DOS ain't dead

Hi all,

my web hoster will disable any PHP versions older than 7.1 on December 1st.
This forum's software is powered by an older PHP version (5.6), so "DOS ain't dead" will be affected.

Over the years I applied some customizations to the code of this installation. Now I have to figure out, how to keep these changes or roll back to standard.

I will start migrating soon to a test installation and will keep you updated.

Cheers,
Robert

---
Forum admin

> I will start migrating soon to a test installation and will keep you
> updated.

Meanwhile I prepared a test system at https://www.bttr-software.de/forum2/.

Please do your testing in forum2 until November 15th!

Important: Anything posted to "forum2" will be lost, when I'm doing the final migration. So please post only test messages to "forum2".

To help you to distinguish between both forum installations, I set forum2's name to "[Test] DOS ain't dead" and its header and footer background colors to red.

Your logins from this forum should also work with "forum2", because I copied the underlying database.
But if you change your password in this forum, it won't be reflected to "forum2" and vice versa.

Testing may include:
- creating new messages
- editing existing messages
- changing your password
- changing your other profil data
- uploading images
- use of special chars like German umlauts

Please post your valuable feedback only here to this thread! -- Thanks in advance!

In "forum2" I changed character encoding from ISO-8859-1 to UTF-8 (Unicode). I therefore had to run many MySQL UPDATE commands to fix database entries with characters now escaped/masked invalidly. -- Not sure what this means for any Arachne or Links users. Are there any?

---
Forum admin

"Not sure what this means for any Arachne or Links users. Are there any?" -- I am a Links user, I did everything with Links 2.20 in the Test forum. I think the characters are displayed correctly, but the forum doesn't accept some characters from me (though I am sure Links sends the correct Unicode characters) and stores question marks instead. This includes the double accent/hungarumlaut letters of Hungarian, most Turkish special letters and almost every special letters of the Slavic languages with latin alphabet.

---
DOS isn't about why. It's about why not.

> "Not sure what this means for any Arachne or Links users. Are there any?"
> -- I am a Links user, I did everything with Links 2.20 in the Test forum. I
> think the characters are displayed correctly, but the forum doesn't accept
> some characters from me (though I am sure Links sends the correct Unicode
> characters) and stores question marks instead. This includes the double
> accent/hungarumlaut letters of Hungarian, most Turkish special letters and
> almost every special letters of the Slavic languages with latin alphabet.

Are you using Links in DOS?
Could you please try it from a Windows or Linux based computer?

From what I know now, there is more to do for full UTF-8 awareness. Thankfully Heiko is working on that at https://github.com/auge8472/My-Little-Forum-1

---
Forum admin

I did. The result is the same. You can see it under the post called "Káposzta" in Forum2. The post originally shows the result when I used Links for DOS. I commented from Windows a few minutes ago and it looks the same.

---
DOS isn't about why. It's about why not.

> I think the characters are displayed correctly, but the forum doesn't
> accept some characters from me (though I am sure Links sends the correct
> Unicode characters) and stores question marks instead.
>
> Could you please try it from a Windows or Linux based computer?

I tried in Linux with Firefox ESR. Same problem. It seems to accept and show the characters correctly ... until you post, then it mangles them all into '?'. (I had thought HTML escapes like ĉ were working correctly, but apparently not.)

---
Know your limits.h

> I did. The result is the same. You can see it under the post called
> "Káposzta" in Forum2. The post originally shows the result when I used
> Links for DOS. I commented from Windows a few minutes ago and it looks the
> same.

Thanks for your feedback.

---
Forum admin

> > Could you please try it from a Windows or Linux based computer?
>
> I tried in Linux with Firefox ESR. Same problem. It seems to accept and
> show the characters correctly ... until you post, then it mangles them all
> into '?'. (I had thought HTML escapes like ĉ were working correctly,
> but apparently not.)

That's a problem how PHP and MySQL interact with each other. I think, Heiko already fixed this in his upcoming v1.8 release. I'm waiting for his "Go!" to test it at forum2.

