line, direct on the command line, but not all that usefully.
If you use it you can't then follow C with any other said commands because there's no easy way to signal that that's the end of the insertion.
And here is the next command.
But you can do it by using multiple minus e quoted strings on the command line.
So I've got a demonstration here.
The demonstration I'll just read out the command.
It says space minus e space open quote one C backslash line remove close quote minus e quote three q lowercase q close quote.
And then the file said demo one dot TXT.
So what that saying is online one replace that line with the words line removed and then online three.
That's the second minus e expression quit.
So what you see is the first line to replace the line removed then the next two lines are printed out normally.
But you can only generate one line through to the C command in this way.
You can't do this in normal in the standard said.
So if you wanted to generate more than one line and I've got an example of trying to do that.
So I've got a one C in my expression backslash asterisk asterisk asterisk asterisk, backslash do not read.
It doesn't generate two lines.
It's just one line containing all that text with the backslash is removed.
so but you can insert escape character so backslash n is valid so that will
generate a line break so you do you can generate multiple lines or the third
example in this the notes the long notes here shows it being used to add the
line censored with asterisk rounded do not read instead of line one of said
demo one dot txt but the c command is best used in a file of said commands and
I've put one together which is called demo five dot said scd which is available
and linked from the the notes available on the hpr website so what I've done in
the notes is I've shown the the said file demo five dot said and I've I've
listed it using the nl command with various arguments to to control the way
that the numbers are put out so you'll see there are seven lines the first
five lines consist of one c so that means on line one operate the c command
replace line and then on the end of the c commands are backslash then a line
of a few hyphons backslash this next line is line three this line has been
censored backslash and then so and so lines four then line five is the last
line and it's a bunch of hyphons is a blank line then three q so basically this
is a similar said program to the ones have already seen and when you run it you
would run it as said space minus f space demo five dot scd space said demo one
dot txt and instead of line one of the file you get this line has been
censored with lines of hyphons around it and so on and so forth you could have
done all that on one command line using backslash in new line things but
that's very much a a GNU said extension so this particular example would work
with standard versions of said on bsd or unique systems or whatever not sure how
many systems don't use GNU said these days but they do exist I'm sure so
after the the c command let's look at the a command this is similar it consists
of the letter a followed by a backslash followed by multiple lines each
followed by a backslash except the last one it'll take any of the address
types that we normally find and what happens here is that the line that's
currently been read in is processed as normal and then it's followed by the
lines associated by with the a command and it happens at the end of the cycle
so basically it's appending this whatever you text you're providing after this
particular line so if you put addresses on the a command then it will only do
this for the lines in question or if you don't put a line number at all it will
apply to every line of the file so I've got an example of using it on the
command line using the one line form as we looked at for the c command and this
one consists of the command said space minus e space quote one a backslash
chickens quote space minus e space quote one q quote space said demo dot t x t
so what this does is it prints out the first line hackaburlic radio and then
it follows it with the word chickens and it only applies to the first line and
the one q says after processing line one stop so it just does it once and you
only see one line of the file the next example shows how you could add lines
containing just a single hyphen after every line of a file I don't know why
you'd want to do that but it's effectively double space so maybe it's a bit
easier to read or something so this one the command is said space minus e
space quote a backslash hyphen quote space minus e space quote three q quote space
said demo one dot t x t so this prints out the first second and third lines of
the file and after each one is a black is a line that only contains a hyphen
in the first first column so you can see that the a command is being applied to
each line and the hyphen is being written out after the line so finally in this
group of insertion command we have the i command this is both I should have
said as far as the a command you won't find that in other versions of all
said it's only the GNU said and similarly with the i command it's only only
available in GNU said and it's got the same structure as the cna command so it's
the is i backslash line one backslash line two and so forth the last line not
containing not ending with the backslash takes addresses in the same way and what
happens is all that the lines covered by the addresses or all lines at the
file if there's no address have the text which has been inserted by the i
command in added in front of them so using the one line form as we looked at
with the c and the a commands you can put escape characters like backslash
in for new line so I've got an example that simply adds the hyphen that the
line contains just a hyphen in front of every line and printed out from the
file and this is this is a very very similarity example I gave for the a command
