Problem with file trimming in Batch

[John Stockton]

I want to copy an HTML page file to another file, omitting both the heading+introduction and the tailpiece. For test I use a line of four hashes as separator, and a test file :

head
####
want
####
tail

My non-working test code is in :

@echo PRUNE.BAT JRS 2021-11-30+ ???

@echo.
@if not exist zzz\nul mkdir zzz
@echo 1111

@SET ZC=0
@echo. Start > ZZZ\ORIG.HTM
@FOR /F "eol=; tokens=*" %%J IN (PRUNE.DAT) DO @(
@echo Comment line='%%J' ZC=%ZC%
if [%ZC%] == [1] (
echo '%%J' >> ZZZ\ORIG.HTM
)
if [%%J] == [####] (
set ZC=1
echo === %ZC%
)
@echo ++
)

@echo 9999

@echo. Ended >> ZZZ\ORIG.HTM
@echo.

type ZZZ\ORIG.HTM

@echo PRUNE.BAT ends.

The apparent problem is that the line set ZC=1 has no effect, so that the wanted part is not copied. ( I intend to change it to set /A ZC=1+%ZC% which should mean that the tail part is not copied). Subsequently I may use a line   as an invisible separator.

What's wrong?


--
(c) John Stockton, near London, UK. Using Google Groups. |

[LangerTom]

I believe that your command
SET ZC=1
is working fine, but checking its result with
echo === %ZC%
is going wrong because both lines are placed between round brackets.
CMD.EXE is reading all code betwwen round brackets at once and is
resolving variables like %ZC% during code reading. At this time your
command
set ZC=1
did not take place yet.
To avoid this switch on DELAYEDEXPANSION and note variables like this; !ZC!

The following code should work (not tested):

HTH Thomas

By the way: working on html with batch is very hard stuff because html
code contains many poison characters like <>?*/"&

<code>

@echo.
SETLOCAL ENABLEDELAYEDEXPANSION

@if not exist zzz\nul mkdir zzz
@echo 1111

@SET ZC=0
@echo. Start > ZZZ\ORIG.HTM
@FOR /F "eol=; tokens=*" %%J IN (PRUNE.DAT) DO @(

@echo Comment line='%%J' ZC=!ZC!
if [!ZC!] == [1] (

echo '%%J' >> ZZZ\ORIG.HTM
)
if [%%J] == [####] (
set ZC=1

echo === !ZC!

)
@echo ++
)

@echo 9999

@echo. Ended >> ZZZ\ORIG.HTM
@echo.

type ZZZ\ORIG.HTM

@echo PRUNE.BAT ends.

</code>

[JJ]

The easiest way is to use Delayed Expansion as other have mentioned, but it
can produce more severe problem if the source data includes `!` character.

Without Delayed Expansion, move the code that use variables which are
updated within the same or parent command groups, into a subroutine. e.g. in
your case, change it like below.

@echo PRUNE.BAT JRS 2021-11-30+ ???

@echo.
@if not exist zzz\nul mkdir zzz
@echo 1111

@SET ZC=0
@echo. Start > ZZZ\ORIG.HTM
@FOR /F "eol=; tokens=*" %%J IN (PRUNE.DAT) DO @(
@echo Comment line='%%J' ZC=%ZC%
if [%ZC%] == [1] (
echo '%%J' >> ZZZ\ORIG.HTM
)
if [%%J] == [####] (
set ZC=1

call :DispZC

)
@echo ++
)

@echo 9999

@echo. Ended >> ZZZ\ORIG.HTM
@echo.

type ZZZ\ORIG.HTM

@echo PRUNE.BAT ends.

goto :eof

:DispZC
echo === %ZC%

[John Stockton]

On Tuesday, 30 November 2021 at 18:24:13 UTC, LangerTom wrote:
> I believe that your command
> SET ZC=1
> is working fine, but checking its result with
> echo === %ZC%
> is going wrong because both lines are placed between round brackets.

Not actually wrong, as it shows that %ZC% is not returning what I want;
and so explains why the next line, if [%ZC%] == [1] ( , is not doing
what I wanted it to do <g>.

> CMD.EXE is reading all code between round brackets at once and is

> resolving variables like %ZC% during code reading. At this time your
> command
> set ZC=1
> did not take place yet.
> To avoid this switch on DELAYEDEXPANSION and note variables like this; !ZC!
>
> The following code should work (not tested):

It does. I copied the whole lot, and it ran. All that I now want is to arrange that the second #### sets ZC to something other than 1, which should be easy (P.S. it was : set /a ZC=1-!ZC!). And I don't want the #### marks to appear in the output as rendered on the screen, but using &nbsp; instead should serve.

I control the input file; I can alter its content in any way that does not adversely affect what appears on the screen when the HTML is rendered. The present part, a pre-processor, just copies the HTML unchanged (I hope), but omitting blocks that I do not want in the output.

> HTH Thomas
>
> By the way: working on html with batch is very hard stuff because html
> code contains many poison characters like <>?*/"&

Understood. No problem here. It is the code which runs next that handles the HTML; it copies the wanted lines, which are those containing the name-strings provided as arguments, and separately lists the contents of the first href=" ... "> in each of those lines for XCOPYing as in my previous thread here. As the input is HTML, if there is a second such link on one of those lines line I will just insert a newline in the input - I think.

Thanks again! |