Well, I guess I am in the minority. I use them all the time. I indent
the branched code and I use a +n for each instruction. Between the
number of + characters and the indention, it's easy to spot. I also
limit it to only 1, 2 or (seldom) 3 instructions being branched.
FOOBAR
BP *+4+4
WOMBAT of 4 bytes
XYZZY
FOOBAR
BP *+4+4+6
WOMBAT of 4 bytes
BATWAM of 6 bytes
XYZZY
Tony Thigpen
Gary Weinhold wrote on 07/06/2015 09:48 AM:
What he coded would have been
FOOBAR
BP *+8
WOMBAT of 4 bytes
XYZZY
His point was there were no labels (or marker comments) for this method
(clutter to his way of reading code), but blank lines in the source for
readability.
On 03/07/2015 1:42 PM, Paul Gilmartin wrote:
On 2015-07-03, at 09:55, Robin Vowels wrote:
From: "Gary Weinhold"
Sent: Friday, July 03, 2015 10:46 PM
We have ended up with many of these in our code because a senior
developer felt that too many single-usage labels cluttered the code
and made it less readable. (He also required that a blank line
follow every branch statement and precede the target of every
branch, whether labeled or not). ...
Simply awful.
Senior? Senile?
It sounds as if he coded:
FOOBAR
BC CC,*+8
WOMBAT
*+8 DS 0H
XYZZY
I'd choose to do nothing of the sort, but if compelled, I'd do quite
the opposite and treat the implied "if" as a basic block:
FOOBAR
BC CC,*+8
WOMBAT
*+8 DS 0H
XYZZY
(Grrr... HLASM allows me to use blank lines for legibility, then
omits them from the SYSPRINT for illegibility| Or is there an
option controlling this? But it's better than FAP, which treated
a blank line as "DC F'0'".)
-- gil