except the hyphen hyphenated line precedes each line so I won't enlarge on
this one I thought I'd finish off with something that might be might be useful
I don't know and personally I've never found a use for these these particular
commands but what I've done is to construct an example that inserts an open
square bracket before every line and a closed square bracket after every
line and it just does it to the first three lines of said demo 1.dxt and it's
simply done by the command said space minus e space quote i backslash open
square bracket quote space minus e space quote a backslash closed square
bracket quote space minus e space quote 3q quote space said demo 1.dxt so
that's going to print out the first three lines of the file then it quit and
in front of each one I'll put an open square bracket after each one I put a
closed square bracket so that's all there is to say about this really it's it's
a bit weird it's it's could be could be useful in some contexts you could use it
to since the addressing type that you can use there is any of the addresses that
we've talked about in episode two you could use it as a way of highlighting a
given line in a file by by putting something something some sort of highlighting
sequence in front and after it or just in front of it or just after it or whichever
you to ever takes your fancy so I guess it could be useful from that point of
view so I'm going to go on to three elements three commands within GNU's
said which are classified as guru level these are pretty obscure the
section only in the manual the GNU said manual is called commands for said
gurus and it says there's an introduction in most cases use of these
commands indicates that you are probably better off programming in something
like awk or pearl but occasionally one is committed to sticking with said
and these commands can enable one to write quite convoluted scripts so
basically it's warning off using them I think but I'm including them here
because quite a number of the examples in the manual use use them so I thought
it would be useful at least to skim over them so you had some sort of
understanding when I go on to explain some of the examples so said commands
are really a sort of programming language very very basic sort of and rather odd
programming language and as such it's possible to use labels and conditional
and unconditional branching within a said script defining a label consists of a
colon you just put a colon on a line by itself followed by a character
sequence and the character sequences the label so colon x is a gives an example
or colon hello or anything that makes you fancy can be used as a as a label
it's no addressing associated with it makes no sense to have any addresses
and it's simply a point a way of marking a point in the script for branch
speaking of which the b command which consists of b followed by an optional label
it causes an unconditional branch to that label so it makes the flow of
execution jump backwards or indeed forwards to the label in the script the b command
can be used without a label in which case it just stops the current cycle
and starts a new one third example we'll look at in this this particular episode
reverse characters of lines it's called uses this then there's the t command
and this takes the form pretty much the same as the b command letter t
lowercase t it was lowercase b as well I should have said
lowercase t followed by a label so it causes a conditional branch to the label
and it happens only if there's been a successful substitution with an s command
since the last input line was read or conditional branch was taken if you admit the label
which you're quite at liberty to do then it causes the next cycle to start and again this is used
in the third example towards the end of the notes so the final category is commands which are
specific to GNU said there's actually quite a number of these I've skipped them all but one
you're probably relieved to hear that you can check them out and I've given pointers to them
in the GNU manual in the in the notes the command I'm going to talk about quite briefly is
the F command this is a capital F and what it does it prints out the file name of the current
input file it does this with a new line at the end of it so I've got an example here which
consists of said space minus e quote one open curly bracket capital F semicolon q lowercase q
close curly bracket quote space said underscore demo one dot txt so what that's actually saying is
online one there's a group that is to be obeyed the curly brackets and the group of commands
consists of the capital F command which prints the name of the file and then a q command which
stops processing because in this example said is running in its standard read and print mode or
to print then the first line of the output is printed so what you see is output that command
is said underscore demo one so that's the name of the file is being processed and it's written
to the output followed by the first line of the file it's as simple as that it is quite useful
though and there is a show in the queue talking about some of the features of the bash scripting
language where I've demonstrated that it's usefulness okay let's get on with the examples from
the GNU manual there's a bunch of these and that there are they are on the whole pretty obscure I
would say I've chosen three which I think are probably a little bit more understandable and seem
to make a bit more sense it's entirely my arbitrary choice I guess first one is to do with centering
lines and it centers all the lines of a file in a width of 80 columns so it just it just
pads them with spaces on the front to make sure each line fits into the center of a given line
now what I've done here is to I've slightly modified some of these to make them a little bit more