Trivia: Some source code for my little forum dates back to 2002 or even earlier. A little wonder it's still running.

---
Forum admin

News: I just looked at my e-mails. I've got an e-mail notification from the comment Rugxulo commented under my post in Forum2: every character show up correctly that are only question marks in the forum. I don't know if it is important, but perhaps a hint, when in the process the encoding is lost.

---
DOS isn't about why. It's about why not.

> News: I just looked at my e-mails. I've got an e-mail notification from the
> comment Rugxulo commented under my post in Forum2: every character show up
> correctly that are only question marks in the forum. I don't know if it is
> important, but perhaps a hint, when in the process the encoding is lost.

Thank you.

In the meantime to keep things simple for me and you, we should just stick to English letters, when posting to "forum2". Then I could happely make the switch before November 30.

Afterwards, when Heiko has finished version 1.8 of my little forum, there will be a new "forum2" for testing UTF-8 stuff and the like.

Any objections?

---
Forum admin

> > News: I just looked at my e-mails. I've got an e-mail notification from
> the
> > comment Rugxulo commented under my post in Forum2: every character show
> up
> > correctly that are only question marks in the forum. I don't know if it
> is
> > important, but perhaps a hint, when in the process the encoding is lost.
>
> Thank you.

The quoted line splitting is a bit flawed. (Normally, I'd just edit it myself or omit it, but just to remind you ....)

> In the meantime to keep things simple for me and you, we should just stick
> to English letters, when posting to "forum2". Then I could happely make the
> switch before November 30.

English letters? You mean digraphs ("u:")? Or is there other functionality you wanted tested there? (I know you're not directly talking to me, just wondering.)

Besides, (just to pretend to be smart) it's a modified Roman alphabet. (The Romans didn't have 'j', 'u', or 'w'. Even Esperanto calls 'w' "gxermana vo", i.e. "Germanic".)

Someone should ask MarcoV (or maybe Eric Auer) for more advice. They're more educated than I am. (I found a book called _Old English and Its Closest Relatives_ for sale online, which basically compares "Gothic, Old Norse, Old Saxon, Old English, Old Frisian, Old Low Franconian, and Old High German"!!) I pointed Eric to Beowulf online a while back, saying it (unsurprisingly??) sounded like Swedish (although I don't speak/read that either).

Sorry if that's somewhat off-topic, but that's what this reminds me of.

> Afterwards, when Heiko has finished version 1.8 of my little forum, there
> will be a new "forum2" for testing UTF-8 stuff and the like.
>
> Any objections?

No, but there was never much non-English posting here, was there? (I don't and shouldn't demand it, obviously.)

> > > News: I just looked at my e-mails. I've got an e-mail notification
> from
> > the
> > > comment Rugxulo commented under my post in Forum2: every character
> show
> > up
> > > correctly that are only question marks in the forum. I don't know if
> it
> > is
> > > important, but perhaps a hint, when in the process the encoding is
> lost.
> >
> > Thank you.
>
> The quoted line splitting is a bit flawed. (Normally, I'd just edit it
> myself or omit it, but just to remind you ....)

This also will be fixed in v1.8.

> > In the meantime to keep things simple for me and you, we should just
> stick
> > to English letters, when posting to "forum2". Then I could happely make
> the
> > switch before November 30.
>
> English letters? You mean digraphs ("u:")? Or is there other functionality
> you wanted tested there? (I know you're not directly talking to me, just
> wondering.)

I just mean: a-zA-Z0-9@#_&-+()/*"':;!?...
Maybe you could call it ASCII.

> > Afterwards, when Heiko has finished version 1.8 of my little forum,
> there
> > will be a new "forum2" for testing UTF-8 stuff and the like.
> >
> > Any objections?
>
> No, but there was never much non-English posting here, was there? (I don't
> and shouldn't demand it, obviously.)

