Unintentional wording regressions relative to 1.2

[Former user]

Originally reported on Google Code with ID 81 ``` I've done a complete read-thru of the 1.3 spec to catch any glaring issues. I noticed this wording change to upc-terms-and-defs.tex, relative to 1.2:

\sterm{affinity}% - logical association between shared objects and threads. Each element of data storage that contains shared objects has affinity to exactly one thread.

I assume this was an unintentional change, as the resulting text is an incomplete sentence that makes no sense (which is why I noticed it). I'm committing a change back to the original 1.2 wording now, but I'm concerned there may be other such unintentional changes that are harder to notice. Can we somehow audit how this happenned? Can we somehow directly diff the text of 1.2 and 1.3 draft 0.9 to find other such problems? ```

Reported by `danbonachea` on 2012-08-10 23:14:19

[Former user]

``` Attached, is a 'diff' file that lists the differences between the original .tex file and the last "pre-split" version of that file. Eliding the lead in to the sterm{} happened in revision 20. Off-hand, I didn't see any other substantial differences, but would appreciate a separate review.

Note that some of the apparent cosmetic changes made HTML rendering more readable. My guess is that with recent changes in the preamble, we might have to rework some of the markup to bring back HTML. However, that is a minor issue at the moment.

```

Reported by `gary.funck` on 2012-08-11 00:12:27

<hr>

• *Attachment: [upc-lang-spec-pre-split-diffs.txt](https://storage.googleapis.com/google-code-attachments/upc-specification/issue-81/comment-1/upc-lang-spec-pre-split-diffs.txt)*

[Former user]

``` I've laboriously read thru this diff and found only one other questionable change:

-\np {\tt \#include <upc.h>}
- {\tt \#include <upc\_collective.h>}
-{\tt void {upc\_all\_reduce}{\em T}(shared void * restrict dst,
-\phantom{void }shared const void * restrict src, upc\_op\_t op, size\_t nelems,
- \phantom{void }size\_t blk\_size, {\em TYPE}(*func)({\em TYPE}, {\em TYPE}),
- \phantom{void }upc\_flag\_t flags);}
- {\tt void {upc\_all\_prefix\_reduce}{\em T}(shared void * restrict dst,
- \phantom{void }shared const void * restrict src, upc\_op\_t op, size\_t nelems,
- \phantom{void }size\_t blk\_size, {\em TYPE}(*func)({\em TYPE}, {\em TYPE}),
- \phantom{void }upc\_flag\_t flags);}
+\npf +\begin{verbatim} +#include <upc.h> +#include <upc_collective.h> +void upc_all_reduce_<<T>>( + shared void * restrict dst, + void shared const void * restrict src, + upc_op_t op, + size_t nelems, + void size_t blk_size, + <<TYPE>>(*func)(<<TYPE>>, <<TYPE>>), + void upc_flag_t flags); +void upc_all_prefix_reduce<<T>>( + shared void * restrict dst, + void shared const void * restrict src, + upc_op_t op, + size_t nelems, + void size_t blk_size, + <<TYPE>>(*func)(<<TYPE>>, <<TYPE>>), + void upc_flag_t flags); +\end{verbatim}

This was probably intended to beautify the formatting, but the resulting output looks terrible and is significantly uglier than the 1.2 version. It looks like a space vs tabs problem in the verbatim. This needs to be fixed.

I was also annoyed to find that someone had re-wrapped some of the lines in the terms section and removed some trailing spaces on lines in appendix B, despite the fact such whitespace in the TeX file is completely irrelevant. I would like to *strongly* prohibit any such useless whitespace changes in the future, no matter how ugly the tex might look - it's far more important to see what has *actually* changed, and commits touching lines that aren't directly relevant to the change being made should be rejected and rolled back without further discussion.

```

Reported by `danbonachea` on 2012-08-11 01:55:08

[Former user]

``` Dan, thanks for the review. The verbatim change likely relates to HTML rendering. I will correct the space vs. tab issue.

Regarding a policy that forbids removal of white space and re-filling paragraphs: agreed.

```

Reported by `gary.funck` on 2012-08-11 02:29:50

[Former user]

``` "The verbatim change likely relates to HTML rendering. I will correct the space vs. tab issue."

I believe this is the only issue which is blocking the release of "Draft 1" (ie the "semantically equivalent to 1.2" version). I have not looked at the HTML output to see what presentation issues may be present there. ```

Reported by `danbonachea` on 2012-08-12 18:51:19

[Former user]

``` Re: comment #2, fixed in revision 79, by converting tabs into spaces.

Also, per the commit message:

"added use of package moreverb which will support tabs directly within a verbatimtab environment and has a listing environment which will add line numbers. Although unused at the moment, this enhanced verbatim package may prove useful in future updates to the spec."

```

Reported by `gary.funck` on 2012-08-12 20:14:27

[Former user]

``` Set default Consensus to "Low". ```

Reported by `gary.funck` on 2012-08-19 23:26:19 - Labels added: Consensus-Low

[Former user]

``` All the problems mentioned were resolved prior to Draft 1. ```

Reported by `danbonachea` on 2012-09-14 18:56:20 - Status changed: `Fixed`