readable and also to reflect that most of my audience are going to be using Linux and things
like paths and so on are going to be different in Linux and in this particular case the file begins
with usual hash mark exclamation mark often called crunch bang or hash bang followed by the path
to said which in my cases slash bin slash said then a space then a minus f you need the minus f
because you need to tell said to read its commands from the file that it's actually been invoked from
and this particular script is available and as part of this this episode and it's called center.set
cnt re spell the British way and I've linked to it from the notes so I've listed it out and I've
used a feature of mark down and pan dock which lets me do a numbered listing so I can talk about
the lines in the notes so the script begins with a group of commands which span from lines four
to nine and I apologize that you really do need to be to have this in front of you when you
when you're working your way through this you're sitting on the bus or something driving home or
something then it's not going to be very easy to deal with but I can't think of a better way
of achieving this so in the group the first line line five of the script is an x command and you
remember that this exchanges the pattern space and the whole space and on line one this is
this operates on line one it means one line will have been read from the file and it will place that
line into the whole space the whole space because it's line one will be empty so now the pattern
space is empty then the next line consists of a substitute command which is an s slash circumflex
dollar slash then 10 spaces so what that saying is the line in the pattern space which consists
merely of the the dummy character start of start of space start of line and the other one
end of line the dollar is to be replaced by 10 spaces so basically it's it's a way of adding 10
spaces to to the buffer line seven is another substitute now there are multiple spaces at 10 spaces
in the in the buffer so this command is s slash circumflex dot asterisk dollar slash then there
are eight ampersands and then a close slash closing slash so what that does is it replaces the 10
spaces in that pattern space buffer by itself the ampersand eight time so it creates 80 spaces then
line eight is another x command so you won't be surprised to know that it swaps the 80 spaces
into the whole space and the pattern space what was in the whole space the the luck first line
is swapped back into the pattern space so that was a brief process of creating 80 spaces which are
stored in the whole space we move on to line 12 which simply contains a y command y command if
you remember is the the way in which you can swap or transliterate was the term used translate
one type of character to another so in the original it contained a contained the word tab TAB
wise slash tab slash space slash but what that what that was just a convention to signify a tab so
I've replaced that in the copy I've handed out to you with back slash t that escape sequence so
in other words it's just going to go through the whole line and replace any currencies of tabs by
spaces suppose it's bit of a cop out in a way because it's it's gonna destroy the effect of any
tabulation that you happen to have on that line but there you go line 13 is a substitute command
which is s slash circumflex space asterisk slash slash so what that's doing is it's removing
all leading spaces from the pattern space buffer line 14 is the equivalent for taking off
trailing spaces it's s slash space asterisk dollar slash slash so any number of spaces preceding
the end of line is is removed so these lines 12 to 14 are going to be executed here on every line
for the moment we're just imagining it happening to line the first line we're now on line 17
this is a capital G command which appends the contents of the whole space to the pattern space
preceded by a new line it just makes copy of the whole space and appends it to the pattern space
well the whole space it's got 80 80 spaces in it so it's going to add to the pattern space a new
line and 80 spaces line 20 is an s command substitute so it's look as s slash circumflex back slash
open parenthesis dot back slash curly open curly bracket 81 back slash close curly bracket back
slash close parenthesis dot asterisk dollar now that regular expression is a grouped one that's
the back slash parentheses and inside these parentheses it's saying dot which is any characters you
will recall 81 of them and it's anchoring them at the start of the line so the first 81 characters
are to be taken off to be to be matched at the start of the line it's actually matching the whole
line but the the grouped characters are the first 81 and the rest of the s command is back slash
one slash so it's replacing the first 81 the whole line by the first 81 character all right it's
basically trimming off anything beyond character 81 that's line 20 I don't know if I said that
so finally we've got line 23 and this one is quite tricky and take it took me a while I sort of
understand about I have to really think hard about what it's doing this is another substitute command
and it goes like this s slash circumflex back slash open parenthesis dot asterisk back slash close
parenthesis back slash in back slash open parenthesis dot asterisk back slash close parenthesis back slash
two and then the slash so let's talk about that regular expression so it's grouping all of the
characters from the start of the line up to the new line that's in the first group then it's
matching the new line and then here's the really tricky bit remember there's there's a bunch of
spaces after the new line and that bunch of spaces are such that everything the whole line is now
81 characters long so that's 80 characters plus the the new line so the remainder of the line
will be spaces after the new line and where by the regular expression has been formed is to say