Yes. One of this forum's rules is to write in English and thankfully everybody does so. It's the "lowest common denominator" for multinational communication on the Internet.

---
Forum admin

> English letters? You mean digraphs ("u:")? Or is there other functionality
> you wanted tested there? (I know you're not directly talking to me, just
> wondering.)

English has diaeresis/trema too, though usually it is used for place-names that don't follow modern English orthography. But yes, English has accents!

>
> Besides, (just to pretend to be smart) it's a modified Roman alphabet. (The
> Romans didn't have 'j', 'u', or 'w'. Even Esperanto calls 'w' "gxermana
> vo", i.e. "Germanic".)
>
> Someone should ask MarcoV (or maybe Eric Auer) for more advice. They're
> more educated than I am. (I found a book called _Old English and Its
> Closest Relatives_ for sale online, which basically compares "Gothic, Old
> Norse, Old Saxon, Old English, Old Frisian, Old Low Franconian, and Old
> High German"!!)

So middle (e.g. Franconian) dialects are discriminated against, while at the same time further removed dialects like High German and even non Western Germanic Gothic and Norse are included. Odd. Everybody seems to forget Luxembourg in Germanic language discussions.

I'm btw no linguist. I just started reading linguist (Germanisti(e)k) news and articles to keep track of a local dialect.

> I pointed Eric to
> Beowulf online a while back, saying it
> (unsurprisingly??) sounded like Swedish (although I don't speak/read that
> either).

Keep in mind that in the time of Beowulf, that many of the current separations of (West-) Germanic languages didn't exist. Coastal Dutch and German were mostly mutually intelligible with English.

Norse is afaik often dragged into the fold, even in discussions about West Germanic, because it has much newer and complete (up to the 12th century) sources for old Epics, while the West Germanic stuff is more like 5th.

Gothic and Burgundian are dead. Gothic is often dragged in because some of the earliest Germanic sources are Gothic. Burgundian sources are so scarce it is probably better to not mention it at all.

maltho thi afrio lito

> > there was never much non-English posting here, was there?
> > (I don't and shouldn't demand it, obviously.)
>
> Yes. One of this forum's rules is to write in English and thankfully
> everybody does so. It's the "lowest common denominator" for multinational
> communication on the Internet.

Debatable! But I won't object. Just saying, some people may prefer another.

> English has diaeresis/trema too, though usually it is used for place-names
> that don't follow modern English orthography. But yes, English has
> accents!

Obviously it's a very big world. The great thing about standards is that there are so many of them! So my experience is, of course, very limited. Having said that, accents are very rare (ignoring loan words or proper names). This keyboard has none, not even indicated with modifier keys. No one in my (small) part of the world uses them (in normal/average/common circumstances).

But other English speakers are free to do what they wish. (ASCII normally means 7-bit because we usually do without. Of course, that minimalism all went away with 8-bit chars and networking and beyond. Even lowly DOS isn't purely 7-bit, if you do the appropriate things.)

> > Someone should ask MarcoV (or maybe Eric Auer) for more advice. They're
> > more educated than I am. (I found a book called _Old English and Its
> > Closest Relatives_ for sale online, which basically compares "Gothic,
> > Old Norse, Old Saxon, Old English, Old Frisian, Old Low Franconian,
> > and Old High German"!!)
>
> So middle (e.g. Franconian) dialects are discriminated against, while at
> the same time further removed dialects like High German and even non
> Western Germanic Gothic and Norse are included. Odd. Everybody seems to
> forget Luxembourg in Germanic language discussions.

It's easy to forget when you're thousands of miles away, separated by ocean, and have no direct links. (Luxembourg is way more obscure and exotic than Germany or the Netherlands.) Not everyone has the nerdy (or scholarly) affection for it that we do.

> I'm btw no linguist. I just started reading linguist (Germanisti(e)k) news
> and articles to keep track of a local dialect.

Amateur! (Me, too.) It's usually good to be open to communication with others. Isolation isn't necessarily great, nor are incompatibilities.

