From: | Josh Soref <jsoref(at)gmail(dot)com> |
---|---|
To: | Heikki Linnakangas <hlinnaka(at)iki(dot)fi>, Andres Freund <andres(at)anarazel(dot)de> |
Cc: | pgsql-hackers(at)postgresql(dot)org |
Subject: | Re: Possible spelling fixes |
Date: | 2017-02-06 10:52:48 |
Message-ID: | 20170206105248.4902991.81306.2285@gmail.com |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-hackers |
Heikki wrote:
> I pushed most of these. Except for the below:
> optimisation -> optimization et al.
> Most of our code is written with the American spelling,
> but the British spelling isn't wrong,
> so I don't want to go around changing them all.
Sure
As you'll see, my approach is to aim for consistency. If you used en-GB 99% of the time, I'd have offered a change to enforce that. I have a personal preference, but there's no obligation, and I understand the potential costs churn entails (I see you backported to branches).
> NUL-terminated -> NULL-terminated
> When we're talking about NUL-terminated strings,
> NUL refers to the NUL ASCII character. NULL usually refers to a NULL pointer.
This wasn't even in my original set (i.e. The dictionary I'm using didn't consider NUL to be misspelled). I ran across it while splitting comments out per Andres and figured I'd offer it as well.
> We're probably not consistent about this,
Hrm, I was going to say "that's correct, you aren't", but rereading, I realize that I'd have to review every instance in order to prove to myself that statement. I could make the weaker argument that "nul-terminated" should be changed to either NUL-terminated or null-terminated . My general approach is to only make changes when I can detect an inconsistency. And I generally tend toward "majority rule".
Here, I think the corpus has 4 spellings, and it sounds like it should only have two, but probably (NUL- and null-) not the two I picked (NULL- and null-).
> but in this context, NUL-terminated isn't wrong, so let's leave them as they are.
But that's OK. My goal in posting these is to encourage people to consider their code.
>> Ooops -> Oops
> "Oops" is more idiomatic, but this doesn't really seem worth changing.
Technically oops is in dictionaries whereas the other isn't, but I understood the intent.
> Maybe "Ooops" indicates a slightly bigger mistake than "oops" :-)
That seemed like the intent. It's certainly not unreasonable to retain it. It's also why I generally offer a queue, so people can reject families of changes.
>> re-entrancy -> reentrancy
> Googling around, I can see both spellings being used.
Both are used, but reentrancy is in most dictionaries (and encyclopedias) and is the form that's used in instruction (certainly it was when I studied in university, and it isn't likely to regress). It's akin to email vs e-mail. Once the dashless form becomes accepted (within a domain [1]), it's the correct form, and the other was merely transitional.
> "Re-entrancy" actually feels more natural to me, although I'm not sure which is more correct.
> Let's leave them as they are.
Sure
>> passthru -> passthrough
> "Passthrough" is clearly the correct spelling (or "pass-through"?),
The former is also present in the codebase. (I didn't look for the latter, for the same reason as the previous note.)
> but "passthru" seems OK in the context, as an informal shorthand.
My goal is consistency. If you always spell a concept a single way, then grepping for that concept is easier and more reliable.
I personally recognize quite a few flavors, because they're usable for talking to Coverity / Purify.
>> - * Temporay we use TSLexeme.flags for inner use...
>> + * Temporary we use TSLexeme.flags for inner use...
> Looking at the code real quick, I couldn't understand the original
meaning of this. Is it:
> * DT_USEASIS is a temporary value we use for something. For what?
> * DT_USEASIS is used temporarily for something.
> Does this mean, "temporarily" until we get around to write the code differently, or does
> it happen temporarily at runtime, or what?
> Just fixing the typo doesn't help much here,
> and I'm not sure if it should be "temporary" or "temporarily" anyway.
Apparently I didn't look at this one much at all. I believe temporarily is the intended word (fwiw, I originally mis-corrected directly as directory, that I did spot before submitting). And probably as a runtime concept.
But I'm not volunteering to fix all comments in the project ;-). After spelling fixes, I'm more likely to try actual bugs / usability issues. I have a specific bug which bit me, but fixing that would require more effort than a spelling pass and more cooperation. I tend to do a spelling pass to determine if the more expensive activity is viable. So far, the project is welcoming :-) so, perhaps I'll manage to write the real fix...
> I wasn't sure if this changes the meaning of the comment slightly.
> An "UPDATE" in all-caps refers to an UPDATE statement,
> is that what's meant here? Or just updating a tuple,
> i.e. should this rather be "skip updating of the tuple" or "skip update of tuple"?
I'm not certain. I do understand that capital UPDATE is special. This one people more familiar with the project will have to resolve.
Fwiw, if it's the former, you could omit the "of".
> This "postsql" refers to the SQL dialect of PostgreSQL,
I had to look up the other dialect from that line to decide it wasn't a spelling error.
> rather than PostgreSQL the project.
> I don't remember seeing it called "postsql" anywhere else, though.
Nothing within the corpus I was changing shared that spelling, otherwise it too would have been changed :)
Oddly, this specific thing feels like a Deja-vu. I wonder if I started a spelling fix series for Postgres a decade ago or something...
> We hardly care about what was an error in postqual anyway,
> though, so perhaps this should be rewritten into something else entirely,
> like "This is not allowed by the SQL standard, but ok on PostgreSQL"
> (assuming that's correct, I'm not 100% sure).
> Or just leave it alone.
I'd encourage you to find something that's meaningful and correct.
> Thanks for the fixes!
You're welcome.
Thanks for the quick handling. Some projects take months. Or never respond.
> I was particularly impressed that you caught the typo in Marcel Kornacker's surname.
My tools identify both spellings as incorrect (and all possibly misspelled words are listed alphabetically), which means that I have the opportunity to choose a correct spelling -- generally I'll Google if I'm concerned because there is insufficient preference within a corpus.
Did you want me to submit emails for the remaining portions from
https://github.com/jsoref/postgres/commits/spelling
From | Date | Subject | |
---|---|---|---|
Next Message | Nikita Glukhov | 2017-02-06 10:55:27 | Re: [PATCH] kNN for btree |
Previous Message | Boris Muratshin | 2017-02-06 10:50:02 | Re: 3D Z-curve spatial index |