all the characters a bunch of characters an arbitrary number of characters is to be placed
into a group and then that group is repeated so that forces the regular expression engine to select
only as many spaces as can represent the group and a reiteration of the group in other words it
it chops the the spaces after the new line into two sections two equal not it amounts then the
rest of the substitution expression consists of back slash two back slash one so it's taken back
slash two which is half of the spaces and is stuck in front of back slash one which is the the
stuff before the new line so the effect of that is to place half of the spaces in front of the
actual text thereby centering it it's still hard to find this hard to explain I can sort of visualize
a bit it's really hard to turn into words I hope you you've got got some understanding from this
but I hope I've made it clear enough that you've you understand it my text perhaps might
explain a bit of the night just tried to do so this script is written to center stuff in 80 columns
if you wanted to use a wider number of of columns then you'd have to change that s command
that generated the 80 spaces and you'd also need to change the expression on line 20 where it uses 81
to compensate for that it's not very flexible however it's clever and it's a useful demonstration
what can be done with said if it were me I would not be using said to do something like this I'm
sure I would come up with something better as the manual itself says using org or pearl or something
so the second example from the GNU manual is something which emulates the unix command
attack which is a reverse version of the cat command and it simply lists a file backwards
named the section reverse lines of file this is quite well documented actually some of the one some
the examples are really really cryptic I find but which is why I've explained this one it is
um it's pretty well done however I thought it would be useful to just drill down into a little
bit more than is in the manual so this one's available in a file call tacktac.set and it's part
of the the bundle of stuff for this show so the first line is the hash bang line and it's using
slash bin slash said and in this case it's not only using minus f it's also using minus n and
that turns off auto printing we'll see why shortly so it's pretty small there are only three
commands so on there's quite a number of comments online seven is an address which is one
and a single command but the one is followed by an exclamation mark so it means all lines other than
the g command depends new line to the contents of the pattern space the capital g that is and then
depends the contents of the whole space to the pattern space so whatever is in for all lines that
are not one take whatever is in the whole space and stick it on the end of the pattern space
then line 10 consists of a dollar and a p and that means on the last line print whatever is in
the pattern space remember we've switched off auto printing so so that will only trigger on the last
line line 13 simply consists of the command lowercase h and what that does is replace the contents of
the whole space with the contents of the the pattern space and this is done it has no address
on it so it's done for every input line so I thought it's worth just explaining the algorithm
in a bit more detail the first line that's read doesn't do anything other than trigger the h command
online 13 and this means that the line is simply stored in the whole space then the second and
subsequent input lines trigger the the capital g command online seven and so if we were dealing
with the the second line for example we would append a new line to the pattern space and then a
penned line one which was stored previously in the whole space to it then the h on line 13 is
invoked and the pattern space which is now in in the order line two then line one is stored in
the whole space again and that way every line gets the reversed hold space stuck on the end of it
and is then stored in the whole space and when the last line is read the g command online seven
we triggered as before and it will append the whole space a contents again so that the pattern space
now holds the entire file in reverse order then because it's the last line the p command will be
invoked for the only time and it'll print everything out so what you'll see is the entirety of the
file which has been stashed away inside buffers in and said printed out I said that it bothered me
slightly that after doing that p command it will also invoke the h command again which is just
a slightly obsessive thing and I'd have been tempted to to have written line 10 with a with a q
on it so it would stop stop it doing that but it really makes no difference at all I'm ashamed
to have you written this now okay here's the the third and last example if you're finding this
a bit much not surprised if you are because it's not easy it's not even easy to read this stuff
when I was writing these notes I was thinking oh yeah yeah wait so obviously it's so clear now
and it was a month or so ago I wrote this and I'm sort of going through it now and now I read it again
my god it's it's really really obscure said it's brilliant said is a fantastic thing but I don't
ever see me using it in this way other than to say to somebody look what I did isn't this clever
and everybody whoa that's amazing but they're asking under their breath anyway I'm saying too much
so reverse characters of lines is the last one and this emulates the standard units command called
RAV which reverses takes each line reverses the characters now I've put the script into a file called
reverse underscore characters dot said and it's available with the bundle for this episode and
I have obviously changed the path so it can actually run you can run it and the other thing I did