> > I pointed Eric to Beowulf online a while back, saying it
> > (unsurprisingly??) sounded like Swedish (although I don't speak/read
> > that either).
>
> Keep in mind that in the time of Beowulf, that many of the current
> separations of (West-) Germanic languages didn't exist. Coastal Dutch and
> German were mostly mutually intelligible with English.

The U.S. DoD's Ada reduced their projects from using "over 450 [languages] in 1983 to 37 by 1996". Variety is good, and you don't want to step on anyone's toes, but there is a limit to how much can be properly supported. (But "Just use English! Now!!" is too gauche! Besides, it's far from simple and elegant ... although you can simplify and standardize any subset, if you try hard enough.)

> Norse is afaik often dragged into the fold, even in discussions about West
> Germanic, because it has much newer and complete (up to the 12th century)
> sources for old Epics, while the West Germanic stuff is more like 5th.

Well, the Nordic languages still exist, but clearly English, German, and Dutch are the most popular Germanic languages nowadays. (Sadly, few truly care even that much, unless forced.)

> Gothic and Burgundian are dead. Gothic is often dragged in because some of
> the earliest Germanic sources are Gothic. Burgundian sources are so scarce
> it is probably better to not mention it at all.

Requiescat in pace.

> maltho thi afrio lito

Gesundheit.

Please be so kind and open a new thread, if you would like to further discuss that language topic. Thanks.

---
Forum admin

I performed final migration from "PHP 5.6 + my little forum 1.7.6" to "PHP 7.3 + my little forum 1.7.8" today.

Notable differences:
- As mentioned in IMPORTANT: Forum migration in November 2019 version 1.7.8 lacks full UTF-8 support. So you better stick to non-accented English letters until I upgrade to version 1.8.0 in December 2019 or January 2020.
- Monospaced text is now slightly darker, because I like it more.
- Bottom link "powered by..." points to a valid URL.
- (Test installation "forum2" is offline now.)

Please report any errors to this thread.

Update 2019-12-06: Fixed "Last login" timestamp 8 users, which got corrupted by me during migration
Update 2019-12-02: My web hoster switched to PHP 7.3.11
Update 2019-11-26: Updated forum to version 1.7.11 to stay on track
Update 2019-11-22: Fixed "Mark thread" function for admins
Update 2019-11-19: Updated forum to version 1.7.9 to fix the admin panel
Update 2019-11-18: Fixed non-working "Admin contact" link

---
Forum admin

> I tried in Linux with Firefox ESR. Same problem. It seems to accept and
> show the characters correctly ... until you post, then it mangles them all
> into '?'. (I had thought HTML escapes like ? were working correctly,
> but apparently not.)

Uh ... did this change (get fixed)?? I think so! But we're not quite on 1.8 yet, so I dunno. Lemme test (HTML Unicode):

c circumflex : & #265; = ?
C circumflex : & #264; = ?
g circumflex : & #285; = ?
G circumflex : & #284; = ?
h circumflex : & #293; = ?
H circumflex : & #292; = ?
j circumflex : & #309; = ?
J circumflex : & #308; = ?
s circumflex : & #349; = ?
S circumflex : & #348; = ?
u breve : & #365; = ?
U breve : & #364; = ?

If you see these accents, good news!

> > (I had thought HTML escapes like ? were working correctly,
> > but apparently not.)

Gosh dangit! Why is this not working here? Look at my original message (and the reply), it (EDIT: lower c with circumflex, c^) shows up there correctly now! But when I try to do it here, it doesn't work! Very strange.

> > I tried in Linux with Firefox ESR. Same problem. It seems to accept and
> > show the characters correctly ... until you post, then it mangles them
> all
> > into '?'. (I had thought HTML escapes like ? were working correctly,
> > but apparently not.)
>
> Uh ... did this change (get fixed)??

No, I didn't change anything on that topic. You have to wait until v1.8.

---
Forum admin