was there were in the original example there are actual new lines used that is you know that the
line is broken by typing a new line in it whereas I've used a backslash ends so in doing this
I've made it a new specific so let's dive into this one the first line is the hash bang line
which is it's just got a minus f on it no no suppression of auto printing and the first command
is online three and it consists of an address and a single lowercase b command the address consists
of a string which which is simply contains two dots inside slashes and then it's negated it's
got a it's got an exclamation mark after it so what it means is any line that doesn't contain
two characters the b command if you remember is an unconditional branch to a label except there's
no label here basically it means ignore or just carry on to the next one so whenever a line
containing fewer than two characters then the unconditional branch simply triggers a new cycle
which will print the line and ignore the rest of the script and carry on with the next line so
a line with only one character in it there's not much point in reversing it is there that's the logic
wow that's a lot ended up being quite a lot to say of a line with just a small number of characters
said like machine covenants and you get to this level anyway let's move on line six is a substitute
command which the regular expression part is a circumflex dot star or asterisk dollar it means
the entirety of the line so we know it's going to have two characters in it or more and it's
replaced by backslash n ampersand backslash n slash and so what that's doing is it's embedding
the line in between two new lines there's a new line placed at the front and one placed at the end
now line ten is a weird one it's it contains the t command which is a conditional branch
and the label is branching to is x well line eleven contains colon x so there's a line ten simply
says jump to the next line and this is documented in the GNU manual with the text this is often
needed to reset the flag that is tested by the t command this is a thing simply to make the t
set itself to a known state and I think that's probably because there was a substitution that
happened just before that t command and remember the t command has to it's condition is that there
has been a success for substitution prior to it being invoked so I guess that makes sense so I
tried taking it out see if it made any difference and it didn't seem too so anyway line eleven is
this label colon x they keep their labels pretty short in these examples in line twelve is a
substitution s slash backslash open parenthesis backslash n dot backslash close parenthesis backslash
open parenthesis dot star that last risk I should say backslash close parenthesis backslash open
parenthesis dot backslash n backslash close parenthesis so that expression was three groups the
first group was the new line and the character after it the second group was all of the characters
other than the ones that match the other groups so basically the characters between the new lines
but excluding the one we already picked up after the new line and then the group three is like group
one in reverse it's the character before the second new line three groups so we simply output
them as the second half of the s backslash three backslash two backslash one whatever was whatever
the character was just after the left most new line becomes the character on beside the right
most new line when they get swapped over so these new lines quite important as markers for which
characters are to be swapped and in swapping them the character to the right of the first new line
becomes the character to the right of the second new line so it's out it becomes outside the new
lines the same for the for the group three the characters that were inside just inside the two
new lines become the characters just outside the two new lines and swapped over does that help
the rest of the line is left alone so you'd say okay okay well it's just swap two characters well
yeah but it's using this labeling thing this jump to label thing because line 13 consists of
another tx so this is one of these conditional jumps two back to label x and if the substitution
was successful it will fire that particular jump so it goes back to label x where it simply
executes that s command on line 12 again so it's going to be looping around
obeying this command as many times as is necessary until there are no more
characters between the new lines so the last line of this this script is simply on line 16 it's
a substitute where it goes slash s slash back slash n slash slash g so in other words remove all
new lines or both new lines and then because we're in auto print mode the thing will be printed
now took my while to get my head around what was happening here and part of the process of
finding out was to put some debug stuff into the script just to be certain that it was doing what
I thought it was doing and I've offered you an alternative version of this script which I've
called reverse underscore characters underscore debug dot said which contains L commands remember
the L command lists the current contents of the pattern space and it does it in a way that makes
it easier to spot what's actually going on puts dollars to show where new lines are and so on and
so forth so I've listed it out but I won't go through it basically I've done it so that I put
the L's in so that you see the line after it's had the initial new lines attached to it remember we
put a new line at the front and the back of the line then I've got an L command in this loop with
the tx through the branching back to the x label so every every time it iterates through that
loop it prints out the result when you run that what I did was I fed it echoed to it the alphabet
and lowercase pipe that to reverse characters underscore debug dot said which is executable and
it demonstrates you can actually see the characters inside of the new lines being flipped such that
outside the new lines and they've been reversed so that ABC etc gets reversed to z y x w I hope
you find that that's useful it's only made it clearer to me as to what was going on maybe you
can visualize these things better than I can and the last yeah the last line in the block of
lines in the in the notes is the the auto-printed result of doing all of the swaps so that's it
that we've done it we've covered the three at least of these very very hairy sand examples actually
there's a lot more hairy ones than that but I'm sure I'm sure you would be switching off very very
fast if I were to do any more and to be honest I have not looked at any more of them and understood
them so last item of business is I said a quiz in the last episode and I said I give you the answer
well the quiz was to take famous said demo 1 dot txt only concentrate on the first line
and take the first letter of each word place it at the end and follow that with the letters a y
so that pig p i g becomes igpe i g pa y this is called pig latin and latin becomes atin lay
I said skip one and two letter words I think I said skip three as well in the in the last episode
but then my example of manipulating the word pig would be silly because it would be skipped
so one and two makes more sense and the reason you you don't want to do anything with with one
letter words is pretty obvious because it would turn the single letter a to a a y which is
make a lot of sense plus also a number of the words in that file the first line of that file
of a capitalized and I said don't bother about them fact that the capital is end up in the wrong
place is not a big deal so here's my answer what I did was said space minus n e so we're we're not
all to printing quote 1 s slash back slash up parentheses back slash b back slash w back slash
close parenthesis back slash open parenthesis back slash w that's locust w I should have said
back slash open curly bracket two comma back slash close curly bracket back slash close
parenthesis slash so that's the whole regular expression let's see if we can explain this
the first group in parenthesis contains back slash b back slash w so back slash b is the thing
that it's the strange boundary marker meaning whenever there's a word the back slash b marks
the the edge of it the transition from non word to word and the back slash w is simply the first
character of a word or a character of a word so we're looking for in this first group the first
character of any word the second group contains back slash w and that's followed by a quantifier
in curly brackets which is two comma and you'll recall that that means I want word characters and
I want two or more of them don't care how many there are in total I have no upper limit but two
has to be at least two so that's that's how we're skipping the two character words so in doing
this we're matching a sequence of the start of a word and the rest of the word and as soon as
the space is encountered the scan for the the multiple word word characters stops so in the group
we have the first letter of a word followed by the rest of the word and then the rest of the s
expression consists of back slash two back slash one a y close slash so that means put them
the second group which is the bit of the word after the first character follow that with the
first character back slash one and then follow that with a y and then the qualifiers to the flag
so I should say for the the s command rg and p it's to do this repeatedly across the line
and when it's finished printed and the rest of the line is space said demo one dot txt so the result
is that you get the words instead of hacker public radio get aca hay oblique pay adior ray then
it starts getting a bit silly because it's it changes hpr to prh a which is just plain silly
but you can't do much more about that with this is quite simple I mean the said solution is simple
and then it is an an are not touched across their two character words then internet internet is turn
into internet EA which is awful I'm sure there I know that there are rules in this pig Latin
that shouldn't allow that to happen you can't really write that sort of stuff in said radio turns
into adior ray which I think is quite nice but shows shows sorry turns into how say which is silly
and podcast do not quite like podcast turns into odd cast pay that turns into hattay to silly release
releases big turns into elisa's ray anyway I know at least one other hpr listener quite
likes pig Latin but this is not the best examples but you know it it answers the the question
so who won well sadly there were no winners because there were no entries come on guys have you
actually got to sleep that's what I said in my notes here probably just as well I'm finishing
this series here because I think I probably sent everyone to sleep several episodes back so
if you've been snoring through this one well good luck to you and hope you had a good sleep
so that's it I finished that's it no more said and I hope you did enjoy it hope somebody got
something out of it anyway okay that's it bye bye
you've been listening to hecka public radio at hecka public radio dot org we are a community podcast
network that releases shows every weekday Monday through Friday today show like all our shows
was contributed by an hpr listener like yourself if you ever thought of recording a podcast
and click on our contribute link to find out how easy it really is hecka public radio was found
by the digital dog pound and the infonomican computer club and it's part of the binary revolution
at binrev.com if you have comments on today's show please email the host directly leave a comment
on the website or record a follow up episode yourself unless otherwise status today's show is
released on the creative comments attribution share a light 3.0 license